Options price a 12% chance of SPY falling 10% in three months
What do option prices imply for the next three months, and does the site's own model add anything to them?
In brief
What the options market prices for one asset, beside the site's own model and its record- Options price a 12% chance that SPY ends three months more than 10% lower.→ Exhibit 1
- In SPY's forecasts the short-horizon band sits furthest from zero, at +0.07 standard units.→ Exhibit 3
- The engine's forecasts for 16 Feb 2027 put SPY's median at +8.0%.→ Exhibit 4
- The model's median path for SPY points +3.9% over three months.→ Exhibit 5
- Its paths called SPY's three-month direction 73% of the time, against 78% for always-up.→ Exhibit 6
| What would change this view | Trigger | Today | Decided by |
|---|---|---|---|
| The model's paths beat always-up | three-month direction called more often than always-up | 73% against 78% (17 Nov 2026)not crossed | each run |
| Market and model point opposite ways | the priced median and the model's median on opposite sides of today's price | priced +2.5%, model +3.9%not crossed | each run |
The reasoning behind each point
- PricedOptions price a 12% chance that SPY ends three months more than 10% lower.
What the market prices: the distribution of the price in 91 days that the session's option prices imply, 17 Nov 2026.
- ComponentsIn SPY's forecasts the short-horizon band sits furthest from zero, at +0.07 standard units.
The option-based forecasts at every horizon, summarised in three smoothed components.
- ForecastsThe engine's forecasts for 16 Feb 2027 put SPY's median at +8.0%.
The engine's forecast fan on 16 Feb 2027: 7,675 paths drawn from its stored forecasts; their 10th to 90th percentile runs from −2.4% to +17.9%, their mean +8.0%.
- ModelThe model's median path for SPY points +3.9% over three months.
The site's model (model-conditional): the median of 4,000 paths given the engine's two-week path; +5.8% without it.
- RecordIts paths called SPY's three-month direction 73% of the time, against 78% for always-up.
92 out-of-sample tests, re-fitted every month.
Next: Skew & term structure — which tail the same option prices favour
Options price a 12% chance of a 10% fall
Which way does the market lean over the next one and three months?Exhibit 1 · Options price a 12% chance that SPY ends three months more than 10% lower.
SPY is listed; it is drawn once the session folder ships its files.
About this exhibit
Exhibit 2 · One-month risk reversal against its three-year range
SPY is listed; it is drawn once the session folder ships its files.
The short-horizon band is doing the most work
Which shifts in the forecasts are doing the work now?Exhibit 3 · In SPY's forecasts the short-horizon band sits furthest from zero, at +0.07 standard units.
SPY is listed; it is drawn once the session folder ships its files.
About this exhibit
The model's paths have not beaten always-up
Where do the engine's forecasts and the model's paths point, and how good have they been?Exhibit 4 · The engine's forecasts for 16 Feb 2027 put SPY's median at +8.0%
SPY is listed; it is drawn once the session folder ships its files.
About this exhibit
Exhibit 5 · The model's median path for SPY points +3.9% over three months
SPY is listed; it is drawn once the session folder ships its files.
About this exhibit
Exhibit 6 · Its paths called SPY's three-month direction 73% of the time
SPY is listed; it is drawn once the session folder ships its files.
About this exhibit
Past entries into today's regime moved a median +10.3% over 13 weeks
What happened after past readings like today's?Exhibit 7 · Price paths after past entries into today's regime
SPY is listed; it is drawn once the session folder ships its files.
Exhibit 8 · Price paths while the regime lasted and after it ended
SPY is listed; it is drawn once the session folder ships its files.
Exhibit 9 · What followed days like today, by whether the regime ended
SPY is listed; it is drawn once the session folder ships its files.
Every covered asset
How do the assets compare on the same measures?
| Asset | Group | 10% fall priced, 3m | Model, 3m | Model alone | Called | Always-up | Fit | Tests |
|---|---|---|---|---|---|---|---|---|
| Equity index | 12% | +3.9% | +5.8% | 73% | 78% | −0.31 | 92 | |
| Equity index | 19% | +4.7% | +7.2% | 71% | 76% | −0.22 | 91 | |
| Equity index | 15% | −2.1% | +2.9% | 48% | 60% | −0.70 | 92 | |
| Equity index | 11% | +4.6% | +5.9% | 68% | 73% | −0.03 | 92 | |
| Equity index | 7% | +4.2% | +4.7% | 67% | 73% | −0.11 | 92 | |
| Sectors | – | +2.1% | +1.9% | 53% | 62% | −0.27 | 92 | |
| Sectors | 9% | +0.8% | +4.5% | 70% | 74% | −0.24 | 50 | |
| Sectors | 17% | +5.1% | +5.8% | 47% | 70% | −0.21 | 30 | |
| Sectors | – | +1.3% | +1.9% | 62% | 75% | −0.13 | 91 | |
| Sectors | – | +2.8% | +5.5% | 60% | 72% | −0.02 | 92 | |
| Sectors | – | +3.8% | +8.6% | 68% | 73% | −0.12 | 92 | |
| Sectors | – | +1.3% | +0.2% | 72% | 74% | −0.01 | 92 | |
| Sectors | – | +1.6% | +1.9% | 62% | 60% | −0.02 | 81 | |
| Sectors | 8% | −2.8% | −2.4% | 59% | 67% | 0.01 | 92 | |
| Sectors | – | +2.1% | −0.9% | 62% | 66% | −0.04 | 91 | |
| Sectors | – | +1.5% | +2.8% | 62% | 68% | −0.23 | 92 | |
| Industries and themes | 31% | −3.5% | +4.3% | 70% | 74% | −0.17 | 91 | |
| Industries and themes | 33% | +1.8% | +9.2% | 62% | 67% | −0.30 | 69 | |
| Industries and themes | – | +4.5% | +8.7% | 64% | 67% | −0.08 | 91 | |
| Industries and themes | – | +2.7% | +3.1% | 46% | 61% | −0.08 | 90 | |
| Industries and themes | – | +1.8% | +2.5% | 59% | 72% | −0.25 | 88 | |
| Industries and themes | 27% | +1.4% | +3.9% | 51% | 56% | −0.15 | 85 | |
| Industries and themes | – | +2.4% | +5.0% | 53% | 65% | −0.19 | 86 | |
| Industries and themes | 29% | +3.4% | +12.4% | 45% | 64% | −0.12 | 86 |
Appendix
How the distribution is drawn from option prices, how the paths are modelled, and what each check holds4,866 checks, all pass
What the options market prices for one asset over the next one and three months, beside the site's own model of where its price goes and that model's record. The distribution comes from the latest session's option prices; the paths are model-conditional, made from the option-based forecasts' components, with and without the engine's two-week path, day by day for three months. The record sits open beside the paths, because a range of paths is only worth what its record shows.
Model-conditional: the paths come from a daily model of the return on the three smoothed components of the option-based forecasts, its lag chosen by the Bayesian information criterion (at most 21 sessions), and 4,000 draws from its forecast distribution with a fixed seed, 63 sessions (three months) out. The conditional paths hold the components to the engine's path for ten sessions — its trend slope decaying at the rate measured from the slope's own history — and leave the return free; after that the model runs on its own dynamics.
The price distribution is what the market prices: per expiry, the forward from put-call parity and a smile in total variance against log-moneyness (the raw SVI form); at 30 and 91 days the smile is interpolated in total variance between the bracketing expiries and the density read from the call prices' curvature across strikes. Two penalties in the fit keep the density a density: the wings' slope bound and no butterfly arbitrage across the drawn strikes. A horizon is printed only when its density sums to one, stays non-negative and is centred on the forward. Its smile must also sit within half a vol point of the quoted at-the-money volatility, read at the forward from the six nearest quotes. A horizon the chain does not bracket is not extrapolated.
The closes are the engine's adjusted series: splits and distributions are carried back, so a return across an ex-date is a total return. The engine's forecasts, its component scores and its fan are made on the same series. The price distribution is read against the session's traded close, the price the options are struck in. The engine's own test for a change of adjustment basis runs on every asset's closes. The lead asset must pass it, an asset with a split-sized break is left out by name, and a smaller one is listed below with its date.
The engine's own test finds a change of adjustment basis in the closes of 11 assets, each smaller than the 25% at which an asset is left out: DIA 21 Sep 2026 (0.02%); TLT 5 Oct 2026 (0.42%); IEF 2 Sep 2026 (0.34%); IEF 5 Oct 2026 (0.34%); TIP 5 Oct 2026 (0.15%); FXA 5 Oct 2026 (0.15%); CSCO 2 Sep 2026 (0.39%); DELL 21 Sep 2026 (0.14%); NOK 28 Sep 2026 (0.38%); GEV 28 Sep 2026 (0.06%); EQT 5 Oct 2026 (0.32%); BX 5 Oct 2026 (0.94%). The engine's test for a change of adjustment basis was not run on 22 assets, so their closes are not tested: FX_AUDCAD, FX_AUDCHF, FX_AUDJPY, FX_AUDX, FX_CADCHF, FX_CADJPY, FX_CADX, FX_CHFJPY, FX_CHFX, FX_EURAUD, FX_EURCAD, FX_EURCHF, FX_EURGBP, FX_EURJPY, FX_EURX, FX_GBPAUD, FX_GBPCAD, FX_GBPCHF, FX_GBPJPY, FX_GBPX, FX_GLDSLV, FX_JPYX. It does not apply to an asset composed as a ratio of two funds. For 49 assets some of the engine's stored forecasts start from a price that is no longer the close on their date. A distribution or a split since they were stored moved the closes and not the stored forecasts, so the engine's scores and fan carry the gap. Largest: FX_GLDSLV 2.58% (45 of 128,610 forecasts); OWL 2.27% (11,023 of 12,068 forecasts); PFE 1.46% (18,522 of 22,689 forecasts); FX_GBPCHF 1.31% (52,014 of 79,641 forecasts); FX_GBPJPY 1.30% (78,393 of 105,993 forecasts); FX_GBPAUD 1.17% (104,871 of 132,498 forecasts); FX_GBPCAD 1.12% (52,014 of 79,641 forecasts); FX_GBPX 0.97% (79,272 of 105,849 forecasts); ….
The risk reversal, the PCA series, the engine's forecasts and the model's paths share one date axis: three years of sessions to the session, then the 63 weekdays of the paths ahead, so a date sits at the same place on each.
The PCA series are the engine's own, copied: each day it projects the option-based forecasts at every horizon onto two fixed sets of loadings. The step PCs (SPC_G1–G3) average the forecasts within three horizon bands, 5–25, 50–100 and 125–250 days. The orthogonal PCs (PC1–PC3) project the whole forecast curve onto level, slope and curvature. The card draws the raw daily score and the engine's smoothed score on either set, over three years.
Model-conditional: the fan is the engine's own, copied and not recomputed. It is the fan the engine's own publishing step draws from the engine's files up to this session. The engine's close after this session was not yet settled, so the step was run on its files cut at the session. The engine takes each stored forecast with more than 21 days to run and draws 25 straight paths from it. Each path runs from the close on the forecast date to an end price shifted by a random error the size of its forecasts' past misses. On each date the fan gathers every path that crosses it and weights later forecasts more, with a half-life of 60 days. The card prints the mean, the median, the 40th–60th, 25th–75th and 10th–90th percentiles and the mean plus and minus two standard deviations, beside the adjusted close. The engine holds every price of its fan between zero and 20 times its latest close. The card runs three years back and, ahead, as far as the paths: three months. A fan whose past closes are not this session's closes is not drawn, and the card says so.
The risk reversal is the engine's own 25-delta reading, put less call implied volatility in vol points: 30 days is read linearly between its 25- and 50-day points, daily, over three years, against its own 10th, 50th and 90th percentiles. Positive means puts cost more than calls.
Model-conditional: the paths come from a daily model of the return on the three smoothed components of the option-based forecasts, its lag chosen by the Bayesian information criterion (at most 21 sessions), and 4,000 draws from its forecast distribution with a fixed seed, 63 sessions (three months) out. The conditional paths hold the components to the engine's path for ten sessions — its trend slope decaying at the rate measured from the slope's own history — and leave the return free; after that the model runs on its own dynamics.
The record re-fits the model on the data to each origin, every 21 sessions from the 504th session (252nd for a shorter history), and scores the 63-session return against the historical average and against always-up.
The episodes are the engine's. Its regime is the sign of each smoothed component, and an episode is a run of days in one regime, out of sample only. The card has two ways to follow the past entries into today's regime, beside the current episode from its own entry close. Extended tracking follows every entry for 65 sessions on the engine's closes, whether or not the regime lasted. Its band, median and mean are taken on each day across every entry with a close that day. Its dash-dot line is the median of any 65 sessions that start between the first entry and the last. In regime only follows each episode for as long as it lasted. Its band, median and mean need five episodes still running. On days with two to four the median and the mean continue, faded, without the band, so the later days rest on the longer episodes only. Each entry's path on the closes equals the engine's own path for as long as its episode ran. Where that cannot be shown for every past entry of an asset, its entries are not followed and the card says why. The lower chart divides each return by the volatility at that entry. That is the standard deviation of the 63 daily log returns before the entry day, annualised, so a calm and a turbulent start are read on one scale.
The two groups are the same entries, split on each day by whether the regime was still running. An episode is in the first group for as long as it lasted and in the second from the session after. Each group's median and its 10th to 90th percentile are drawn on days when it holds five or more entries. Which group an entry would fall in was not known on the day, so the split is hindsight. It describes what the regime's end went with, and is not a rule to act on.
The table counts what followed sessions like today's, on every covered asset together. A session is like today's when the asset was in the same regime. It counts only when the sign of its three components is the engine's regime that day. The rows sort those sessions by the distance of the component nearest zero. A component's distance is its score divided by that score's own standard deviation over every session before it (126 or more). For each row the table gives the share of sessions on which the regime ended within 21 sessions. It also gives the mean return over those 21 sessions when the regime ended, when it did not and on all of them. Returns are in units of the asset's usual move: the daily volatility of the 63 sessions before the session × √21. The last 21 sessions of a regime still running are left out, because their outcome is not known yet. A share or a mean over fewer than 30 sessions is not printed. The sessions overlap and the assets move together, so the rows are counts of what happened, not independent tests and not a forecast.
Days to expiry are calendar days (τ = days ÷ 365). The site's realised volatility is √252 × the sample standard deviation of daily log closes, overnight returns included (the author's decision of 3 Oct 2026 (D-2026-10-03-19)); the options engine states its own as annualised by None on trading days, overnight return excluded. The quotes are the vendor's end-of-day set for the session, stamped 16:00 New York, 17 Nov 2026; the options engine's results for a session are read only when they are complete and were written after the close had settled (the author's decision of 4 Oct 2026 (D-2026-10-04-02)).
Option data: marketdata.app. Every figure derived from it on this page is computed here.
This page follows the author's decision of 4 Oct 2026 (D-2026-10-04-01) and the author's decision of 3 Oct 2026 (D-2026-10-03-21) and the author's decision of 4 Oct 2026 (D-2026-10-04-05) and the author's decision of 4 Oct 2026 (D-2026-10-04-06) and the author's decision of 4 Oct 2026 (D-2026-10-04-08) and the author's decision of 7 Oct 2026 (D-2026-10-07-SKEW1).
How the numbers are made
session_folder ──▶ core.imports: the newest complete folder on or before the run day; every file's SHA-256 against checks.json ──▶ the verified folder the verified folder ──▶ per asset and expiry: the forward from put-call parity and the smile in total variance; the price distribution at 30 and 91 days from the call prices' curvature ──▶ what the market prices the verified folder ──▶ a daily model of the return on the three components; its 63-session paths alone and given the engine's ten-session path; the expanding 63-session record ──▶ the model's paths and record the verified folder ──▶ the engine's own published fan, copied: on each date every path its stored forecasts draw across it, their mean, median and 10th–90th range; drawn only where its past closes are the folder's ──▶ the engine's forecasts the verified folder ──▶ the engine's 25-delta risk reversal at 30 days (linear between its 25- and 50-day points), three years ──▶ the market's lean the verified folder ──▶ the engine's regime episodes: the past entries into today's regime, each followed 65 sessions on the engine's closes and for as long as its episode lasted, the current episode from its entry close, and each path in units of the volatility at its entry (63 sessions before it) ──▶ what followed similar states the verified folder ──▶ the same entries by whether the regime had ended; and, over every covered asset, the sessions in today's regime by the nearest component's distance from zero: how often the regime ended within 21 sessions and the return that went with it ──▶ the two groups and the fork
Sources
- licensedOptions engine session folder: the latest session's end-of-day option chains, the option-based forecasts' component scores and trend slopes, and adjusted closes
marketdata.app
Reproduction checks
| Check | Published | Computed | Tolerance | Result |
|---|---|---|---|---|
| Outlook folder 2026-11-17 · _SUCCESS: present: the engine writes it last, and a folder without it is not read | 1 | 1 | 0 | pass |
| Outlook folder 2026-11-17 · chain_slice.parquet: SHA-256 equals the hash checks.json lists for it | 1 | 1 | 0 | pass |
| Outlook folder 2026-11-17 · closes.parquet: SHA-256 equals the hash checks.json lists for it | 1 | 1 | 0 | pass |
| Outlook folder 2026-11-17 · components.parquet: SHA-256 equals the hash checks.json lists for it | 1 | 1 | 0 | pass |
| Outlook folder 2026-11-17 · episode_paths.parquet: SHA-256 equals the hash checks.json lists for it | 1 | 1 | 0 | pass |
| Outlook folder 2026-11-17 · episodes.parquet: SHA-256 equals the hash checks.json lists for it | 1 | 1 | 0 | pass |
| Outlook folder 2026-11-17 · fan.parquet: SHA-256 equals the hash checks.json lists for it | 1 | 1 | 0 | pass |
| Outlook folder 2026-11-17 · risk_reversal.parquet: SHA-256 equals the hash checks.json lists for it | 1 | 1 | 0 | pass |
| Outlook folder 2026-11-17 · checks.json: session_date is the folder's name | 1 | 1 | 0 | pass |
| Outlook folder 2026-11-17 · checks.json: session_date is a weekday (a US session) | 1 | 1 | 0 | pass |
| Outlook folder 2026-11-17 · checks.json: session_date is on or before the run day 2026-11-20 | 1 | 1 | 0 | pass |
| Outlook folder 2026-11-17 · checks.json: quote_time_utc is at or after the session's 16:00 New York close (the vendor's end-of-day quotes) | 1 | 1 | 0 | pass |
| Outlook folder 2026-11-17 · checks.json: written_utc is at or after the session's 16:00 New York close plus 15 minutes (written once the session had settled) | 1 | 1 | 0 | pass |
| Outlook folder 2026-11-17 · checks.json: written_utc is not after the moment of the check | 1 | 1 | 0 | pass |
| Outlook folder 2026-11-17 · checks.json: quote_time_utc is not after the moment of the check | 1 | 1 | 0 | pass |
| Outlook folder 2026-11-17 · checks.json: every one of the export's gate checks passed, standing in for the engine's own (2 listed) | 1 | 1 | 0 | pass |
| The session folder ships chain_slice.parquet, which this page reads | 1 | 1 | 0 | pass |
| The session folder ships components.parquet, which this page reads | 1 | 1 | 0 | pass |
| The session folder ships closes.parquet, which this page reads | 1 | 1 | 0 | pass |
| The session folder ships fan.parquet, which this page reads (D-OPT-OL4) | 1 | 1 | 0 | pass |
| The session folder ships episodes.parquet, which this page reads (D-OPT-OL4) | 1 | 1 | 0 | pass |
| The session folder ships episode_paths.parquet, which this page reads (D-OPT-OL4) | 1 | 1 | 0 | pass |
| The session folder ships risk_reversal.parquet, which this page reads (D-OPT-OL4) | 1 | 1 | 0 | pass |
| The session folder states that its closes are the engine's adjusted series, with the traded close beside them (asked of every session from 11 Nov 2026; a session before it is read as written) | 1 | 1 | 0 | pass |
| The session folder's closes carry the traded close, the price the options are struck in | 1 | 1 | 0 | pass |
| SPY, 30 days: the price distribution's mass (target 1.0, tolerance 0.001) | 1 | 0.999999970204 | 0.001 | pass |
| SPY, 30 days: the price distribution's no negative (target 0.0, tolerance 0.005) | 0 | -1.558e-7 | 0.005 | pass |
| SPY, 30 days: the price distribution's mean forward (target 0.0, tolerance 0.001) | 0 | 2.8251e-8 | 0.001 | pass |
| SPY, 30 days: the price distribution's atm iv (target 0.0, tolerance 0.005) | 0 | -0.000354824423 | 0.005 | pass |
| SPY, 91 days: the price distribution's mass (target 1.0, tolerance 0.001) | 1 | 0.999999969594 | 0.001 | pass |
| SPY, 91 days: the price distribution's no negative (target 0.0, tolerance 0.005) | 0 | 0 | 0.005 | pass |
| SPY, 91 days: the price distribution's mean forward (target 0.0, tolerance 0.001) | 0 | 2.996e-8 | 0.001 | pass |
| SPY, 91 days: the price distribution's atm iv (target 0.0, tolerance 0.005) | 0 | -0.000874982682 | 0.005 | pass |
| SPY, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| SPY, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| SPY, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| SPY, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| SPY: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| SPY: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| SPY: the 30-day risk reversal is read only on days carrying both grid points (days dropped: 0; none invented) | 0 | 0 | 0 | pass |
| SPY, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| SPY, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| SPY: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 9 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| SPY: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| SPY: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| SPY: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 8 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| SPY: the entries followed never grow in number from one session to the next, and all 8 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| SPY: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (756 stretches) | 0.05635858813 | 0.05635858813 | 1e-12 | pass |
| SPY: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| SPY: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| SPY: a component's distance from zero uses only the sessions before it (the history cut at 11 dates against the full series; dates that differ) | 0 | 0 | 0 | pass |
| SPY: the fork's usual move is the page's volatility at entry, the 63 sessions' log returns before the day (at its 119 episode entries; max abs, annualised) | 0 | 0 | 1e-12 | pass |
| SPY: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 25 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| SPY: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| SPY: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| SPY: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 24 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| SPY: the entries followed never grow in number from one session to the next, and all 24 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| SPY: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (2028 stretches) | 0.049324051822 | 0.049324051822 | 1e-12 | pass |
| SPY: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| SPY: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| QQQ, 30 days: the price distribution's mass (target 1.0, tolerance 0.001) | 1 | 0.999999936105 | 0.001 | pass |
| QQQ, 30 days: the price distribution's no negative (target 0.0, tolerance 0.005) | 0 | -3.1887e-8 | 0.005 | pass |
| QQQ, 30 days: the price distribution's mean forward (target 0.0, tolerance 0.001) | 0 | 6.2579e-8 | 0.001 | pass |
| QQQ, 30 days: the price distribution's atm iv (target 0.0, tolerance 0.005) | 0 | -0.001915183628 | 0.005 | pass |
| QQQ, 91 days: the price distribution's mass (target 1.0, tolerance 0.001) | 1 | 0.999998780284 | 0.001 | pass |
| QQQ, 91 days: the price distribution's no negative (target 0.0, tolerance 0.005) | 0 | 0 | 0.005 | pass |
| QQQ, 91 days: the price distribution's mean forward (target 0.0, tolerance 0.001) | 0 | 0.000001201535 | 0.001 | pass |
| QQQ, 91 days: the price distribution's atm iv (target 0.0, tolerance 0.005) | 0 | -0.000937601284 | 0.005 | pass |
| QQQ, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| QQQ, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| QQQ, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| QQQ, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| QQQ: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| QQQ: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| QQQ: the 30-day risk reversal is read only on days carrying both grid points (days dropped: 0; none invented) | 0 | 0 | 0 | pass |
| QQQ, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| QQQ, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| QQQ: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 8 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| QQQ: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| QQQ: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| QQQ: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 7 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| QQQ: the entries followed never grow in number from one session to the next, and all 7 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| QQQ: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (1718 stretches) | 0.06054522611 | 0.06054522611 | 1e-12 | pass |
| QQQ: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| QQQ: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| QQQ: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 30 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| QQQ: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| QQQ: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| QQQ: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 29 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| QQQ: the entries followed never grow in number from one session to the next, and all 29 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| QQQ: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (2141 stretches) | 0.056622802399 | 0.056622802399 | 1e-12 | pass |
| QQQ: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| QQQ: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| IWM, 30 days: the price distribution's mass (target 1.0, tolerance 0.001) | 1 | 0.999999991196 | 0.001 | pass |
| IWM, 30 days: the price distribution's no negative (target 0.0, tolerance 0.005) | 0 | 0 | 0.005 | pass |
| IWM, 30 days: the price distribution's mean forward (target 0.0, tolerance 0.001) | 0 | 7.76e-9 | 0.001 | pass |
| IWM, 30 days: the price distribution's atm iv (target 0.0, tolerance 0.005) | 0 | -0.001296561054 | 0.005 | pass |
| IWM, 91 days: the price distribution's mass (target 1.0, tolerance 0.001) | 1 | 0.999999682163 | 0.001 | pass |
| IWM, 91 days: the price distribution's no negative (target 0.0, tolerance 0.005) | 0 | -8.78263e-7 | 0.005 | pass |
| IWM, 91 days: the price distribution's mean forward (target 0.0, tolerance 0.001) | 0 | 2.95416e-7 | 0.001 | pass |
| IWM, 91 days: the price distribution's atm iv (target 0.0, tolerance 0.005) | 0 | -0.003120265145 | 0.005 | pass |
| IWM, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| IWM, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| IWM, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| IWM, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| IWM: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| IWM: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| IWM: the 30-day risk reversal is read only on days carrying both grid points (days dropped: 0; none invented) | 0 | 0 | 0 | pass |
| IWM, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| IWM, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| IWM: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 16 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| IWM: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| IWM: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| IWM: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 15 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| IWM: the entries followed never grow in number from one session to the next, and all 15 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| IWM: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (1627 stretches) | 0.035108673185 | 0.035108673185 | 1e-12 | pass |
| IWM: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| IWM: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| IWM: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 21 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| IWM: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| IWM: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| IWM: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 20 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| IWM: the entries followed never grow in number from one session to the next, and all 20 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| IWM: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (1824 stretches) | 0.033061669918 | 0.033061669918 | 1e-12 | pass |
| IWM: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| IWM: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| DIA, 30 days: the price distribution's mass (target 1.0, tolerance 0.001) | 1 | 0.999999998247 | 0.001 | pass |
| DIA, 30 days: the price distribution's no negative (target 0.0, tolerance 0.005) | 0 | 0 | 0.005 | pass |
| DIA, 30 days: the price distribution's mean forward (target 0.0, tolerance 0.001) | 0 | -6.485e-9 | 0.001 | pass |
| DIA, 30 days: the price distribution's atm iv (target 0.0, tolerance 0.005) | 0 | -0.003538981418 | 0.005 | pass |
| DIA, 91 days: the price distribution's mass (target 1.0, tolerance 0.001) | 1 | 0.999999975478 | 0.001 | pass |
| DIA, 91 days: the price distribution's no negative (target 0.0, tolerance 0.005) | 0 | 0 | 0.005 | pass |
| DIA, 91 days: the price distribution's mean forward (target 0.0, tolerance 0.001) | 0 | 2.3467e-8 | 0.001 | pass |
| DIA, 91 days: the price distribution's atm iv (target 0.0, tolerance 0.005) | 0 | -0.000664713707 | 0.005 | pass |
| DIA, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| DIA, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| DIA, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| DIA, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| DIA: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| DIA: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| DIA: the 30-day risk reversal is read only on days carrying both grid points (days dropped: 0; none invented) | 0 | 0 | 0 | pass |
| DIA, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| DIA, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| DIA: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 20 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| DIA: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| DIA: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| DIA: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 19 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| DIA: the entries followed never grow in number from one session to the next, and all 19 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| DIA: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (2037 stretches) | 0.034264730058 | 0.034264730058 | 1e-12 | pass |
| DIA: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| DIA: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| DIA: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 29 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| DIA: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| DIA: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| DIA: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 28 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| DIA: the entries followed never grow in number from one session to the next, and all 28 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| DIA: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (2141 stretches) | 0.034367136904 | 0.034367136904 | 1e-12 | pass |
| DIA: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| DIA: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| RSP, 91 days: the price distribution's mass (target 1.0, tolerance 0.001) | 1 | 0.99996014147 | 0.001 | pass |
| RSP, 91 days: the price distribution's no negative (target 0.0, tolerance 0.005) | 0 | 0 | 0.005 | pass |
| RSP, 91 days: the price distribution's mean forward (target 0.0, tolerance 0.001) | 0 | 0.000039372053 | 0.001 | pass |
| RSP, 91 days: the price distribution's atm iv (target 0.0, tolerance 0.005) | 0 | -0.004007043937 | 0.005 | pass |
| RSP, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| RSP, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| RSP, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| RSP, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| RSP: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| RSP: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| RSP: the 30-day risk reversal is read only on days carrying both grid points (days dropped: 0; none invented) | 0 | 0 | 0 | pass |
| RSP, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| RSP, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| RSP: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 21 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| RSP: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| RSP: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| RSP: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 20 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| RSP: the entries followed never grow in number from one session to the next, and all 20 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| RSP: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (1998 stretches) | 0.035289315971 | 0.035289315971 | 1e-12 | pass |
| RSP: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| RSP: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| RSP: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 6 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| RSP: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| RSP: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| RSP: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 5 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| RSP: the entries followed never grow in number from one session to the next, and all 5 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| RSP: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (2011 stretches) | 0.034721196395 | 0.034721196395 | 1e-12 | pass |
| RSP: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| RSP: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| XLB, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| XLB, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| XLB, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| XLB, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| XLB: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| XLB: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| XLB: the 30-day risk reversal is read only on days carrying both grid points (days dropped: 0; none invented) | 0 | 0 | 0 | pass |
| XLB, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| XLB, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| XLB: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 23 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| XLB: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| XLB: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| XLB: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 22 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| XLB: the entries followed never grow in number from one session to the next, and all 22 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| XLB: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (2162 stretches) | 0.022706091849 | 0.022706091849 | 1e-12 | pass |
| XLB: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| XLB: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| XLB: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 26 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| XLB: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| XLB: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| XLB: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 25 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| XLB: the entries followed never grow in number from one session to the next, and all 25 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| XLB: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (1756 stretches) | 0.025048026687 | 0.025048026687 | 1e-12 | pass |
| XLB: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| XLB: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| XLC, 91 days: the price distribution's mass (target 1.0, tolerance 0.001) | 1 | 0.999820392936 | 0.001 | pass |
| XLC, 91 days: the price distribution's no negative (target 0.0, tolerance 0.005) | 0 | 0 | 0.005 | pass |
| XLC, 91 days: the price distribution's mean forward (target 0.0, tolerance 0.001) | 0 | 0.000177072215 | 0.001 | pass |
| XLC, 91 days: the price distribution's atm iv (target 0.0, tolerance 0.005) | 0 | -0.000340518279 | 0.005 | pass |
| XLC, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| XLC, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| XLC, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| XLC, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| XLC: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| XLC: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| XLC: the 30-day risk reversal is read only on days carrying both grid points (days dropped: 0; none invented) | 0 | 0 | 0 | pass |
| XLC, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| XLC, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| XLC: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 16 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| XLC: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| XLC: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| XLC: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 15 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| XLC: the entries followed never grow in number from one session to the next, and all 15 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| XLC: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (1196 stretches) | 0.038376809809 | 0.038376809809 | 1e-12 | pass |
| XLC: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| XLC: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| XLC: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 19 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| XLC: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| XLC: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| XLC: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 18 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| XLC: the entries followed never grow in number from one session to the next, and all 18 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| XLC: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (753 stretches) | 0.048867338525 | 0.048867338525 | 1e-12 | pass |
| XLC: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| XLC: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| XLE, 91 days: the price distribution's mass (target 1.0, tolerance 0.001) | 1 | 0.999357043165 | 0.001 | pass |
| XLE, 91 days: the price distribution's no negative (target 0.0, tolerance 0.005) | 0 | 0 | 0.005 | pass |
| XLE, 91 days: the price distribution's mean forward (target 0.0, tolerance 0.001) | 0 | 0.00063723285 | 0.001 | pass |
| XLE, 91 days: the price distribution's atm iv (target 0.0, tolerance 0.005) | 0 | -0.000699376036 | 0.005 | pass |
| XLE, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| XLE, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| XLE, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| XLE, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| XLE: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| XLE: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| XLE: the 30-day risk reversal is read only on days carrying both grid points (days dropped: 0; none invented) | 0 | 0 | 0 | pass |
| XLE, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| XLE, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| XLE: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 8 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| XLE: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| XLE: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| XLE: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 7 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| XLE: the entries followed never grow in number from one session to the next, and all 7 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| XLE: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (330 stretches) | -0.009426126436 | -0.009426126436 | 1e-12 | pass |
| XLE: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| XLE: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| XLE: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 16 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| XLE: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| XLE: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| XLE: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 15 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| XLE: the entries followed never grow in number from one session to the next, and all 15 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| XLE: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (609 stretches) | 0.033046691519 | 0.033046691519 | 1e-12 | pass |
| XLE: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| XLE: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| XLF, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| XLF, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| XLF, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| XLF, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| XLF: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| XLF: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| XLF: the 30-day risk reversal is read only on days carrying both grid points (days dropped: 0; none invented) | 0 | 0 | 0 | pass |
| XLF, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| XLF, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| XLF: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 25 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| XLF: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| XLF: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| XLF: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 24 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| XLF: the entries followed never grow in number from one session to the next, and all 24 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| XLF: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (2136 stretches) | 0.036043592525 | 0.036043592525 | 1e-12 | pass |
| XLF: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| XLF: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| XLF: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 30 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| XLF: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| XLF: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| XLF: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 29 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| XLF: the entries followed never grow in number from one session to the next, and all 29 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| XLF: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (2127 stretches) | 0.036330671434 | 0.036330671434 | 1e-12 | pass |
| XLF: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| XLF: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| XLI, 30 days: the price distribution's mass (target 1.0, tolerance 0.001) | 1 | 0.999998784745 | 0.001 | pass |
| XLI, 30 days: the price distribution's no negative (target 0.0, tolerance 0.005) | 0 | 0 | 0.005 | pass |
| XLI, 30 days: the price distribution's mean forward (target 0.0, tolerance 0.001) | 0 | 1.83596e-7 | 0.001 | pass |
| XLI, 30 days: the price distribution's atm iv (target 0.0, tolerance 0.005) | 0 | 0.000543789406 | 0.005 | pass |
| XLI, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| XLI, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| XLI, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| XLI, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| XLI: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| XLI: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| XLI: the 30-day risk reversal is read only on days carrying both grid points (days dropped: 0; none invented) | 0 | 0 | 0 | pass |
| XLI, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| XLI, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| XLI: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 31 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| XLI: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| XLI: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| XLI: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 30 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| XLI: the entries followed never grow in number from one session to the next, and all 30 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| XLI: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (2143 stretches) | 0.034097249601 | 0.034097249601 | 1e-12 | pass |
| XLI: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| XLI: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| XLI: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 30 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| XLI: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| XLI: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| XLI: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 29 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| XLI: the entries followed never grow in number from one session to the next, and all 29 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| XLI: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (2142 stretches) | 0.034101954614 | 0.034101954614 | 1e-12 | pass |
| XLI: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| XLI: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| XLK, 30 days: the price distribution's mass (target 1.0, tolerance 0.001) | 1 | 0.99999999954 | 0.001 | pass |
| XLK, 30 days: the price distribution's no negative (target 0.0, tolerance 0.005) | 0 | -3.335e-9 | 0.005 | pass |
| XLK, 30 days: the price distribution's mean forward (target 0.0, tolerance 0.001) | 0 | -2.3536e-8 | 0.001 | pass |
| XLK, 30 days: the price distribution's atm iv (target 0.0, tolerance 0.005) | 0 | 0.004506153666 | 0.005 | pass |
| XLK, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| XLK, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| XLK, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| XLK, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| XLK: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| XLK: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| XLK: the 30-day risk reversal is read only on days carrying both grid points (days dropped: 0; none invented) | 0 | 0 | 0 | pass |
| XLK, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| XLK, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| XLK: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 22 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| XLK: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| XLK: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| XLK: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 21 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| XLK: the entries followed never grow in number from one session to the next, and all 21 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| XLK: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (2145 stretches) | 0.064550203711 | 0.064550203711 | 1e-12 | pass |
| XLK: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| XLK: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| XLK: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 30 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| XLK: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| XLK: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| XLK: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 29 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| XLK: the entries followed never grow in number from one session to the next, and all 29 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| XLK: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (2130 stretches) | 0.065823424602 | 0.065823424602 | 1e-12 | pass |
| XLK: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| XLK: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| XLP, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| XLP, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| XLP, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| XLP, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| XLP: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| XLP: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| XLP: the 30-day risk reversal is read only on days carrying both grid points (days dropped: 0; none invented) | 0 | 0 | 0 | pass |
| XLP, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| XLP, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| XLP: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 29 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| XLP: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| XLP: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| XLP: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 28 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| XLP: the entries followed never grow in number from one session to the next, and all 28 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| XLP: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (2162 stretches) | 0.025829948188 | 0.025829948188 | 1e-12 | pass |
| XLP: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| XLP: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| XLP: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 36 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| XLP: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| XLP: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| XLP: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 35 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| XLP: the entries followed never grow in number from one session to the next, and all 35 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| XLP: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (2014 stretches) | 0.026854771289 | 0.026854771289 | 1e-12 | pass |
| XLP: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| XLP: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| XLRE, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| XLRE, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| XLRE, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| XLRE, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| XLRE: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| XLRE: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| XLRE: the 30-day risk reversal is read only on days carrying both grid points (days dropped: 0; none invented) | 0 | 0 | 0 | pass |
| XLRE, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| XLRE, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| XLRE: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 28 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| XLRE: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| XLRE: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| XLRE: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 27 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| XLRE: the entries followed never grow in number from one session to the next, and all 27 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| XLRE: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (1944 stretches) | 0.020775743606 | 0.020775743606 | 1e-12 | pass |
| XLRE: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| XLRE: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| XLRE: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 36 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| XLRE: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| XLRE: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| XLRE: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 35 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| XLRE: the entries followed never grow in number from one session to the next, and all 35 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| XLRE: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (1951 stretches) | 0.020995744145 | 0.020995744145 | 1e-12 | pass |
| XLRE: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| XLRE: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| XLU, 91 days: the price distribution's mass (target 1.0, tolerance 0.001) | 1 | 0.999982948764 | 0.001 | pass |
| XLU, 91 days: the price distribution's no negative (target 0.0, tolerance 0.005) | 0 | 0 | 0.005 | pass |
| XLU, 91 days: the price distribution's mean forward (target 0.0, tolerance 0.001) | 0 | 0.000014150848 | 0.001 | pass |
| XLU, 91 days: the price distribution's atm iv (target 0.0, tolerance 0.005) | 0 | -0.004173737988 | 0.005 | pass |
| XLU, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| XLU, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| XLU, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| XLU, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| XLU: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| XLU: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| XLU: the 30-day risk reversal is read only on days carrying both grid points (days dropped: 0; none invented) | 0 | 0 | 0 | pass |
| XLU, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| XLU, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| XLU: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 20 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| XLU: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| XLU: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| XLU: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 19 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| XLU: the entries followed never grow in number from one session to the next, and all 19 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| XLU: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (1759 stretches) | 0.035880444992 | 0.035880444992 | 1e-12 | pass |
| XLU: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| XLU: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| XLU: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 14 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| XLU: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| XLU: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| XLU: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 13 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| XLU: the entries followed never grow in number from one session to the next, and all 13 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| XLU: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (1528 stretches) | 0.026940177811 | 0.026940177811 | 1e-12 | pass |
| XLU: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| XLU: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| XLV, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| XLV, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| XLV, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| XLV, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| XLV: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| XLV: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| XLV: the 30-day risk reversal is read only on days carrying both grid points (days dropped: 0; none invented) | 0 | 0 | 0 | pass |
| XLV, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| XLV, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| XLV: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 12 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| XLV: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| XLV: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| XLV: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 11 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| XLV: the entries followed never grow in number from one session to the next, and all 11 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| XLV: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (1466 stretches) | 0.022660205581 | 0.022660205581 | 1e-12 | pass |
| XLV: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| XLV: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| XLV: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 32 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| XLV: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| XLV: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| XLV: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 31 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| XLV: the entries followed never grow in number from one session to the next, and all 31 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| XLV: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (2155 stretches) | 0.023326724162 | 0.023326724162 | 1e-12 | pass |
| XLV: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| XLV: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| XLY, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| XLY, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| XLY, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| XLY, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| XLY: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| XLY: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| XLY: the 30-day risk reversal is read only on days carrying both grid points (days dropped: 0; none invented) | 0 | 0 | 0 | pass |
| XLY, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| XLY, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| XLY: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 27 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| XLY: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| XLY: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| XLY: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 26 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| XLY: the entries followed never grow in number from one session to the next, and all 26 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| XLY: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (2069 stretches) | 0.043690236423 | 0.043690236423 | 1e-12 | pass |
| XLY: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| XLY: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| XLY: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 30 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| XLY: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| XLY: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| XLY: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 29 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| XLY: the entries followed never grow in number from one session to the next, and all 29 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| XLY: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (2084 stretches) | 0.043003674176 | 0.043003674176 | 1e-12 | pass |
| XLY: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| XLY: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| SMH, 91 days: the price distribution's mass (target 1.0, tolerance 0.001) | 1 | 0.999998891909 | 0.001 | pass |
| SMH, 91 days: the price distribution's no negative (target 0.0, tolerance 0.005) | 0 | -0.000223669794 | 0.005 | pass |
| SMH, 91 days: the price distribution's mean forward (target 0.0, tolerance 0.001) | 0 | -0.000098764639 | 0.001 | pass |
| SMH, 91 days: the price distribution's atm iv (target 0.0, tolerance 0.005) | 0 | 0.001656343567 | 0.005 | pass |
| SMH, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| SMH, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| SMH, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| SMH, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| SMH: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| SMH: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| SMH: the 30-day risk reversal is read only on days carrying both grid points (days dropped: 0; none invented) | 0 | 0 | 0 | pass |
| SMH, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| SMH, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| SMH: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 32 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| SMH: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| SMH: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| SMH: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 31 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| SMH: the entries followed never grow in number from one session to the next, and all 31 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| SMH: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (2162 stretches) | 0.082874643808 | 0.082874643808 | 1e-12 | pass |
| SMH: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| SMH: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| SMH: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 26 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| SMH: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| SMH: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| SMH: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 25 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| SMH: the entries followed never grow in number from one session to the next, and all 25 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| SMH: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (2146 stretches) | 0.084008577056 | 0.084008577056 | 1e-12 | pass |
| SMH: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| SMH: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| SOXX, 91 days: the price distribution's mass (target 1.0, tolerance 0.001) | 1 | 0.999999969281 | 0.001 | pass |
| SOXX, 91 days: the price distribution's no negative (target 0.0, tolerance 0.005) | 0 | 0 | 0.005 | pass |
| SOXX, 91 days: the price distribution's mean forward (target 0.0, tolerance 0.001) | 0 | -3.7973e-8 | 0.001 | pass |
| SOXX, 91 days: the price distribution's atm iv (target 0.0, tolerance 0.005) | 0 | -0.003315550886 | 0.005 | pass |
| SOXX, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| SOXX, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| SOXX, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| SOXX, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| SOXX: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| SOXX: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| SOXX: the 30-day risk reversal is read only on days carrying both grid points (days dropped: 0; none invented) | 0 | 0 | 0 | pass |
| SOXX, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| SOXX, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| SOXX: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 13 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| SOXX: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| SOXX: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| SOXX: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 12 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| SOXX: the entries followed never grow in number from one session to the next, and all 12 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| SOXX: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (1619 stretches) | 0.093416101847 | 0.093416101847 | 1e-12 | pass |
| SOXX: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| SOXX: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| SOXX: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 21 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| SOXX: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| SOXX: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| SOXX: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 20 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| SOXX: the entries followed never grow in number from one session to the next, and all 20 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| SOXX: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (1626 stretches) | 0.094665002422 | 0.094665002422 | 1e-12 | pass |
| SOXX: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| SOXX: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| IGV, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| IGV, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| IGV, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| IGV, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| IGV: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| IGV: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| IGV: the 30-day risk reversal is read only on days carrying both grid points (days dropped: 0; none invented) | 0 | 0 | 0 | pass |
| IGV, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| IGV, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| IGV: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 8 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| IGV: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| IGV: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| IGV: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 7 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| IGV: the entries followed never grow in number from one session to the next, and all 7 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| IGV: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (890 stretches) | 0.058004169611 | 0.058004169611 | 1e-12 | pass |
| IGV: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| IGV: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| IGV: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 27 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| IGV: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| IGV: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| IGV: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 26 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| IGV: the entries followed never grow in number from one session to the next, and all 26 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| IGV: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (1944 stretches) | 0.053796570221 | 0.053796570221 | 1e-12 | pass |
| IGV: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| IGV: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| KRE, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| KRE, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| KRE, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| KRE, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| KRE: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| KRE: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| KRE: the 30-day risk reversal is read only on days carrying both grid points (days dropped: 0; none invented) | 0 | 0 | 0 | pass |
| KRE, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| KRE, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| KRE: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 27 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| KRE: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| KRE: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| KRE: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 26 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| KRE: the entries followed never grow in number from one session to the next, and all 26 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| KRE: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (2106 stretches) | 0.021406667288 | 0.021406667288 | 1e-12 | pass |
| KRE: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| KRE: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| KRE: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 16 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| KRE: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| KRE: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| KRE: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 15 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| KRE: the entries followed never grow in number from one session to the next, and all 15 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| KRE: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (1728 stretches) | 0.027402754921 | 0.027402754921 | 1e-12 | pass |
| KRE: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| KRE: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| KIE, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| KIE, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| KIE, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| KIE, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| KIE: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| KIE: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| KIE: the 30-day risk reversal is read only on days carrying both grid points (days dropped: 0; none invented) | 0 | 0 | 0 | pass |
| KIE, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| KIE, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| KIE: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 28 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| KIE: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| KIE: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| KIE: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 27 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| KIE: the entries followed never grow in number from one session to the next, and all 27 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| KIE: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (1835 stretches) | 0.035766723109 | 0.035766723109 | 1e-12 | pass |
| KIE: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| KIE: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| KIE: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 28 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| KIE: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| KIE: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| KIE: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 27 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| KIE: the entries followed never grow in number from one session to the next, and all 27 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| KIE: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (1985 stretches) | 0.036515233032 | 0.036515233032 | 1e-12 | pass |
| KIE: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| KIE: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| XBI, 30 days: the price distribution's mass (target 1.0, tolerance 0.001) | 1 | 0.999999998148 | 0.001 | pass |
| XBI, 30 days: the price distribution's no negative (target 0.0, tolerance 0.005) | 0 | -4.356e-9 | 0.005 | pass |
| XBI, 30 days: the price distribution's mean forward (target 0.0, tolerance 0.001) | 0 | -1.30397e-7 | 0.001 | pass |
| XBI, 30 days: the price distribution's atm iv (target 0.0, tolerance 0.005) | 0 | -0.003358869012 | 0.005 | pass |
| XBI, 91 days: the price distribution's mass (target 1.0, tolerance 0.001) | 1 | 0.999999787666 | 0.001 | pass |
| XBI, 91 days: the price distribution's no negative (target 0.0, tolerance 0.005) | 0 | 0 | 0.005 | pass |
| XBI, 91 days: the price distribution's mean forward (target 0.0, tolerance 0.001) | 0 | 1.82388e-7 | 0.001 | pass |
| XBI, 91 days: the price distribution's atm iv (target 0.0, tolerance 0.005) | 0 | -0.004140518123 | 0.005 | pass |
| XBI, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| XBI, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| XBI, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| XBI, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| XBI: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| XBI: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| XBI: the 30-day risk reversal is read only on days carrying both grid points (days dropped: 0; none invented) | 0 | 0 | 0 | pass |
| XBI, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| XBI, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| XBI: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 7 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| XBI: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| XBI: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| XBI: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 6 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| XBI: the entries followed never grow in number from one session to the next, and all 6 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| XBI: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (1532 stretches) | -0.010436862176 | -0.010436862176 | 1e-12 | pass |
| XBI: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| XBI: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| XBI: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 24 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| XBI: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| XBI: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| XBI: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 23 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| XBI: the entries followed never grow in number from one session to the next, and all 23 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| XBI: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (2111 stretches) | 0.008134059318 | 0.008134059318 | 1e-12 | pass |
| XBI: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| XBI: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| XHB, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| XHB, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| XHB, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| XHB, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| XHB: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| XHB: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| XHB: the 30-day risk reversal is read only on days carrying both grid points (days dropped: 0; none invented) | 0 | 0 | 0 | pass |
| XHB, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| XHB, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| XHB: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 18 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| XHB: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| XHB: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| XHB: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 17 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| XHB: the entries followed never grow in number from one session to the next, and all 17 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| XHB: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (1880 stretches) | 0.051397486538 | 0.051397486538 | 1e-12 | pass |
| XHB: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| XHB: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| XHB: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 18 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| XHB: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| XHB: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| XHB: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 17 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| XHB: the entries followed never grow in number from one session to the next, and all 17 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| XHB: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (1893 stretches) | 0.052143087735 | 0.052143087735 | 1e-12 | pass |
| XHB: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| XHB: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| XME, 91 days: the price distribution's mass (target 1.0, tolerance 0.001) | 1 | 0.999995554248 | 0.001 | pass |
| XME, 91 days: the price distribution's no negative (target 0.0, tolerance 0.005) | 0 | 0 | 0.005 | pass |
| XME, 91 days: the price distribution's mean forward (target 0.0, tolerance 0.001) | 0 | -0.000420388056 | 0.001 | pass |
| XME, 91 days: the price distribution's atm iv (target 0.0, tolerance 0.005) | 0 | 0.004939408781 | 0.005 | pass |
| XME, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| XME, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| XME, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| XME, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| XME: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| XME: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| XME: the 30-day risk reversal is read only on days carrying both grid points (days dropped: 0; none invented) | 0 | 0 | 0 | pass |
| XME, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| XME, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| XME: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 24 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| XME: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| XME: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| XME: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 23 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| XME: the entries followed never grow in number from one session to the next, and all 23 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| XME: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (2093 stretches) | 0.039186319391 | 0.039186319391 | 1e-12 | pass |
| XME: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| XME: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| XME: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 24 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| XME: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| XME: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| XME: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 23 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| XME: the entries followed never grow in number from one session to the next, and all 23 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| XME: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (2012 stretches) | 0.042135083842 | 0.042135083842 | 1e-12 | pass |
| XME: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| XME: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| XOP, 30 days: the price distribution's mass (target 1.0, tolerance 0.001) | 1 | 0.999999973093 | 0.001 | pass |
| XOP, 30 days: the price distribution's no negative (target 0.0, tolerance 0.005) | 0 | 0 | 0.005 | pass |
| XOP, 30 days: the price distribution's mean forward (target 0.0, tolerance 0.001) | 0 | 2.6402e-8 | 0.001 | pass |
| XOP, 30 days: the price distribution's atm iv (target 0.0, tolerance 0.005) | 0 | 0.000643235891 | 0.005 | pass |
| XOP, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| XOP, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| XOP, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| XOP, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| XOP: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| XOP: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| XOP: the 30-day risk reversal is read only on days carrying both grid points (days dropped: 0; none invented) | 0 | 0 | 0 | pass |
| XOP, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| XOP, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| XOP: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 26 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| XOP: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| XOP: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| XOP: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 25 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| XOP: the entries followed never grow in number from one session to the next, and all 25 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| XOP: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (2142 stretches) | 0.019656216339 | 0.019656216339 | 1e-12 | pass |
| XOP: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| XOP: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| XOP: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 26 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| XOP: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| XOP: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| XOP: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 25 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| XOP: the entries followed never grow in number from one session to the next, and all 25 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| XOP: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (2103 stretches) | 0.01663557831 | 0.01663557831 | 1e-12 | pass |
| XOP: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| XOP: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| XRT, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| XRT, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| XRT, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| XRT, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| XRT: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| XRT: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| XRT: the 30-day risk reversal is read only on days carrying both grid points (days dropped: 0; none invented) | 0 | 0 | 0 | pass |
| XRT, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| XRT, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| XRT: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 24 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| XRT: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| XRT: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| XRT: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 23 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| XRT: the entries followed never grow in number from one session to the next, and all 23 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| XRT: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (2112 stretches) | 0.024108837908 | 0.024108837908 | 1e-12 | pass |
| XRT: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| XRT: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| XRT: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 25 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| XRT: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| XRT: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| XRT: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 24 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| XRT: the entries followed never grow in number from one session to the next, and all 24 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| XRT: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (2115 stretches) | 0.023969805221 | 0.023969805221 | 1e-12 | pass |
| XRT: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| XRT: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| ITA, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| ITA, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| ITA, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| ITA, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| ITA: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| ITA: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| ITA: the 30-day risk reversal is read only on days carrying both grid points (days dropped: 0; none invented) | 0 | 0 | 0 | pass |
| ITA, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| ITA, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| ITA: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 14 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| ITA: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| ITA: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| ITA: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 13 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| ITA: the entries followed never grow in number from one session to the next, and all 13 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| ITA: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (1211 stretches) | 0.046284264195 | 0.046284264195 | 1e-12 | pass |
| ITA: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| ITA: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| ITA: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 13 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| ITA: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| ITA: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| ITA: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 12 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| ITA: the entries followed never grow in number from one session to the next, and all 12 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| ITA: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (1072 stretches) | 0.0506833336 | 0.0506833336 | 1e-12 | pass |
| ITA: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| ITA: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| DTCR, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| DTCR, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| DTCR, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| DTCR, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| DTCR: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| DTCR: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| DTCR: the 30-day risk reversal is read only on days carrying both grid points (days dropped: 0; none invented) | 0 | 0 | 0 | pass |
| DTCR, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| DTCR, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| DTCR: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 1 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| DTCR: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| DTCR: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 1 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| DTCR: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| URNM, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| URNM, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| URNM, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| URNM, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| URNM: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| URNM: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| URNM: the 30-day risk reversal is read only on days carrying both grid points (days dropped: 0; none invented) | 0 | 0 | 0 | pass |
| URNM, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| URNM, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| URNM: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 8 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| URNM: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| URNM: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| URNM: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 7 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| URNM: the entries followed never grow in number from one session to the next, and all 7 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| URNM: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (595 stretches) | -0.018590079872 | -0.018590079872 | 1e-12 | pass |
| URNM: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| URNM: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| URNM: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 7 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| URNM: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| URNM: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| URNM: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 6 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| URNM: the entries followed never grow in number from one session to the next, and all 6 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| URNM: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (497 stretches) | 0.05814003431 | 0.05814003431 | 1e-12 | pass |
| URNM: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| URNM: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| EWY, 30 days: the price distribution's mass (target 1.0, tolerance 0.001) | 1 | 0.999999933849 | 0.001 | pass |
| EWY, 30 days: the price distribution's no negative (target 0.0, tolerance 0.005) | 0 | 0 | 0.005 | pass |
| EWY, 30 days: the price distribution's mean forward (target 0.0, tolerance 0.001) | 0 | -0.000001947464 | 0.001 | pass |
| EWY, 30 days: the price distribution's atm iv (target 0.0, tolerance 0.005) | 0 | 0.004012290095 | 0.005 | pass |
| EWY, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| EWY, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| EWY, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| EWY, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| EWY: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| EWY: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| EWY: the 30-day risk reversal is read only on days carrying both grid points (days dropped: 0; none invented) | 0 | 0 | 0 | pass |
| EWY, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| EWY, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| EWY: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 18 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| EWY: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| EWY: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| EWY: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 17 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| EWY: the entries followed never grow in number from one session to the next, and all 17 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| EWY: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (2012 stretches) | 0.012256206682 | 0.012256206682 | 1e-12 | pass |
| EWY: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| EWY: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| EWY: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 6 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| EWY: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| EWY: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| EWY: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 5 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| EWY: the entries followed never grow in number from one session to the next, and all 5 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| EWY: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (925 stretches) | -0.036457606583 | -0.036457606583 | 1e-12 | pass |
| EWY: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| EWY: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| FXI, 30 days: the price distribution's mass (target 1.0, tolerance 0.001) | 1 | 0.99998714644 | 0.001 | pass |
| FXI, 30 days: the price distribution's no negative (target 0.0, tolerance 0.005) | 0 | 0 | 0.005 | pass |
| FXI, 30 days: the price distribution's mean forward (target 0.0, tolerance 0.001) | 0 | 0.00000653288 | 0.001 | pass |
| FXI, 30 days: the price distribution's atm iv (target 0.0, tolerance 0.005) | 0 | -0.004377403082 | 0.005 | pass |
| FXI, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| FXI, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| FXI, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| FXI, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| FXI: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| FXI: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| FXI: the 30-day risk reversal is read only on days carrying both grid points (days dropped: 0; none invented) | 0 | 0 | 0 | pass |
| FXI, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| FXI, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| FXI: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 22 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| FXI: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| FXI: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| FXI: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 21 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| FXI: the entries followed never grow in number from one session to the next, and all 21 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| FXI: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (2138 stretches) | -0.015875244105 | -0.015875244105 | 1e-12 | pass |
| FXI: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| FXI: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| FXI: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 11 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| FXI: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| FXI: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| FXI: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 10 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| FXI: the entries followed never grow in number from one session to the next, and all 10 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| FXI: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (1676 stretches) | -0.003943919606 | -0.003943919606 | 1e-12 | pass |
| FXI: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| FXI: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| YINN, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| YINN, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| YINN, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| YINN, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| YINN: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| YINN: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| YINN: the 30-day risk reversal is read only on days carrying both grid points (days dropped: 0; none invented) | 0 | 0 | 0 | pass |
| YINN, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| YINN, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| YINN: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 22 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| YINN: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| YINN: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| YINN: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 21 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| YINN: the entries followed never grow in number from one session to the next, and all 21 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| YINN: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (2140 stretches) | -0.114173064968 | -0.114173064968 | 1e-12 | pass |
| YINN: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| YINN: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| YINN: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 18 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| YINN: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| YINN: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| YINN: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 17 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| YINN: the entries followed never grow in number from one session to the next, and all 17 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| YINN: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (1827 stretches) | -0.101199617476 | -0.101199617476 | 1e-12 | pass |
| YINN: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| YINN: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| KORU, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| KORU, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| KORU, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| KORU, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| KORU: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| KORU: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| KORU: the 30-day risk reversal is read only on days carrying both grid points (days dropped: 0; none invented) | 0 | 0 | 0 | pass |
| KORU, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| KORU, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| KORU: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 10 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| KORU: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| KORU: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| KORU: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 9 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| KORU: the entries followed never grow in number from one session to the next, and all 9 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| KORU: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (603 stretches) | 0.123878632008 | 0.123878632008 | 1e-12 | pass |
| KORU: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| KORU: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| KORU: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 7 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| KORU: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| KORU: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| KORU: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 6 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| KORU: the entries followed never grow in number from one session to the next, and all 6 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| KORU: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (311 stretches) | -0.087208689233 | -0.087208689233 | 1e-12 | pass |
| KORU: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| KORU: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| SOXL, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| SOXL, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| SOXL, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| SOXL, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| SOXL: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| SOXL: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| SOXL: the 30-day risk reversal is read only on days carrying both grid points (days dropped: 0; none invented) | 0 | 0 | 0 | pass |
| SOXL, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| SOXL, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| SOXL: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 15 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| SOXL: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| SOXL: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| SOXL: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 14 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| SOXL: the entries followed never grow in number from one session to the next, and all 14 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| SOXL: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (2131 stretches) | 0.127280623375 | 0.127280623375 | 1e-12 | pass |
| SOXL: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| SOXL: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| SOXL: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 22 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| SOXL: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| SOXL: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| SOXL: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 21 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| SOXL: the entries followed never grow in number from one session to the next, and all 21 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| SOXL: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (2133 stretches) | 0.127112806271 | 0.127112806271 | 1e-12 | pass |
| SOXL: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| SOXL: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| SPXL, 91 days: the price distribution's mass (target 1.0, tolerance 0.001) | 1 | 0.999850442455 | 0.001 | pass |
| SPXL, 91 days: the price distribution's no negative (target 0.0, tolerance 0.005) | 0 | 0 | 0.005 | pass |
| SPXL, 91 days: the price distribution's mean forward (target 0.0, tolerance 0.001) | 0 | 0.000147829992 | 0.001 | pass |
| SPXL, 91 days: the price distribution's atm iv (target 0.0, tolerance 0.005) | 0 | 0.002838296105 | 0.005 | pass |
| SPXL, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| SPXL, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| SPXL, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| SPXL, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| SPXL: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| SPXL: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| SPXL: the 30-day risk reversal is read only on days carrying both grid points (days dropped: 0; none invented) | 0 | 0 | 0 | pass |
| SPXL, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| SPXL, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| SPXL: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 8 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| SPXL: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| SPXL: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| SPXL: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 7 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| SPXL: the entries followed never grow in number from one session to the next, and all 7 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| SPXL: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (1632 stretches) | 0.126726648558 | 0.126726648558 | 1e-12 | pass |
| SPXL: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| SPXL: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| SPXL: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 24 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| SPXL: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| SPXL: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| SPXL: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 23 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| SPXL: the entries followed never grow in number from one session to the next, and all 23 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| SPXL: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (2031 stretches) | 0.112921374202 | 0.112921374202 | 1e-12 | pass |
| SPXL: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| SPXL: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| TQQQ, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| TQQQ, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| TQQQ, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| TQQQ, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| TQQQ: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| TQQQ: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| TQQQ: the 30-day risk reversal is read only on days carrying both grid points (days dropped: 0; none invented) | 0 | 0 | 0 | pass |
| TQQQ, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| TQQQ, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| TQQQ: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 25 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| TQQQ: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| TQQQ: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| TQQQ: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 24 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| TQQQ: the entries followed never grow in number from one session to the next, and all 24 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| TQQQ: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (2053 stretches) | 0.124454930869 | 0.124454930869 | 1e-12 | pass |
| TQQQ: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| TQQQ: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| TQQQ: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 29 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| TQQQ: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| TQQQ: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| TQQQ: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 28 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| TQQQ: the entries followed never grow in number from one session to the next, and all 28 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| TQQQ: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (2124 stretches) | 0.118281697726 | 0.118281697726 | 1e-12 | pass |
| TQQQ: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| TQQQ: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| SQQQ, 30 days: the price distribution's mass (target 1.0, tolerance 0.001) | 1 | 0.999769131532 | 0.001 | pass |
| SQQQ, 30 days: the price distribution's no negative (target 0.0, tolerance 0.005) | 0 | -0.000860600282 | 0.005 | pass |
| SQQQ, 30 days: the price distribution's mean forward (target 0.0, tolerance 0.001) | 0 | 0.000222220002 | 0.001 | pass |
| SQQQ, 30 days: the price distribution's atm iv (target 0.0, tolerance 0.005) | 0 | -0.003263682566 | 0.005 | pass |
| SQQQ, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| SQQQ, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| SQQQ, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| SQQQ, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| SQQQ: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| SQQQ: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| SQQQ: the 30-day risk reversal is read only on days carrying both grid points (days dropped: 0; none invented) | 0 | 0 | 0 | pass |
| SQQQ, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| SQQQ, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| SQQQ: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 23 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| SQQQ: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| SQQQ: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| SQQQ: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 22 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| SQQQ: the entries followed never grow in number from one session to the next, and all 22 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| SQQQ: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (2142 stretches) | -0.191506253575 | -0.191506253575 | 1e-12 | pass |
| SQQQ: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| SQQQ: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| SQQQ: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 34 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| SQQQ: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| SQQQ: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| SQQQ: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 33 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| SQQQ: the entries followed never grow in number from one session to the next, and all 33 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| SQQQ: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (2141 stretches) | -0.191490569724 | -0.191490569724 | 1e-12 | pass |
| SQQQ: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| SQQQ: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| PSQ, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| PSQ, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| PSQ, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| PSQ, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| PSQ: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| PSQ: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| PSQ: the 30-day risk reversal is read only on days carrying both grid points (days dropped: 0; none invented) | 0 | 0 | 0 | pass |
| PSQ, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| PSQ, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| PSQ: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 28 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| PSQ: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| PSQ: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| PSQ: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 27 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| PSQ: the entries followed never grow in number from one session to the next, and all 27 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| PSQ: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (2141 stretches) | -0.054123929375 | -0.054123929375 | 1e-12 | pass |
| PSQ: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| PSQ: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| PSQ: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 35 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| PSQ: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| PSQ: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| PSQ: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 34 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| PSQ: the entries followed never grow in number from one session to the next, and all 34 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| PSQ: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (2140 stretches) | -0.054060323825 | -0.054060323825 | 1e-12 | pass |
| PSQ: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| PSQ: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| UVXY, 30 days: the price distribution's mass (target 1.0, tolerance 0.001) | 1 | 0.99999374438 | 0.001 | pass |
| UVXY, 30 days: the price distribution's no negative (target 0.0, tolerance 0.005) | 0 | 0 | 0.005 | pass |
| UVXY, 30 days: the price distribution's mean forward (target 0.0, tolerance 0.001) | 0 | -0.000577194299 | 0.001 | pass |
| UVXY, 30 days: the price distribution's atm iv (target 0.0, tolerance 0.005) | 0 | -0.003951732883 | 0.005 | pass |
| UVXY, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| UVXY, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| UVXY, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| UVXY, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| UVXY: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| UVXY: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| UVXY: the 30-day risk reversal is read only on days carrying both grid points (days dropped: 1; none invented) | 0 | 0 | 0 | pass |
| UVXY, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| UVXY, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| UVXY: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 22 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| UVXY: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| UVXY: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| UVXY: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 21 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| UVXY: the entries followed never grow in number from one session to the next, and all 21 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| UVXY: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (2056 stretches) | -0.290382228404 | -0.290382228404 | 1e-12 | pass |
| UVXY: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| UVXY: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| UVXY: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 29 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| UVXY: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| UVXY: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| UVXY: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 28 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| UVXY: the entries followed never grow in number from one session to the next, and all 28 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| UVXY: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (2036 stretches) | -0.285714285714 | -0.285714285714 | 1e-12 | pass |
| UVXY: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| UVXY: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| TLT, 30 days: the price distribution's mass (target 1.0, tolerance 0.001) | 1 | 0.99999999496 | 0.001 | pass |
| TLT, 30 days: the price distribution's no negative (target 0.0, tolerance 0.005) | 0 | -1.81e-10 | 0.005 | pass |
| TLT, 30 days: the price distribution's mean forward (target 0.0, tolerance 0.001) | 0 | -3.7564e-8 | 0.001 | pass |
| TLT, 30 days: the price distribution's atm iv (target 0.0, tolerance 0.005) | 0 | -0.003013182752 | 0.005 | pass |
| TLT, 91 days: the price distribution's mass (target 1.0, tolerance 0.001) | 1 | 0.999999864705 | 0.001 | pass |
| TLT, 91 days: the price distribution's no negative (target 0.0, tolerance 0.005) | 0 | -0.000038357891 | 0.005 | pass |
| TLT, 91 days: the price distribution's mean forward (target 0.0, tolerance 0.001) | 0 | 1.32991e-7 | 0.001 | pass |
| TLT, 91 days: the price distribution's atm iv (target 0.0, tolerance 0.005) | 0 | -0.002274509525 | 0.005 | pass |
| TLT, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| TLT, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| TLT, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| TLT, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| TLT: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| TLT: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| TLT: the 30-day risk reversal is read only on days carrying both grid points (days dropped: 0; none invented) | 0 | 0 | 0 | pass |
| TLT, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| TLT, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| TLT: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 10 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| TLT: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| TLT: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| TLT: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 9 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| TLT: the entries followed never grow in number from one session to the next, and all 9 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| TLT: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (2117 stretches) | -0.004013723816 | -0.004013723816 | 1e-12 | pass |
| TLT: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| TLT: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| TLT: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 31 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| TLT: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| TLT: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| TLT: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 30 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| TLT: the entries followed never grow in number from one session to the next, and all 30 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| TLT: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (2135 stretches) | -0.003214948666 | -0.003214948666 | 1e-12 | pass |
| TLT: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| TLT: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| IEF, 30 days: the price distribution's mass (target 1.0, tolerance 0.001) | 1 | 0.999992973834 | 0.001 | pass |
| IEF, 30 days: the price distribution's no negative (target 0.0, tolerance 0.005) | 0 | -0.000030829764 | 0.005 | pass |
| IEF, 30 days: the price distribution's mean forward (target 0.0, tolerance 0.001) | 0 | -2.27733e-7 | 0.001 | pass |
| IEF, 30 days: the price distribution's atm iv (target 0.0, tolerance 0.005) | 0 | -0.004703590579 | 0.005 | pass |
| IEF, 91 days: the price distribution's mass (target 1.0, tolerance 0.001) | 1 | 0.999999992776 | 0.001 | pass |
| IEF, 91 days: the price distribution's no negative (target 0.0, tolerance 0.005) | 0 | 0 | 0.005 | pass |
| IEF, 91 days: the price distribution's mean forward (target 0.0, tolerance 0.001) | 0 | 5.258e-9 | 0.001 | pass |
| IEF, 91 days: the price distribution's atm iv (target 0.0, tolerance 0.005) | 0 | -0.001393470073 | 0.005 | pass |
| IEF, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| IEF, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| IEF, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| IEF, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| IEF: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| IEF: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| IEF: the 30-day risk reversal is read only on days carrying both grid points (days dropped: 0; none invented) | 0 | 0 | 0 | pass |
| IEF, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| IEF, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| IEF: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 32 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| IEF: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| IEF: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| IEF: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 31 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| IEF: the entries followed never grow in number from one session to the next, and all 31 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| IEF: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (2092 stretches) | 0.003291937871 | 0.003291937871 | 1e-12 | pass |
| IEF: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| IEF: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| IEF: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 22 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| IEF: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| IEF: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| IEF: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 21 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| IEF: the entries followed never grow in number from one session to the next, and all 21 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| IEF: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (1300 stretches) | -0.002873856054 | -0.002873856054 | 1e-12 | pass |
| IEF: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| IEF: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| TIP, 91 days: the price distribution's mass (target 1.0, tolerance 0.001) | 1 | 0.999999992067 | 0.001 | pass |
| TIP, 91 days: the price distribution's no negative (target 0.0, tolerance 0.005) | 0 | 0 | 0.005 | pass |
| TIP, 91 days: the price distribution's mean forward (target 0.0, tolerance 0.001) | 0 | 2.48e-10 | 0.001 | pass |
| TIP, 91 days: the price distribution's atm iv (target 0.0, tolerance 0.005) | 0 | 0.000935738537 | 0.005 | pass |
| TIP, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| TIP, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| TIP, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| TIP, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| TIP: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| TIP: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| TIP: the 30-day risk reversal is read only on days carrying both grid points (days dropped: 0; none invented) | 0 | 0 | 0 | pass |
| TIP, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| TIP, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| TIP: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 17 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| TIP: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| TIP: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| TIP: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 16 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| TIP: the entries followed never grow in number from one session to the next, and all 16 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| TIP: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (2064 stretches) | 0.010228553452 | 0.010228553452 | 1e-12 | pass |
| TIP: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| TIP: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| TIP: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 26 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| TIP: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| TIP: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| TIP: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 25 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| TIP: the entries followed never grow in number from one session to the next, and all 25 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| TIP: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (2155 stretches) | 0.009358843597 | 0.009358843597 | 1e-12 | pass |
| TIP: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| TIP: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| HYG, 91 days: the price distribution's mass (target 1.0, tolerance 0.001) | 1 | 0.999999994843 | 0.001 | pass |
| HYG, 91 days: the price distribution's no negative (target 0.0, tolerance 0.005) | 0 | 0 | 0.005 | pass |
| HYG, 91 days: the price distribution's mean forward (target 0.0, tolerance 0.001) | 0 | -7.22e-10 | 0.001 | pass |
| HYG, 91 days: the price distribution's atm iv (target 0.0, tolerance 0.005) | 0 | -0.000754423919 | 0.005 | pass |
| HYG, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| HYG, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| HYG, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| HYG, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| HYG: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| HYG: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| HYG: the 30-day risk reversal is read only on days carrying both grid points (days dropped: 0; none invented) | 0 | 0 | 0 | pass |
| HYG, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| HYG, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| HYG: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 7 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| HYG: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| HYG: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| HYG: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 6 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| HYG: the entries followed never grow in number from one session to the next, and all 6 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| HYG: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (1217 stretches) | 0.01265023714 | 0.01265023714 | 1e-12 | pass |
| HYG: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| HYG: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| HYG: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 12 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| HYG: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| HYG: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| HYG: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 11 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| HYG: the entries followed never grow in number from one session to the next, and all 11 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| HYG: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (1546 stretches) | 0.014991650052 | 0.014991650052 | 1e-12 | pass |
| HYG: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| HYG: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| HYGH, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| HYGH, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| HYGH, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| HYGH, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| HYGH: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| HYGH: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| HYGH: the 30-day risk reversal is read only on days carrying both grid points (days dropped: 2; none invented) | 0 | 0 | 0 | pass |
| HYGH, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| HYGH, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| HYGH: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 17 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| HYGH: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| HYGH: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| HYGH: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 16 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| HYGH: the entries followed never grow in number from one session to the next, and all 16 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| HYGH: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (1295 stretches) | 0.018515263584 | 0.018515263584 | 1e-12 | pass |
| HYGH: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| HYGH: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| HYGH: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 15 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| HYGH: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| HYGH: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| HYGH: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 14 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| HYGH: the entries followed never grow in number from one session to the next, and all 14 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| HYGH: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (1111 stretches) | 0.020706575431 | 0.020706575431 | 1e-12 | pass |
| HYGH: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| HYGH: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| LQD, 91 days: the price distribution's mass (target 1.0, tolerance 0.001) | 1 | 0.999999968318 | 0.001 | pass |
| LQD, 91 days: the price distribution's no negative (target 0.0, tolerance 0.005) | 0 | 0 | 0.005 | pass |
| LQD, 91 days: the price distribution's mean forward (target 0.0, tolerance 0.001) | 0 | 2.6457e-8 | 0.001 | pass |
| LQD, 91 days: the price distribution's atm iv (target 0.0, tolerance 0.005) | 0 | 0.001099616791 | 0.005 | pass |
| LQD, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| LQD, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| LQD, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| LQD, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| LQD: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| LQD: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| LQD: the 30-day risk reversal is read only on days carrying both grid points (days dropped: 0; none invented) | 0 | 0 | 0 | pass |
| LQD, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| LQD, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| LQD: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 10 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| LQD: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| LQD: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| LQD: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 9 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| LQD: the entries followed never grow in number from one session to the next, and all 9 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| LQD: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (1574 stretches) | 0.001025951948 | 0.001025951948 | 1e-12 | pass |
| LQD: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| LQD: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| LQD: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 18 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| LQD: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| LQD: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| LQD: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 17 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| LQD: the entries followed never grow in number from one session to the next, and all 17 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| LQD: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (1648 stretches) | 0.001856263112 | 0.001856263112 | 1e-12 | pass |
| LQD: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| LQD: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| LQDH, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| LQDH, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| LQDH, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| LQDH, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| LQDH: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| LQDH: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| LQDH: the 30-day risk reversal is read only on days carrying both grid points (days dropped: 2; none invented) | 0 | 0 | 0 | pass |
| LQDH, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| LQDH, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| LQDH: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 9 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| LQDH: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| LQDH: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| LQDH: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 8 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| LQDH: the entries followed never grow in number from one session to the next, and all 8 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| LQDH: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (1080 stretches) | 0.016664560391 | 0.016664560391 | 1e-12 | pass |
| LQDH: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| LQDH: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| LQDH: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 13 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| LQDH: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| LQDH: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| LQDH: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 12 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| LQDH: the entries followed never grow in number from one session to the next, and all 12 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| LQDH: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (1266 stretches) | 0.015258029731 | 0.015258029731 | 1e-12 | pass |
| LQDH: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| LQDH: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| GLD, 30 days: the price distribution's mass (target 1.0, tolerance 0.001) | 1 | 0.99999999801 | 0.001 | pass |
| GLD, 30 days: the price distribution's no negative (target 0.0, tolerance 0.005) | 0 | -9.66e-10 | 0.005 | pass |
| GLD, 30 days: the price distribution's mean forward (target 0.0, tolerance 0.001) | 0 | -6.6122e-8 | 0.001 | pass |
| GLD, 30 days: the price distribution's atm iv (target 0.0, tolerance 0.005) | 0 | 0.001217991359 | 0.005 | pass |
| GLD, 91 days: the price distribution's mass (target 1.0, tolerance 0.001) | 1 | 0.999979958456 | 0.001 | pass |
| GLD, 91 days: the price distribution's no negative (target 0.0, tolerance 0.005) | 0 | -0.000003887169 | 0.005 | pass |
| GLD, 91 days: the price distribution's mean forward (target 0.0, tolerance 0.001) | 0 | -0.000265557353 | 0.001 | pass |
| GLD, 91 days: the price distribution's atm iv (target 0.0, tolerance 0.005) | 0 | -0.001126880874 | 0.005 | pass |
| GLD, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| GLD, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| GLD, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| GLD, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| GLD: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| GLD: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| GLD: the 30-day risk reversal is read only on days carrying both grid points (days dropped: 0; none invented) | 0 | 0 | 0 | pass |
| GLD, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| GLD, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| GLD: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 19 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| GLD: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| GLD: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| GLD: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 18 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| GLD: the entries followed never grow in number from one session to the next, and all 18 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| GLD: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (1935 stretches) | 0.045097921782 | 0.045097921782 | 1e-12 | pass |
| GLD: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| GLD: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| GLD: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 25 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| GLD: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| GLD: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| GLD: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 24 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| GLD: the entries followed never grow in number from one session to the next, and all 24 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| GLD: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (2085 stretches) | 0.038361994594 | 0.038361994594 | 1e-12 | pass |
| GLD: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| GLD: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| SLV, 30 days: the price distribution's mass (target 1.0, tolerance 0.001) | 1 | 0.999999628275 | 0.001 | pass |
| SLV, 30 days: the price distribution's no negative (target 0.0, tolerance 0.005) | 0 | 0 | 0.005 | pass |
| SLV, 30 days: the price distribution's mean forward (target 0.0, tolerance 0.001) | 0 | -0.00003230077 | 0.001 | pass |
| SLV, 30 days: the price distribution's atm iv (target 0.0, tolerance 0.005) | 0 | -0.00343206097 | 0.005 | pass |
| SLV, 91 days: the price distribution's mass (target 1.0, tolerance 0.001) | 1 | 0.999993868774 | 0.001 | pass |
| SLV, 91 days: the price distribution's no negative (target 0.0, tolerance 0.005) | 0 | 0 | 0.005 | pass |
| SLV, 91 days: the price distribution's mean forward (target 0.0, tolerance 0.001) | 0 | -0.000062877631 | 0.001 | pass |
| SLV, 91 days: the price distribution's atm iv (target 0.0, tolerance 0.005) | 0 | -0.000978044948 | 0.005 | pass |
| SLV, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| SLV, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| SLV, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| SLV, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| SLV: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| SLV: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| SLV: the 30-day risk reversal is read only on days carrying both grid points (days dropped: 0; none invented) | 0 | 0 | 0 | pass |
| SLV, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| SLV, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| SLV: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 18 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| SLV: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| SLV: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| SLV: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 17 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| SLV: the entries followed never grow in number from one session to the next, and all 17 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| SLV: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (1672 stretches) | 0.022549958649 | 0.022549958649 | 1e-12 | pass |
| SLV: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| SLV: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| SLV: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 26 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| SLV: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| SLV: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| SLV: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 25 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| SLV: the entries followed never grow in number from one session to the next, and all 25 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| SLV: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (2025 stretches) | 0.024259721727 | 0.024259721727 | 1e-12 | pass |
| SLV: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| SLV: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| PALL, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| PALL, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| PALL: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| PALL: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| PALL: the 30-day risk reversal is read only on days carrying both grid points (days dropped: 0; none invented) | 0 | 0 | 0 | pass |
| PALL, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| PALL, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| PPLT, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| PPLT, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| PPLT: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| PPLT: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| PPLT: the 30-day risk reversal is read only on days carrying both grid points (days dropped: 0; none invented) | 0 | 0 | 0 | pass |
| PPLT, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| PPLT, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| CPER, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| CPER, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| CPER, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| CPER, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| CPER: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| CPER: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| CPER: the 30-day risk reversal is read only on days carrying both grid points (days dropped: 0; none invented) | 0 | 0 | 0 | pass |
| CPER, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| CPER, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| CPER: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 8 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| CPER: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| CPER: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| CPER: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 7 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| CPER: the entries followed never grow in number from one session to the next, and all 7 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| CPER: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (689 stretches) | 0.053715967623 | 0.053715967623 | 1e-12 | pass |
| CPER: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| CPER: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| CPER: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 14 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| CPER: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| CPER: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| CPER: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 13 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| CPER: the entries followed never grow in number from one session to the next, and all 13 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| CPER: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (718 stretches) | 0.048546536611 | 0.048546536611 | 1e-12 | pass |
| CPER: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| CPER: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| CORN, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| CORN, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| CORN, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| CORN, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| CORN: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| CORN: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| CORN: the 30-day risk reversal is read only on days carrying both grid points (days dropped: 0; none invented) | 0 | 0 | 0 | pass |
| CORN, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| CORN, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| CORN: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 26 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| CORN: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| CORN: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| CORN: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 25 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| CORN: the entries followed never grow in number from one session to the next, and all 25 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| CORN: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (1996 stretches) | -0.008798223986 | -0.008798223986 | 1e-12 | pass |
| CORN: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| CORN: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| CORN: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 27 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| CORN: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| CORN: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| CORN: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 26 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| CORN: the entries followed never grow in number from one session to the next, and all 26 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| CORN: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (2003 stretches) | -0.00873839432 | -0.00873839432 | 1e-12 | pass |
| CORN: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| CORN: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| UNG, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| UNG, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| UNG, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| UNG, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| UNG: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| UNG: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| UNG: the 30-day risk reversal is read only on days carrying both grid points (days dropped: 0; none invented) | 0 | 0 | 0 | pass |
| UNG, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| UNG, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| UNG: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 27 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| UNG: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| UNG: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| UNG: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 26 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| UNG: the entries followed never grow in number from one session to the next, and all 26 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| UNG: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (2161 stretches) | -0.055134041477 | -0.055134041477 | 1e-12 | pass |
| UNG: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| UNG: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| UNG: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 21 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| UNG: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| UNG: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| UNG: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 20 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| UNG: the entries followed never grow in number from one session to the next, and all 20 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| UNG: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (2066 stretches) | -0.05205922419 | -0.05205922419 | 1e-12 | pass |
| UNG: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| UNG: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| USO, 30 days: the price distribution's mass (target 1.0, tolerance 0.001) | 1 | 0.999999999661 | 0.001 | pass |
| USO, 30 days: the price distribution's no negative (target 0.0, tolerance 0.005) | 0 | -8.06e-10 | 0.005 | pass |
| USO, 30 days: the price distribution's mean forward (target 0.0, tolerance 0.001) | 0 | -4.908e-9 | 0.001 | pass |
| USO, 30 days: the price distribution's atm iv (target 0.0, tolerance 0.005) | 0 | -0.002231883718 | 0.005 | pass |
| USO, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| USO, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| USO, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| USO, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| USO: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| USO: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| USO: the 30-day risk reversal is read only on days carrying both grid points (days dropped: 0; none invented) | 0 | 0 | 0 | pass |
| USO, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| USO, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| USO: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 27 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| USO: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| USO: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| USO: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 26 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| USO: the entries followed never grow in number from one session to the next, and all 26 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| USO: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (1853 stretches) | 0.023882681564 | 0.023882681564 | 1e-12 | pass |
| USO: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| USO: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| USO: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 26 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| USO: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| USO: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| USO: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 25 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| USO: the entries followed never grow in number from one session to the next, and all 25 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| USO: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (1670 stretches) | 0.02999397687 | 0.02999397687 | 1e-12 | pass |
| USO: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| USO: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| UUP, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| UUP, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| UUP, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| UUP, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| UUP: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| UUP: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| UUP: the 30-day risk reversal is read only on days carrying both grid points (days dropped: 0; none invented) | 0 | 0 | 0 | pass |
| UUP, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| UUP, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| UUP: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 31 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| UUP: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| UUP: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| UUP: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 30 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| UUP: the entries followed never grow in number from one session to the next, and all 30 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| UUP: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (1840 stretches) | 0.012302761177 | 0.012302761177 | 1e-12 | pass |
| UUP: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| UUP: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| UUP: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 25 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| UUP: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| UUP: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| UUP: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 24 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| UUP: the entries followed never grow in number from one session to the next, and all 24 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| UUP: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (2065 stretches) | 0.01393692817 | 0.01393692817 | 1e-12 | pass |
| UUP: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| UUP: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| FXA, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| FXA, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| FXA, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| FXA, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| FXA: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| FXA: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| FXA: the 30-day risk reversal is read only on days carrying both grid points (days dropped: 0; none invented) | 0 | 0 | 0 | pass |
| FXA, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| FXA, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| FXA: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 21 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| FXA: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| FXA: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| FXA: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 20 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| FXA: the entries followed never grow in number from one session to the next, and all 20 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| FXA: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (1869 stretches) | 0.001035943014 | 0.001035943014 | 1e-12 | pass |
| FXA: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| FXA: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| FXA: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 32 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| FXA: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| FXA: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| FXA: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 31 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| FXA: the entries followed never grow in number from one session to the next, and all 31 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| FXA: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (2133 stretches) | -0.00157487715 | -0.00157487715 | 1e-12 | pass |
| FXA: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| FXA: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| FXB, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| FXB, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| FXB, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| FXB, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| FXB: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| FXB: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| FXB: the 30-day risk reversal is read only on days carrying both grid points (days dropped: 0; none invented) | 0 | 0 | 0 | pass |
| FXB, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| FXB, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| FXB: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 17 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| FXB: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| FXB: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| FXB: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 16 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| FXB: the entries followed never grow in number from one session to the next, and all 16 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| FXB: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (1957 stretches) | 0.006831592195 | 0.006831592195 | 1e-12 | pass |
| FXB: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| FXB: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| FXB: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 23 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| FXB: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| FXB: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| FXB: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 22 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| FXB: the entries followed never grow in number from one session to the next, and all 22 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| FXB: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (2085 stretches) | 0.005210494356 | 0.005210494356 | 1e-12 | pass |
| FXB: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| FXB: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| FXC, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| FXC, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| FXC, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| FXC, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| FXC: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| FXC: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| FXC: the 30-day risk reversal is read only on days carrying both grid points (days dropped: 0; none invented) | 0 | 0 | 0 | pass |
| FXC, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| FXC, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| FXC: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 30 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| FXC: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| FXC: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| FXC: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 29 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| FXC: the entries followed never grow in number from one session to the next, and all 29 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| FXC: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (2098 stretches) | -0.00088392301 | -0.00088392301 | 1e-12 | pass |
| FXC: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| FXC: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| FXC: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 33 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| FXC: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| FXC: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| FXC: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 32 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| FXC: the entries followed never grow in number from one session to the next, and all 32 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| FXC: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (2143 stretches) | -0.000526776346 | -0.000526776346 | 1e-12 | pass |
| FXC: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| FXC: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| FXE, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| FXE, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| FXE, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| FXE, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| FXE: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| FXE: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| FXE: the 30-day risk reversal is read only on days carrying both grid points (days dropped: 0; none invented) | 0 | 0 | 0 | pass |
| FXE, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| FXE, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| FXE: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 19 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| FXE: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| FXE: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| FXE: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 18 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| FXE: the entries followed never grow in number from one session to the next, and all 18 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| FXE: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (2041 stretches) | -0.00365639997 | -0.00365639997 | 1e-12 | pass |
| FXE: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| FXE: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| FXE: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 25 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| FXE: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| FXE: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| FXE: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 24 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| FXE: the entries followed never grow in number from one session to the next, and all 24 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| FXE: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (2094 stretches) | -0.003626252486 | -0.003626252486 | 1e-12 | pass |
| FXE: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| FXE: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| FXF, 30 days: the price distribution's mass (target 1.0, tolerance 0.001) | 1 | 0.999999996756 | 0.001 | pass |
| FXF, 30 days: the price distribution's no negative (target 0.0, tolerance 0.005) | 0 | -0.0006332818 | 0.005 | pass |
| FXF, 30 days: the price distribution's mean forward (target 0.0, tolerance 0.001) | 0 | -5.6237e-8 | 0.001 | pass |
| FXF, 30 days: the price distribution's atm iv (target 0.0, tolerance 0.005) | 0 | 0.004495171263 | 0.005 | pass |
| FXF, 91 days: the price distribution's mass (target 1.0, tolerance 0.001) | 1 | 0.99999997147 | 0.001 | pass |
| FXF, 91 days: the price distribution's no negative (target 0.0, tolerance 0.005) | 0 | 0 | 0.005 | pass |
| FXF, 91 days: the price distribution's mean forward (target 0.0, tolerance 0.001) | 0 | 8.756e-9 | 0.001 | pass |
| FXF, 91 days: the price distribution's atm iv (target 0.0, tolerance 0.005) | 0 | -0.003115756643 | 0.005 | pass |
| FXF, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| FXF, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| FXF, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| FXF, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| FXF: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| FXF: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| FXF: the 30-day risk reversal is read only on days carrying both grid points (days dropped: 0; none invented) | 0 | 0 | 0 | pass |
| FXF, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| FXF, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| FXF: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 26 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| FXF: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| FXF: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| FXF: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 25 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| FXF: the entries followed never grow in number from one session to the next, and all 25 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| FXF: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (2011 stretches) | 0.002664534219 | 0.002664534219 | 1e-12 | pass |
| FXF: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| FXF: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| FXF: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 34 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| FXF: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| FXF: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| FXF: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 33 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| FXF: the entries followed never grow in number from one session to the next, and all 33 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| FXF: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (2115 stretches) | 0.001290837528 | 0.001290837528 | 1e-12 | pass |
| FXF: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| FXF: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| FXY, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| FXY, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| FXY, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| FXY, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| FXY: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| FXY: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| FXY: the 30-day risk reversal is read only on days carrying both grid points (days dropped: 0; none invented) | 0 | 0 | 0 | pass |
| FXY, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| FXY, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| FXY: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 24 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| FXY: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| FXY: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| FXY: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 23 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| FXY: the entries followed never grow in number from one session to the next, and all 23 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| FXY: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (2145 stretches) | -0.014311676511 | -0.014311676511 | 1e-12 | pass |
| FXY: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| FXY: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| FXY: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 4 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| FXY: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| FXY: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| FXY: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 3 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| FXY: the entries followed never grow in number from one session to the next, and all 3 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| FXY: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (1383 stretches) | -0.02639508256 | -0.02639508256 | 1e-12 | pass |
| FXY: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| FXY: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| FX_AUDCAD, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| FX_AUDCAD, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| FX_AUDCAD, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| FX_AUDCAD, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| FX_AUDCAD: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| FX_AUDCAD: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| FX_AUDCAD, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| FX_AUDCAD, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| FX_AUDCAD: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 31 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| FX_AUDCAD: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| FX_AUDCAD: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| FX_AUDCAD: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 30 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| FX_AUDCAD: the entries followed never grow in number from one session to the next, and all 30 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| FX_AUDCAD: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (2131 stretches) | -0.002725574201 | -0.002725574201 | 1e-12 | pass |
| FX_AUDCAD: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| FX_AUDCAD: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| FX_AUDCAD: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 31 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| FX_AUDCAD: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| FX_AUDCAD: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| FX_AUDCAD: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 30 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| FX_AUDCAD: the entries followed never grow in number from one session to the next, and all 30 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| FX_AUDCAD: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (1908 stretches) | -0.006106231017 | -0.006106231017 | 1e-12 | pass |
| FX_AUDCAD: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| FX_AUDCAD: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| FX_AUDCHF, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| FX_AUDCHF, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| FX_AUDCHF, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| FX_AUDCHF, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| FX_AUDCHF: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| FX_AUDCHF: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| FX_AUDCHF, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| FX_AUDCHF, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| FX_AUDCHF: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 28 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| FX_AUDCHF: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| FX_AUDCHF: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| FX_AUDCHF: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 27 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| FX_AUDCHF: the entries followed never grow in number from one session to the next, and all 27 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| FX_AUDCHF: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (1835 stretches) | -0.012533670957 | -0.012533670957 | 1e-12 | pass |
| FX_AUDCHF: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| FX_AUDCHF: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| FX_AUDCHF: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 25 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| FX_AUDCHF: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| FX_AUDCHF: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| FX_AUDCHF: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 24 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| FX_AUDCHF: the entries followed never grow in number from one session to the next, and all 24 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| FX_AUDCHF: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (1919 stretches) | -0.010332107454 | -0.010332107454 | 1e-12 | pass |
| FX_AUDCHF: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| FX_AUDCHF: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| FX_AUDJPY, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| FX_AUDJPY, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| FX_AUDJPY, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| FX_AUDJPY, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| FX_AUDJPY: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| FX_AUDJPY: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| FX_AUDJPY, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| FX_AUDJPY, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| FX_AUDJPY: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 24 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| FX_AUDJPY: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| FX_AUDJPY: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| FX_AUDJPY: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 23 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| FX_AUDJPY: the entries followed never grow in number from one session to the next, and all 23 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| FX_AUDJPY: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (2104 stretches) | 0.010102662941 | 0.010102662941 | 1e-12 | pass |
| FX_AUDJPY: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| FX_AUDJPY: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| FX_AUDJPY: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 19 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| FX_AUDJPY: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| FX_AUDJPY: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| FX_AUDJPY: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 18 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| FX_AUDJPY: the entries followed never grow in number from one session to the next, and all 18 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| FX_AUDJPY: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (2112 stretches) | 0.009823470357 | 0.009823470357 | 1e-12 | pass |
| FX_AUDJPY: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| FX_AUDJPY: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| FX_AUDX, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| FX_AUDX, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| FX_AUDX, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| FX_AUDX, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| FX_AUDX: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| FX_AUDX: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| FX_AUDX, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| FX_AUDX, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| FX_AUDX: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 29 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| FX_AUDX: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| FX_AUDX: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| FX_AUDX: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 28 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| FX_AUDX: the entries followed never grow in number from one session to the next, and all 28 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| FX_AUDX: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (1960 stretches) | -0.003252702857 | -0.003252702857 | 1e-12 | pass |
| FX_AUDX: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| FX_AUDX: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| FX_AUDX: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 32 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| FX_AUDX: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| FX_AUDX: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| FX_AUDX: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 31 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| FX_AUDX: the entries followed never grow in number from one session to the next, and all 31 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| FX_AUDX: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (2131 stretches) | -0.000329225771 | -0.000329225771 | 1e-12 | pass |
| FX_AUDX: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| FX_AUDX: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| FX_CADCHF, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| FX_CADCHF, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| FX_CADCHF, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| FX_CADCHF, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| FX_CADCHF: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| FX_CADCHF: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| FX_CADCHF, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| FX_CADCHF, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| FX_CADCHF: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 25 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| FX_CADCHF: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| FX_CADCHF: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| FX_CADCHF: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 24 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| FX_CADCHF: the entries followed never grow in number from one session to the next, and all 24 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| FX_CADCHF: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (2063 stretches) | -0.003618245707 | -0.003618245707 | 1e-12 | pass |
| FX_CADCHF: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| FX_CADCHF: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| FX_CADCHF: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 32 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| FX_CADCHF: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| FX_CADCHF: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| FX_CADCHF: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 31 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| FX_CADCHF: the entries followed never grow in number from one session to the next, and all 31 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| FX_CADCHF: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (2056 stretches) | -0.003733376788 | -0.003733376788 | 1e-12 | pass |
| FX_CADCHF: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| FX_CADCHF: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| FX_CADJPY, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| FX_CADJPY, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| FX_CADJPY, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| FX_CADJPY, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| FX_CADJPY: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| FX_CADJPY: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| FX_CADJPY, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| FX_CADJPY, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| FX_CADJPY: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 22 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| FX_CADJPY: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| FX_CADJPY: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| FX_CADJPY: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 21 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| FX_CADJPY: the entries followed never grow in number from one session to the next, and all 21 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| FX_CADJPY: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (2068 stretches) | 0.016670156437 | 0.016670156437 | 1e-12 | pass |
| FX_CADJPY: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| FX_CADJPY: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| FX_CADJPY: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 19 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| FX_CADJPY: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| FX_CADJPY: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| FX_CADJPY: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 18 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| FX_CADJPY: the entries followed never grow in number from one session to the next, and all 18 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| FX_CADJPY: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (2131 stretches) | 0.015125721272 | 0.015125721272 | 1e-12 | pass |
| FX_CADJPY: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| FX_CADJPY: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| FX_CADX, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| FX_CADX, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| FX_CADX, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| FX_CADX, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| FX_CADX: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| FX_CADX: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| FX_CADX, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| FX_CADX, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| FX_CADX: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 30 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| FX_CADX: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| FX_CADX: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| FX_CADX: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 29 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| FX_CADX: the entries followed never grow in number from one session to the next, and all 29 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| FX_CADX: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (2124 stretches) | 0.001326218248 | 0.001326218248 | 1e-12 | pass |
| FX_CADX: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| FX_CADX: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| FX_CADX: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 40 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| FX_CADX: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| FX_CADX: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| FX_CADX: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 39 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| FX_CADX: the entries followed never grow in number from one session to the next, and all 39 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| FX_CADX: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (2131 stretches) | 0.001311147666 | 0.001311147666 | 1e-12 | pass |
| FX_CADX: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| FX_CADX: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| FX_CHFJPY, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| FX_CHFJPY, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| FX_CHFJPY, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| FX_CHFJPY, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| FX_CHFJPY: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| FX_CHFJPY: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| FX_CHFJPY, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| FX_CHFJPY, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| FX_CHFJPY: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 26 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| FX_CHFJPY: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| FX_CHFJPY: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| FX_CHFJPY: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 25 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| FX_CHFJPY: the entries followed never grow in number from one session to the next, and all 25 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| FX_CHFJPY: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (1711 stretches) | 0.012948460673 | 0.012948460673 | 1e-12 | pass |
| FX_CHFJPY: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| FX_CHFJPY: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| FX_CHFJPY: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 27 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| FX_CHFJPY: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| FX_CHFJPY: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| FX_CHFJPY: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 26 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| FX_CHFJPY: the entries followed never grow in number from one session to the next, and all 26 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| FX_CHFJPY: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (2080 stretches) | 0.01564486616 | 0.01564486616 | 1e-12 | pass |
| FX_CHFJPY: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| FX_CHFJPY: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| FX_CHFX, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| FX_CHFX, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| FX_CHFX, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| FX_CHFX, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| FX_CHFX: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| FX_CHFX: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| FX_CHFX, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| FX_CHFX, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| FX_CHFX: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 31 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| FX_CHFX: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| FX_CHFX: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| FX_CHFX: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 30 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| FX_CHFX: the entries followed never grow in number from one session to the next, and all 30 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| FX_CHFX: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (1842 stretches) | 0.007522851182 | 0.007522851182 | 1e-12 | pass |
| FX_CHFX: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| FX_CHFX: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| FX_CHFX: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 25 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| FX_CHFX: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| FX_CHFX: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| FX_CHFX: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 24 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| FX_CHFX: the entries followed never grow in number from one session to the next, and all 24 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| FX_CHFX: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (1998 stretches) | 0.006189165855 | 0.006189165855 | 1e-12 | pass |
| FX_CHFX: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| FX_CHFX: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| FX_EURAUD, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| FX_EURAUD, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| FX_EURAUD, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| FX_EURAUD, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| FX_EURAUD: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| FX_EURAUD: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| FX_EURAUD, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| FX_EURAUD, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| FX_EURAUD: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 29 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| FX_EURAUD: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| FX_EURAUD: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| FX_EURAUD: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 28 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| FX_EURAUD: the entries followed never grow in number from one session to the next, and all 28 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| FX_EURAUD: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (2017 stretches) | 0.001461523471 | 0.001461523471 | 1e-12 | pass |
| FX_EURAUD: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| FX_EURAUD: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| FX_EURAUD: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 29 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| FX_EURAUD: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| FX_EURAUD: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| FX_EURAUD: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 28 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| FX_EURAUD: the entries followed never grow in number from one session to the next, and all 28 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| FX_EURAUD: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (2131 stretches) | -0.000159025515 | -0.000159025515 | 1e-12 | pass |
| FX_EURAUD: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| FX_EURAUD: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| FX_EURCAD, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| FX_EURCAD, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| FX_EURCAD, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| FX_EURCAD, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| FX_EURCAD: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| FX_EURCAD: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| FX_EURCAD, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| FX_EURCAD, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| FX_EURCAD: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 30 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| FX_EURCAD: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| FX_EURCAD: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| FX_EURCAD: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 29 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| FX_EURCAD: the entries followed never grow in number from one session to the next, and all 29 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| FX_EURCAD: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (1647 stretches) | 0.001880250539 | 0.001880250539 | 1e-12 | pass |
| FX_EURCAD: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| FX_EURCAD: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| FX_EURCAD: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 18 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| FX_EURCAD: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| FX_EURCAD: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| FX_EURCAD: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 17 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| FX_EURCAD: the entries followed never grow in number from one session to the next, and all 17 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| FX_EURCAD: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (1638 stretches) | 0.001536616822 | 0.001536616822 | 1e-12 | pass |
| FX_EURCAD: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| FX_EURCAD: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| FX_EURCHF, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| FX_EURCHF, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| FX_EURCHF, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| FX_EURCHF, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| FX_EURCHF: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| FX_EURCHF: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| FX_EURCHF, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| FX_EURCHF, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| FX_EURCHF: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 29 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| FX_EURCHF: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| FX_EURCHF: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| FX_EURCHF: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 28 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| FX_EURCHF: the entries followed never grow in number from one session to the next, and all 28 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| FX_EURCHF: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (1917 stretches) | -0.005764549425 | -0.005764549425 | 1e-12 | pass |
| FX_EURCHF: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| FX_EURCHF: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| FX_EURCHF: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 17 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| FX_EURCHF: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| FX_EURCHF: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| FX_EURCHF: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 16 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| FX_EURCHF: the entries followed never grow in number from one session to the next, and all 16 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| FX_EURCHF: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (1788 stretches) | -0.003372019061 | -0.003372019061 | 1e-12 | pass |
| FX_EURCHF: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| FX_EURCHF: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| FX_EURGBP, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| FX_EURGBP, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| FX_EURGBP, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| FX_EURGBP, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| FX_EURGBP: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| FX_EURGBP: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| FX_EURGBP, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| FX_EURGBP, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| FX_EURGBP: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 15 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| FX_EURGBP: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| FX_EURGBP: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| FX_EURGBP: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 14 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| FX_EURGBP: the entries followed never grow in number from one session to the next, and all 14 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| FX_EURGBP: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (2036 stretches) | -0.004659834136 | -0.004659834136 | 1e-12 | pass |
| FX_EURGBP: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| FX_EURGBP: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| FX_EURGBP: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 23 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| FX_EURGBP: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| FX_EURGBP: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| FX_EURGBP: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 22 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| FX_EURGBP: the entries followed never grow in number from one session to the next, and all 22 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| FX_EURGBP: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (2130 stretches) | -0.005627731901 | -0.005627731901 | 1e-12 | pass |
| FX_EURGBP: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| FX_EURGBP: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| FX_EURJPY, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| FX_EURJPY, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| FX_EURJPY, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| FX_EURJPY, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| FX_EURJPY: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| FX_EURJPY: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| FX_EURJPY, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| FX_EURJPY, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| FX_EURJPY: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 33 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| FX_EURJPY: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| FX_EURJPY: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| FX_EURJPY: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 32 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| FX_EURJPY: the entries followed never grow in number from one session to the next, and all 32 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| FX_EURJPY: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (1909 stretches) | 0.014017012516 | 0.014017012516 | 1e-12 | pass |
| FX_EURJPY: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| FX_EURJPY: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| FX_EURJPY: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 28 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| FX_EURJPY: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| FX_EURJPY: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| FX_EURJPY: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 27 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| FX_EURJPY: the entries followed never grow in number from one session to the next, and all 27 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| FX_EURJPY: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (2131 stretches) | 0.01187056124 | 0.01187056124 | 1e-12 | pass |
| FX_EURJPY: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| FX_EURJPY: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| FX_EURX, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| FX_EURX, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| FX_EURX, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| FX_EURX, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| FX_EURX: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| FX_EURX: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| FX_EURX, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| FX_EURX, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| FX_EURX: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 27 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| FX_EURX: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| FX_EURX: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| FX_EURX: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 26 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| FX_EURX: the entries followed never grow in number from one session to the next, and all 26 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| FX_EURX: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (1909 stretches) | -0.001380112986 | -0.001380112986 | 1e-12 | pass |
| FX_EURX: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| FX_EURX: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| FX_EURX: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 32 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| FX_EURX: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| FX_EURX: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| FX_EURX: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 31 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| FX_EURX: the entries followed never grow in number from one session to the next, and all 31 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| FX_EURX: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (2083 stretches) | -0.001670265046 | -0.001670265046 | 1e-12 | pass |
| FX_EURX: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| FX_EURX: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| FX_GBPAUD, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| FX_GBPAUD, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| FX_GBPAUD, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| FX_GBPAUD, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| FX_GBPAUD: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| FX_GBPAUD: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| FX_GBPAUD, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| FX_GBPAUD, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| FX_GBPAUD: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 31 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| FX_GBPAUD: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| FX_GBPAUD: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| FX_GBPAUD: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 30 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| FX_GBPAUD: the entries followed never grow in number from one session to the next, and all 30 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| FX_GBPAUD: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (2121 stretches) | 0.004597482398 | 0.004597482398 | 1e-12 | pass |
| FX_GBPAUD: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| FX_GBPAUD: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| FX_GBPAUD: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 29 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| FX_GBPAUD: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| FX_GBPAUD: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| FX_GBPAUD: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 28 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| FX_GBPAUD: the entries followed never grow in number from one session to the next, and all 28 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| FX_GBPAUD: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (1911 stretches) | 0.00821673185 | 0.00821673185 | 1e-12 | pass |
| FX_GBPAUD: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| FX_GBPAUD: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| FX_GBPCAD, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| FX_GBPCAD, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| FX_GBPCAD, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| FX_GBPCAD, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| FX_GBPCAD: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| FX_GBPCAD: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| FX_GBPCAD, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| FX_GBPCAD, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| FX_GBPCAD: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 4 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| FX_GBPCAD: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| FX_GBPCAD: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| FX_GBPCAD: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 3 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| FX_GBPCAD: the entries followed never grow in number from one session to the next, and all 3 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| FX_GBPCAD: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (1364 stretches) | 0.005629314895 | 0.005629314895 | 1e-12 | pass |
| FX_GBPCAD: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| FX_GBPCAD: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| FX_GBPCAD: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 31 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| FX_GBPCAD: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| FX_GBPCAD: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| FX_GBPCAD: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 30 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| FX_GBPCAD: the entries followed never grow in number from one session to the next, and all 30 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| FX_GBPCAD: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (2110 stretches) | 0.006660479561 | 0.006660479561 | 1e-12 | pass |
| FX_GBPCAD: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| FX_GBPCAD: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| FX_GBPCHF, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| FX_GBPCHF, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| FX_GBPCHF, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| FX_GBPCHF, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| FX_GBPCHF: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| FX_GBPCHF: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| FX_GBPCHF, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| FX_GBPCHF, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| FX_GBPCHF: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 21 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| FX_GBPCHF: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| FX_GBPCHF: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| FX_GBPCHF: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 20 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| FX_GBPCHF: the entries followed never grow in number from one session to the next, and all 20 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| FX_GBPCHF: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (1859 stretches) | -0.00183592026 | -0.00183592026 | 1e-12 | pass |
| FX_GBPCHF: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| FX_GBPCHF: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| FX_GBPCHF: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 25 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| FX_GBPCHF: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| FX_GBPCHF: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| FX_GBPCHF: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 24 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| FX_GBPCHF: the entries followed never grow in number from one session to the next, and all 24 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| FX_GBPCHF: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (1854 stretches) | -0.001691203144 | -0.001691203144 | 1e-12 | pass |
| FX_GBPCHF: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| FX_GBPCHF: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| FX_GBPJPY, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| FX_GBPJPY, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| FX_GBPJPY, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| FX_GBPJPY, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| FX_GBPJPY: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| FX_GBPJPY: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| FX_GBPJPY, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| FX_GBPJPY, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| FX_GBPJPY: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 33 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| FX_GBPJPY: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| FX_GBPJPY: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| FX_GBPJPY: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 32 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| FX_GBPJPY: the entries followed never grow in number from one session to the next, and all 32 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| FX_GBPJPY: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (2096 stretches) | 0.019200857738 | 0.019200857738 | 1e-12 | pass |
| FX_GBPJPY: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| FX_GBPJPY: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| FX_GBPJPY: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 23 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| FX_GBPJPY: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| FX_GBPJPY: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| FX_GBPJPY: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 22 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| FX_GBPJPY: the entries followed never grow in number from one session to the next, and all 22 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| FX_GBPJPY: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (2106 stretches) | 0.019056407293 | 0.019056407293 | 1e-12 | pass |
| FX_GBPJPY: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| FX_GBPJPY: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| FX_GBPX, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| FX_GBPX, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| FX_GBPX, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| FX_GBPX, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| FX_GBPX: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| FX_GBPX: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| FX_GBPX, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| FX_GBPX, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| FX_GBPX: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 7 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| FX_GBPX: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| FX_GBPX: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| FX_GBPX: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 6 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| FX_GBPX: the entries followed never grow in number from one session to the next, and all 6 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| FX_GBPX: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (557 stretches) | 0.006153776732 | 0.006153776732 | 1e-12 | pass |
| FX_GBPX: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| FX_GBPX: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| FX_GBPX: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 8 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| FX_GBPX: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| FX_GBPX: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| FX_GBPX: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 7 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| FX_GBPX: the entries followed never grow in number from one session to the next, and all 7 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| FX_GBPX: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (1904 stretches) | 0.005982100431 | 0.005982100431 | 1e-12 | pass |
| FX_GBPX: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| FX_GBPX: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| FX_GLDSLV, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| FX_GLDSLV, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| FX_GLDSLV, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| FX_GLDSLV, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| FX_GLDSLV: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| FX_GLDSLV: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| FX_GLDSLV, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| FX_GLDSLV, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| FX_GLDSLV: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 20 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| FX_GLDSLV: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| FX_GLDSLV: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| FX_GLDSLV: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 19 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| FX_GLDSLV: the entries followed never grow in number from one session to the next, and all 19 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| FX_GLDSLV: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (2090 stretches) | 0.014326340601 | 0.014326340601 | 1e-12 | pass |
| FX_GLDSLV: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| FX_GLDSLV: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| FX_GLDSLV: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 29 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| FX_GLDSLV: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| FX_GLDSLV: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| FX_GLDSLV: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 28 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| FX_GLDSLV: the entries followed never grow in number from one session to the next, and all 28 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| FX_GLDSLV: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (2057 stretches) | 0.011905263739 | 0.011905263739 | 1e-12 | pass |
| FX_GLDSLV: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| FX_GLDSLV: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| FX_JPYX, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| FX_JPYX, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| FX_JPYX, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| FX_JPYX, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| FX_JPYX: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| FX_JPYX: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| FX_JPYX, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| FX_JPYX, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| FX_JPYX: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 32 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| FX_JPYX: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| FX_JPYX: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| FX_JPYX: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 31 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| FX_JPYX: the entries followed never grow in number from one session to the next, and all 31 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| FX_JPYX: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (2055 stretches) | -0.019361997298 | -0.019361997298 | 1e-12 | pass |
| FX_JPYX: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| FX_JPYX: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| FX_JPYX: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 21 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| FX_JPYX: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| FX_JPYX: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| FX_JPYX: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 20 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| FX_JPYX: the entries followed never grow in number from one session to the next, and all 20 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| FX_JPYX: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (2110 stretches) | -0.017477893408 | -0.017477893408 | 1e-12 | pass |
| FX_JPYX: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| FX_JPYX: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| IBIT, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| IBIT, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| IBIT: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| IBIT: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| IBIT: the 30-day risk reversal is read only on days carrying both grid points (days dropped: 0; none invented) | 0 | 0 | 0 | pass |
| IBIT, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| IBIT, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| MSTR, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| MSTR, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| MSTR, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| MSTR, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| MSTR: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| MSTR: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| MSTR: the 30-day risk reversal is read only on days carrying both grid points (days dropped: 0; none invented) | 0 | 0 | 0 | pass |
| MSTR, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| MSTR, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| MSTR: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 16 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| MSTR: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| MSTR: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| MSTR: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 15 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| MSTR: the entries followed never grow in number from one session to the next, and all 15 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| MSTR: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (1916 stretches) | 0.046447252946 | 0.046447252946 | 1e-12 | pass |
| MSTR: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| MSTR: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| MSTR: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 15 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| MSTR: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| MSTR: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| MSTR: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 14 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| MSTR: the entries followed never grow in number from one session to the next, and all 14 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| MSTR: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (1502 stretches) | 0.106421378365 | 0.106421378365 | 1e-12 | pass |
| MSTR: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| MSTR: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| COIN, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| COIN, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| COIN, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| COIN, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| COIN: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| COIN: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| COIN: the 30-day risk reversal is read only on days carrying both grid points (days dropped: 0; none invented) | 0 | 0 | 0 | pass |
| COIN, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| COIN, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| COIN: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 6 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| COIN: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| COIN: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| COIN: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 5 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| COIN: the entries followed never grow in number from one session to the next, and all 5 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| COIN: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (416 stretches) | -0.03156682223 | -0.03156682223 | 1e-12 | pass |
| COIN: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| COIN: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| COIN: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 12 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| COIN: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| COIN: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| COIN: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 11 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| COIN: the entries followed never grow in number from one session to the next, and all 11 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| COIN: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (567 stretches) | -0.090213260135 | -0.090213260135 | 1e-12 | pass |
| COIN: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| COIN: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| NVDA, 30 days: the price distribution's mass (target 1.0, tolerance 0.001) | 1 | 0.999999959101 | 0.001 | pass |
| NVDA, 30 days: the price distribution's no negative (target 0.0, tolerance 0.005) | 0 | -8.706e-9 | 0.005 | pass |
| NVDA, 30 days: the price distribution's mean forward (target 0.0, tolerance 0.001) | 0 | -0.000003248371 | 0.001 | pass |
| NVDA, 30 days: the price distribution's atm iv (target 0.0, tolerance 0.005) | 0 | -0.003458868634 | 0.005 | pass |
| NVDA, 91 days: the price distribution's mass (target 1.0, tolerance 0.001) | 1 | 0.999961905417 | 0.001 | pass |
| NVDA, 91 days: the price distribution's no negative (target 0.0, tolerance 0.005) | 0 | 0 | 0.005 | pass |
| NVDA, 91 days: the price distribution's mean forward (target 0.0, tolerance 0.001) | 0 | -0.000010773301 | 0.001 | pass |
| NVDA, 91 days: the price distribution's atm iv (target 0.0, tolerance 0.005) | 0 | -0.001861552815 | 0.005 | pass |
| NVDA, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| NVDA, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| NVDA, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| NVDA, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| NVDA: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| NVDA: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| NVDA: the 30-day risk reversal is read only on days carrying both grid points (days dropped: 0; none invented) | 0 | 0 | 0 | pass |
| NVDA, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| NVDA, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| NVDA: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 20 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| NVDA: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| NVDA: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| NVDA: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 19 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| NVDA: the entries followed never grow in number from one session to the next, and all 19 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| NVDA: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (1649 stretches) | 0.182863471279 | 0.182863471279 | 1e-12 | pass |
| NVDA: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| NVDA: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| NVDA: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 18 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| NVDA: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| NVDA: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| NVDA: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 17 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| NVDA: the entries followed never grow in number from one session to the next, and all 17 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| NVDA: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (2060 stretches) | 0.12816286723 | 0.12816286723 | 1e-12 | pass |
| NVDA: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| NVDA: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| AMD, 30 days: the price distribution's mass (target 1.0, tolerance 0.001) | 1 | 0.999999886776 | 0.001 | pass |
| AMD, 30 days: the price distribution's no negative (target 0.0, tolerance 0.005) | 0 | -7.509e-9 | 0.005 | pass |
| AMD, 30 days: the price distribution's mean forward (target 0.0, tolerance 0.001) | 0 | -0.00000919208 | 0.001 | pass |
| AMD, 30 days: the price distribution's atm iv (target 0.0, tolerance 0.005) | 0 | 0.001425526196 | 0.005 | pass |
| AMD, 91 days: the price distribution's mass (target 1.0, tolerance 0.001) | 1 | 0.999999750894 | 0.001 | pass |
| AMD, 91 days: the price distribution's no negative (target 0.0, tolerance 0.005) | 0 | 0 | 0.005 | pass |
| AMD, 91 days: the price distribution's mean forward (target 0.0, tolerance 0.001) | 0 | 2.45477e-7 | 0.001 | pass |
| AMD, 91 days: the price distribution's atm iv (target 0.0, tolerance 0.005) | 0 | -0.000991902227 | 0.005 | pass |
| AMD, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| AMD, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| AMD, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| AMD, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| AMD: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| AMD: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| AMD: the 30-day risk reversal is read only on days carrying both grid points (days dropped: 0; none invented) | 0 | 0 | 0 | pass |
| AMD, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| AMD, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| AMD: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 13 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| AMD: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| AMD: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| AMD: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 12 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| AMD: the entries followed never grow in number from one session to the next, and all 12 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| AMD: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (1630 stretches) | 0.086295103123 | 0.086295103123 | 1e-12 | pass |
| AMD: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| AMD: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| AMD: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 20 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| AMD: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| AMD: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| AMD: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 19 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| AMD: the entries followed never grow in number from one session to the next, and all 19 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| AMD: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (1623 stretches) | 0.085313967136 | 0.085313967136 | 1e-12 | pass |
| AMD: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| AMD: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| AVGO, 30 days: the price distribution's mass (target 1.0, tolerance 0.001) | 1 | 0.999999996483 | 0.001 | pass |
| AVGO, 30 days: the price distribution's no negative (target 0.0, tolerance 0.005) | 0 | 0 | 0.005 | pass |
| AVGO, 30 days: the price distribution's mean forward (target 0.0, tolerance 0.001) | 0 | 3.301e-9 | 0.001 | pass |
| AVGO, 30 days: the price distribution's atm iv (target 0.0, tolerance 0.005) | 0 | -0.002714019412 | 0.005 | pass |
| AVGO, 91 days: the price distribution's mass (target 1.0, tolerance 0.001) | 1 | 0.999999994959 | 0.001 | pass |
| AVGO, 91 days: the price distribution's no negative (target 0.0, tolerance 0.005) | 0 | 0 | 0.005 | pass |
| AVGO, 91 days: the price distribution's mean forward (target 0.0, tolerance 0.001) | 0 | 4.928e-9 | 0.001 | pass |
| AVGO, 91 days: the price distribution's atm iv (target 0.0, tolerance 0.005) | 0 | 0.000319491071 | 0.005 | pass |
| AVGO, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| AVGO, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| AVGO, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| AVGO, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| AVGO: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| AVGO: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| AVGO: the 30-day risk reversal is read only on days carrying both grid points (days dropped: 0; none invented) | 0 | 0 | 0 | pass |
| AVGO, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| AVGO, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| AVGO: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 24 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| AVGO: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| AVGO: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| AVGO: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 23 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| AVGO: the entries followed never grow in number from one session to the next, and all 23 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| AVGO: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (2108 stretches) | 0.100921719107 | 0.100921719107 | 1e-12 | pass |
| AVGO: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| AVGO: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| AVGO: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 27 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| AVGO: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| AVGO: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| AVGO: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 26 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| AVGO: the entries followed never grow in number from one session to the next, and all 26 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| AVGO: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (2119 stretches) | 0.100232302271 | 0.100232302271 | 1e-12 | pass |
| AVGO: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| AVGO: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| TSM, 30 days: the price distribution's mass (target 1.0, tolerance 0.001) | 1 | 0.999946827374 | 0.001 | pass |
| TSM, 30 days: the price distribution's no negative (target 0.0, tolerance 0.005) | 0 | 0 | 0.005 | pass |
| TSM, 30 days: the price distribution's mean forward (target 0.0, tolerance 0.001) | 0 | 0.000052591818 | 0.001 | pass |
| TSM, 30 days: the price distribution's atm iv (target 0.0, tolerance 0.005) | 0 | -0.004249962211 | 0.005 | pass |
| TSM, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| TSM, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| TSM, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| TSM, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| TSM: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| TSM: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| TSM: the 30-day risk reversal is read only on days carrying both grid points (days dropped: 0; none invented) | 0 | 0 | 0 | pass |
| TSM, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| TSM, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| TSM: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 17 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| TSM: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| TSM: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| TSM: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 16 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| TSM: the entries followed never grow in number from one session to the next, and all 16 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| TSM: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (2162 stretches) | 0.082039937373 | 0.082039937373 | 1e-12 | pass |
| TSM: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| TSM: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| TSM: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 27 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| TSM: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| TSM: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| TSM: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 26 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| TSM: the entries followed never grow in number from one session to the next, and all 26 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| TSM: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (2162 stretches) | 0.082039937373 | 0.082039937373 | 1e-12 | pass |
| TSM: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| TSM: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| INTC, 30 days: the price distribution's mass (target 1.0, tolerance 0.001) | 1 | 0.99999900151 | 0.001 | pass |
| INTC, 30 days: the price distribution's no negative (target 0.0, tolerance 0.005) | 0 | 0 | 0.005 | pass |
| INTC, 30 days: the price distribution's mean forward (target 0.0, tolerance 0.001) | 0 | -0.000090685864 | 0.001 | pass |
| INTC, 30 days: the price distribution's atm iv (target 0.0, tolerance 0.005) | 0 | -0.004429028006 | 0.005 | pass |
| INTC, 91 days: the price distribution's mass (target 1.0, tolerance 0.001) | 1 | 0.999994421743 | 0.001 | pass |
| INTC, 91 days: the price distribution's no negative (target 0.0, tolerance 0.005) | 0 | 0 | 0.005 | pass |
| INTC, 91 days: the price distribution's mean forward (target 0.0, tolerance 0.001) | 0 | -0.000601158835 | 0.001 | pass |
| INTC, 91 days: the price distribution's atm iv (target 0.0, tolerance 0.005) | 0 | 0.001357455084 | 0.005 | pass |
| INTC, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| INTC, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| INTC, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| INTC, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| INTC: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| INTC: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| INTC: the 30-day risk reversal is read only on days carrying both grid points (days dropped: 0; none invented) | 0 | 0 | 0 | pass |
| INTC, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| INTC, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| INTC: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 18 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| INTC: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| INTC: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| INTC: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 17 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| INTC: the entries followed never grow in number from one session to the next, and all 17 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| INTC: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (1673 stretches) | -0.029884061905 | -0.029884061905 | 1e-12 | pass |
| INTC: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| INTC: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| INTC: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 33 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| INTC: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| INTC: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| INTC: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 32 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| INTC: the entries followed never grow in number from one session to the next, and all 32 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| INTC: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (2158 stretches) | -0.003697256512 | -0.003697256512 | 1e-12 | pass |
| INTC: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| INTC: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| QCOM, 30 days: the price distribution's mass (target 1.0, tolerance 0.001) | 1 | 0.999999805165 | 0.001 | pass |
| QCOM, 30 days: the price distribution's no negative (target 0.0, tolerance 0.005) | 0 | -8.913e-9 | 0.005 | pass |
| QCOM, 30 days: the price distribution's mean forward (target 0.0, tolerance 0.001) | 0 | -0.000015779649 | 0.001 | pass |
| QCOM, 30 days: the price distribution's atm iv (target 0.0, tolerance 0.005) | 0 | -0.003836295479 | 0.005 | pass |
| QCOM, 91 days: the price distribution's mass (target 1.0, tolerance 0.001) | 1 | 0.99999999551 | 0.001 | pass |
| QCOM, 91 days: the price distribution's no negative (target 0.0, tolerance 0.005) | 0 | 0 | 0.005 | pass |
| QCOM, 91 days: the price distribution's mean forward (target 0.0, tolerance 0.001) | 0 | 4.358e-9 | 0.001 | pass |
| QCOM, 91 days: the price distribution's atm iv (target 0.0, tolerance 0.005) | 0 | -0.001249347866 | 0.005 | pass |
| QCOM, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| QCOM, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| QCOM, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| QCOM, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| QCOM: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| QCOM: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| QCOM: the 30-day risk reversal is read only on days carrying both grid points (days dropped: 0; none invented) | 0 | 0 | 0 | pass |
| QCOM, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| QCOM, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| QCOM: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 5 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| QCOM: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| QCOM: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| QCOM: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 4 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| QCOM: the entries followed never grow in number from one session to the next, and all 4 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| QCOM: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (144 stretches) | 0.061382919529 | 0.061382919529 | 1e-12 | pass |
| QCOM: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| QCOM: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| QCOM: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 33 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| QCOM: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| QCOM: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| QCOM: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 32 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| QCOM: the entries followed never grow in number from one session to the next, and all 32 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| QCOM: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (2100 stretches) | 0.037049650925 | 0.037049650925 | 1e-12 | pass |
| QCOM: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| QCOM: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| MU, 30 days: the price distribution's mass (target 1.0, tolerance 0.001) | 1 | 0.999999943933 | 0.001 | pass |
| MU, 30 days: the price distribution's no negative (target 0.0, tolerance 0.005) | 0 | 0 | 0.005 | pass |
| MU, 30 days: the price distribution's mean forward (target 0.0, tolerance 0.001) | 0 | -0.000004349516 | 0.001 | pass |
| MU, 30 days: the price distribution's atm iv (target 0.0, tolerance 0.005) | 0 | -0.004026188248 | 0.005 | pass |
| MU, 91 days: the price distribution's mass (target 1.0, tolerance 0.001) | 1 | 0.999997824546 | 0.001 | pass |
| MU, 91 days: the price distribution's no negative (target 0.0, tolerance 0.005) | 0 | 0 | 0.005 | pass |
| MU, 91 days: the price distribution's mean forward (target 0.0, tolerance 0.001) | 0 | -0.000187101654 | 0.001 | pass |
| MU, 91 days: the price distribution's atm iv (target 0.0, tolerance 0.005) | 0 | 0.000994308197 | 0.005 | pass |
| MU, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| MU, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| MU, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| MU, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| MU: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| MU: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| MU: the 30-day risk reversal is read only on days carrying both grid points (days dropped: 0; none invented) | 0 | 0 | 0 | pass |
| MU, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| MU, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| MU: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 26 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| MU: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| MU: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| MU: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 25 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| MU: the entries followed never grow in number from one session to the next, and all 25 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| MU: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (2061 stretches) | 0.059256771729 | 0.059256771729 | 1e-12 | pass |
| MU: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| MU: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| MU: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 23 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| MU: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| MU: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| MU: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 22 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| MU: the entries followed never grow in number from one session to the next, and all 22 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| MU: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (1820 stretches) | 0.021596110892 | 0.021596110892 | 1e-12 | pass |
| MU: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| MU: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| AMAT, 30 days: the price distribution's mass (target 1.0, tolerance 0.001) | 1 | 0.999998946573 | 0.001 | pass |
| AMAT, 30 days: the price distribution's no negative (target 0.0, tolerance 0.005) | 0 | -0.000362988396 | 0.005 | pass |
| AMAT, 30 days: the price distribution's mean forward (target 0.0, tolerance 0.001) | 0 | -0.000092785581 | 0.001 | pass |
| AMAT, 30 days: the price distribution's atm iv (target 0.0, tolerance 0.005) | 0 | -0.003291332573 | 0.005 | pass |
| AMAT, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| AMAT, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| AMAT, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| AMAT, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| AMAT: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| AMAT: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| AMAT: the 30-day risk reversal is read only on days carrying both grid points (days dropped: 0; none invented) | 0 | 0 | 0 | pass |
| AMAT, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| AMAT, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| AMAT: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 19 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| AMAT: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| AMAT: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| AMAT: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 18 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| AMAT: the entries followed never grow in number from one session to the next, and all 18 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| AMAT: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (2162 stretches) | 0.081335177336 | 0.081335177336 | 1e-12 | pass |
| AMAT: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| AMAT: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| AMAT: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 28 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| AMAT: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| AMAT: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| AMAT: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 27 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| AMAT: the entries followed never grow in number from one session to the next, and all 27 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| AMAT: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (2139 stretches) | 0.08315543584 | 0.08315543584 | 1e-12 | pass |
| AMAT: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| AMAT: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| LRCX, 91 days: the price distribution's mass (target 1.0, tolerance 0.001) | 1 | 0.999999994532 | 0.001 | pass |
| LRCX, 91 days: the price distribution's no negative (target 0.0, tolerance 0.005) | 0 | 0 | 0.005 | pass |
| LRCX, 91 days: the price distribution's mean forward (target 0.0, tolerance 0.001) | 0 | 5.38e-9 | 0.001 | pass |
| LRCX, 91 days: the price distribution's atm iv (target 0.0, tolerance 0.005) | 0 | 0.003075155679 | 0.005 | pass |
| LRCX, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| LRCX, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| LRCX, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| LRCX, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| LRCX: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| LRCX: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| LRCX: the 30-day risk reversal is read only on days carrying both grid points (days dropped: 0; none invented) | 0 | 0 | 0 | pass |
| LRCX, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| LRCX, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| LRCX: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 14 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| LRCX: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| LRCX: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| LRCX: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 13 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| LRCX: the entries followed never grow in number from one session to the next, and all 13 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| LRCX: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (2077 stretches) | 0.091962482557 | 0.091962482557 | 1e-12 | pass |
| LRCX: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| LRCX: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| LRCX: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 28 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| LRCX: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| LRCX: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| LRCX: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 27 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| LRCX: the entries followed never grow in number from one session to the next, and all 27 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| LRCX: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (2149 stretches) | 0.085610009786 | 0.085610009786 | 1e-12 | pass |
| LRCX: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| LRCX: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| MRVL, 30 days: the price distribution's mass (target 1.0, tolerance 0.001) | 1 | 0.999999995914 | 0.001 | pass |
| MRVL, 30 days: the price distribution's no negative (target 0.0, tolerance 0.005) | 0 | 0 | 0.005 | pass |
| MRVL, 30 days: the price distribution's mean forward (target 0.0, tolerance 0.001) | 0 | 3.96e-9 | 0.001 | pass |
| MRVL, 30 days: the price distribution's atm iv (target 0.0, tolerance 0.005) | 0 | -0.000135953785 | 0.005 | pass |
| MRVL, 91 days: the price distribution's mass (target 1.0, tolerance 0.001) | 1 | 0.999999620767 | 0.001 | pass |
| MRVL, 91 days: the price distribution's no negative (target 0.0, tolerance 0.005) | 0 | 0 | 0.005 | pass |
| MRVL, 91 days: the price distribution's mean forward (target 0.0, tolerance 0.001) | 0 | 3.73658e-7 | 0.001 | pass |
| MRVL, 91 days: the price distribution's atm iv (target 0.0, tolerance 0.005) | 0 | -0.000182670309 | 0.005 | pass |
| MRVL, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| MRVL, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| MRVL, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| MRVL, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| MRVL: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| MRVL: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| MRVL: the 30-day risk reversal is read only on days carrying both grid points (days dropped: 0; none invented) | 0 | 0 | 0 | pass |
| MRVL, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| MRVL, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| MRVL: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 25 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| MRVL: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| MRVL: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| MRVL: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 24 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| MRVL: the entries followed never grow in number from one session to the next, and all 24 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| MRVL: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (2144 stretches) | 0.068727638419 | 0.068727638419 | 1e-12 | pass |
| MRVL: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| MRVL: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| MRVL: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 22 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| MRVL: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| MRVL: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| MRVL: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 21 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| MRVL: the entries followed never grow in number from one session to the next, and all 21 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| MRVL: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (2141 stretches) | 0.068818733702 | 0.068818733702 | 1e-12 | pass |
| MRVL: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| MRVL: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| CRDO, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| CRDO, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| CRDO, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| CRDO, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| CRDO: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| CRDO: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| CRDO: the 30-day risk reversal is read only on days carrying both grid points (days dropped: 0; none invented) | 0 | 0 | 0 | pass |
| CRDO, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| CRDO, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| CRDO: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 7 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| CRDO: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| CRDO: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| CRDO: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 6 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| CRDO: the entries followed never grow in number from one session to the next, and all 6 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| CRDO: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (373 stretches) | 0.30885833711 | 0.30885833711 | 1e-12 | pass |
| CRDO: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| CRDO: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| CRDO: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 11 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| CRDO: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| CRDO: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| CRDO: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 10 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| CRDO: the entries followed never grow in number from one session to the next, and all 10 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| CRDO: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (373 stretches) | 0.30885833711 | 0.30885833711 | 1e-12 | pass |
| CRDO: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| CRDO: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| SMCI, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| SMCI, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| SMCI, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| SMCI, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| SMCI: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| SMCI: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| SMCI: the 30-day risk reversal is read only on days carrying both grid points (days dropped: 0; none invented) | 0 | 0 | 0 | pass |
| SMCI, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| SMCI, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| SMCI: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 22 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| SMCI: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| SMCI: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| SMCI: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 21 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| SMCI: the entries followed never grow in number from one session to the next, and all 21 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| SMCI: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (2129 stretches) | 0.092524509804 | 0.092524509804 | 1e-12 | pass |
| SMCI: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| SMCI: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| SMCI: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 25 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| SMCI: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| SMCI: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| SMCI: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 24 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| SMCI: the entries followed never grow in number from one session to the next, and all 24 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| SMCI: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (2108 stretches) | 0.089663357133 | 0.089663357133 | 1e-12 | pass |
| SMCI: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| SMCI: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| LITE, 30 days: the price distribution's mass (target 1.0, tolerance 0.001) | 1 | 0.999998626297 | 0.001 | pass |
| LITE, 30 days: the price distribution's no negative (target 0.0, tolerance 0.005) | 0 | -0.000060038505 | 0.005 | pass |
| LITE, 30 days: the price distribution's mean forward (target 0.0, tolerance 0.001) | 0 | -0.000124959439 | 0.001 | pass |
| LITE, 30 days: the price distribution's atm iv (target 0.0, tolerance 0.005) | 0 | 0.000264557296 | 0.005 | pass |
| LITE, 91 days: the price distribution's mass (target 1.0, tolerance 0.001) | 1 | 0.999995146398 | 0.001 | pass |
| LITE, 91 days: the price distribution's no negative (target 0.0, tolerance 0.005) | 0 | 0 | 0.005 | pass |
| LITE, 91 days: the price distribution's mean forward (target 0.0, tolerance 0.001) | 0 | -0.000532022118 | 0.001 | pass |
| LITE, 91 days: the price distribution's atm iv (target 0.0, tolerance 0.005) | 0 | -0.000364226522 | 0.005 | pass |
| LITE, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| LITE, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| LITE, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| LITE, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| LITE: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| LITE: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| LITE: the 30-day risk reversal is read only on days carrying both grid points (days dropped: 0; none invented) | 0 | 0 | 0 | pass |
| LITE, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| LITE, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| LITE: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 14 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| LITE: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| LITE: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| LITE: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 13 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| LITE: the entries followed never grow in number from one session to the next, and all 13 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| LITE: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (1587 stretches) | 0.042179995824 | 0.042179995824 | 1e-12 | pass |
| LITE: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| LITE: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| LITE: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 10 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| LITE: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| LITE: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| LITE: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 9 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| LITE: the entries followed never grow in number from one session to the next, and all 9 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| LITE: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (1712 stretches) | 0.051325257902 | 0.051325257902 | 1e-12 | pass |
| LITE: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| LITE: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| STX, 30 days: the price distribution's mass (target 1.0, tolerance 0.001) | 1 | 0.999998350003 | 0.001 | pass |
| STX, 30 days: the price distribution's no negative (target 0.0, tolerance 0.005) | 0 | -0.000068437305 | 0.005 | pass |
| STX, 30 days: the price distribution's mean forward (target 0.0, tolerance 0.001) | 0 | -0.000152768485 | 0.001 | pass |
| STX, 30 days: the price distribution's atm iv (target 0.0, tolerance 0.005) | 0 | -0.000077000816 | 0.005 | pass |
| STX, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| STX, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| STX, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| STX, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| STX: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| STX: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| STX: the 30-day risk reversal is read only on days carrying both grid points (days dropped: 0; none invented) | 0 | 0 | 0 | pass |
| STX, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| STX, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| STX: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 24 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| STX: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| STX: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| STX: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 23 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| STX: the entries followed never grow in number from one session to the next, and all 23 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| STX: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (2014 stretches) | 0.070144674264 | 0.070144674264 | 1e-12 | pass |
| STX: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| STX: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| STX: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 25 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| STX: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| STX: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| STX: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 24 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| STX: the entries followed never grow in number from one session to the next, and all 24 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| STX: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (1698 stretches) | 0.044379877366 | 0.044379877366 | 1e-12 | pass |
| STX: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| STX: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| WDC, 30 days: the price distribution's mass (target 1.0, tolerance 0.001) | 1 | 0.999999064427 | 0.001 | pass |
| WDC, 30 days: the price distribution's no negative (target 0.0, tolerance 0.005) | 0 | -0.000024868152 | 0.005 | pass |
| WDC, 30 days: the price distribution's mean forward (target 0.0, tolerance 0.001) | 0 | -0.000082371879 | 0.001 | pass |
| WDC, 30 days: the price distribution's atm iv (target 0.0, tolerance 0.005) | 0 | 0.001512749832 | 0.005 | pass |
| WDC, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| WDC, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| WDC, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| WDC, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| WDC: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| WDC: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| WDC: the 30-day risk reversal is read only on days carrying both grid points (days dropped: 0; none invented) | 0 | 0 | 0 | pass |
| WDC, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| WDC, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| WDC: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 25 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| WDC: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| WDC: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| WDC: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 24 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| WDC: the entries followed never grow in number from one session to the next, and all 24 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| WDC: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (2008 stretches) | 0.056455949492 | 0.056455949492 | 1e-12 | pass |
| WDC: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| WDC: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| WDC: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 17 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| WDC: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| WDC: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| WDC: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 16 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| WDC: the entries followed never grow in number from one session to the next, and all 16 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| WDC: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (1800 stretches) | -0.012062612435 | -0.012062612435 | 1e-12 | pass |
| WDC: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| WDC: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| MSFT, 30 days: the price distribution's mass (target 1.0, tolerance 0.001) | 1 | 0.999999997519 | 0.001 | pass |
| MSFT, 30 days: the price distribution's no negative (target 0.0, tolerance 0.005) | 0 | 0 | 0.005 | pass |
| MSFT, 30 days: the price distribution's mean forward (target 0.0, tolerance 0.001) | 0 | 2.082e-9 | 0.001 | pass |
| MSFT, 30 days: the price distribution's atm iv (target 0.0, tolerance 0.005) | 0 | -0.000992879158 | 0.005 | pass |
| MSFT, 91 days: the price distribution's mass (target 1.0, tolerance 0.001) | 1 | 0.999999988636 | 0.001 | pass |
| MSFT, 91 days: the price distribution's no negative (target 0.0, tolerance 0.005) | 0 | 0 | 0.005 | pass |
| MSFT, 91 days: the price distribution's mean forward (target 0.0, tolerance 0.001) | 0 | 1.1129e-8 | 0.001 | pass |
| MSFT, 91 days: the price distribution's atm iv (target 0.0, tolerance 0.005) | 0 | -0.000661252179 | 0.005 | pass |
| MSFT, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| MSFT, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| MSFT, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| MSFT, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| MSFT: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| MSFT: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| MSFT: the 30-day risk reversal is read only on days carrying both grid points (days dropped: 0; none invented) | 0 | 0 | 0 | pass |
| MSFT, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| MSFT, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| MSFT: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 12 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| MSFT: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| MSFT: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| MSFT: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 11 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| MSFT: the entries followed never grow in number from one session to the next, and all 11 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| MSFT: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (1811 stretches) | 0.054131324998 | 0.054131324998 | 1e-12 | pass |
| MSFT: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| MSFT: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| MSFT: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 26 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| MSFT: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| MSFT: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| MSFT: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 25 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| MSFT: the entries followed never grow in number from one session to the next, and all 25 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| MSFT: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (2108 stretches) | 0.058738132963 | 0.058738132963 | 1e-12 | pass |
| MSFT: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| MSFT: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| AAPL, 30 days: the price distribution's mass (target 1.0, tolerance 0.001) | 1 | 0.999999999011 | 0.001 | pass |
| AAPL, 30 days: the price distribution's no negative (target 0.0, tolerance 0.005) | 0 | -0.000010181451 | 0.005 | pass |
| AAPL, 30 days: the price distribution's mean forward (target 0.0, tolerance 0.001) | 0 | -1.1484e-8 | 0.001 | pass |
| AAPL, 30 days: the price distribution's atm iv (target 0.0, tolerance 0.005) | 0 | -0.001114294356 | 0.005 | pass |
| AAPL, 91 days: the price distribution's mass (target 1.0, tolerance 0.001) | 1 | 0.999999995168 | 0.001 | pass |
| AAPL, 91 days: the price distribution's no negative (target 0.0, tolerance 0.005) | 0 | 0 | 0.005 | pass |
| AAPL, 91 days: the price distribution's mean forward (target 0.0, tolerance 0.001) | 0 | 4.541e-9 | 0.001 | pass |
| AAPL, 91 days: the price distribution's atm iv (target 0.0, tolerance 0.005) | 0 | -0.001022772393 | 0.005 | pass |
| AAPL, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| AAPL, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| AAPL, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| AAPL, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| AAPL: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| AAPL: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| AAPL: the 30-day risk reversal is read only on days carrying both grid points (days dropped: 0; none invented) | 0 | 0 | 0 | pass |
| AAPL, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| AAPL, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| AAPL: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 10 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| AAPL: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| AAPL: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| AAPL: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 9 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| AAPL: the entries followed never grow in number from one session to the next, and all 9 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| AAPL: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (1575 stretches) | 0.052114412501 | 0.052114412501 | 1e-12 | pass |
| AAPL: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| AAPL: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| AAPL: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 18 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| AAPL: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| AAPL: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| AAPL: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 17 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| AAPL: the entries followed never grow in number from one session to the next, and all 17 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| AAPL: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (2015 stretches) | 0.06639507269 | 0.06639507269 | 1e-12 | pass |
| AAPL: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| AAPL: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| GOOG, 30 days: the price distribution's mass (target 1.0, tolerance 0.001) | 1 | 0.99999999818 | 0.001 | pass |
| GOOG, 30 days: the price distribution's no negative (target 0.0, tolerance 0.005) | 0 | 0 | 0.005 | pass |
| GOOG, 30 days: the price distribution's mean forward (target 0.0, tolerance 0.001) | 0 | 1.235e-9 | 0.001 | pass |
| GOOG, 30 days: the price distribution's atm iv (target 0.0, tolerance 0.005) | 0 | -0.001600335594 | 0.005 | pass |
| GOOG, 91 days: the price distribution's mass (target 1.0, tolerance 0.001) | 1 | 0.999999998924 | 0.001 | pass |
| GOOG, 91 days: the price distribution's no negative (target 0.0, tolerance 0.005) | 0 | -7.808e-9 | 0.005 | pass |
| GOOG, 91 days: the price distribution's mean forward (target 0.0, tolerance 0.001) | 0 | -7.8573e-8 | 0.001 | pass |
| GOOG, 91 days: the price distribution's atm iv (target 0.0, tolerance 0.005) | 0 | 0.000645750513 | 0.005 | pass |
| GOOG, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| GOOG, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| GOOG, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| GOOG, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| GOOG: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| GOOG: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| GOOG: the 30-day risk reversal is read only on days carrying both grid points (days dropped: 0; none invented) | 0 | 0 | 0 | pass |
| GOOG, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| GOOG, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| GOOG: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 20 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| GOOG: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| GOOG: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| GOOG: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 19 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| GOOG: the entries followed never grow in number from one session to the next, and all 19 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| GOOG: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (1844 stretches) | 0.081153146172 | 0.081153146172 | 1e-12 | pass |
| GOOG: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| GOOG: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| GOOG: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 21 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| GOOG: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| GOOG: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| GOOG: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 20 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| GOOG: the entries followed never grow in number from one session to the next, and all 20 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| GOOG: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (1927 stretches) | 0.080970537262 | 0.080970537262 | 1e-12 | pass |
| GOOG: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| GOOG: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| META, 30 days: the price distribution's mass (target 1.0, tolerance 0.001) | 1 | 0.99999999832 | 0.001 | pass |
| META, 30 days: the price distribution's no negative (target 0.0, tolerance 0.005) | 0 | 0 | 0.005 | pass |
| META, 30 days: the price distribution's mean forward (target 0.0, tolerance 0.001) | 0 | 1.353e-9 | 0.001 | pass |
| META, 30 days: the price distribution's atm iv (target 0.0, tolerance 0.005) | 0 | -0.00116024717 | 0.005 | pass |
| META, 91 days: the price distribution's mass (target 1.0, tolerance 0.001) | 1 | 0.999999990469 | 0.001 | pass |
| META, 91 days: the price distribution's no negative (target 0.0, tolerance 0.005) | 0 | 0 | 0.005 | pass |
| META, 91 days: the price distribution's mean forward (target 0.0, tolerance 0.001) | 0 | 9.075e-9 | 0.001 | pass |
| META, 91 days: the price distribution's atm iv (target 0.0, tolerance 0.005) | 0 | -0.000674242317 | 0.005 | pass |
| META, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| META, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| META, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| META, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| META: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| META: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| META: the 30-day risk reversal is read only on days carrying both grid points (days dropped: 0; none invented) | 0 | 0 | 0 | pass |
| META, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| META, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| META: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 18 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| META: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| META: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| META: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 17 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| META: the entries followed never grow in number from one session to the next, and all 17 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| META: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (1904 stretches) | 0.057177897933 | 0.057177897933 | 1e-12 | pass |
| META: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| META: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| META: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 20 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| META: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| META: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| META: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 19 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| META: the entries followed never grow in number from one session to the next, and all 19 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| META: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (1814 stretches) | 0.053181496616 | 0.053181496616 | 1e-12 | pass |
| META: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| META: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| AMZN, 30 days: the price distribution's mass (target 1.0, tolerance 0.001) | 1 | 0.999999157812 | 0.001 | pass |
| AMZN, 30 days: the price distribution's no negative (target 0.0, tolerance 0.005) | 0 | -1.3381e-8 | 0.005 | pass |
| AMZN, 30 days: the price distribution's mean forward (target 0.0, tolerance 0.001) | 0 | 7.0903e-7 | 0.001 | pass |
| AMZN, 30 days: the price distribution's atm iv (target 0.0, tolerance 0.005) | 0 | -0.001013005328 | 0.005 | pass |
| AMZN, 91 days: the price distribution's mass (target 1.0, tolerance 0.001) | 1 | 0.999999204115 | 0.001 | pass |
| AMZN, 91 days: the price distribution's no negative (target 0.0, tolerance 0.005) | 0 | 0 | 0.005 | pass |
| AMZN, 91 days: the price distribution's mean forward (target 0.0, tolerance 0.001) | 0 | 7.84821e-7 | 0.001 | pass |
| AMZN, 91 days: the price distribution's atm iv (target 0.0, tolerance 0.005) | 0 | -0.000366194356 | 0.005 | pass |
| AMZN, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| AMZN, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| AMZN, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| AMZN, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| AMZN: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| AMZN: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| AMZN: the 30-day risk reversal is read only on days carrying both grid points (days dropped: 0; none invented) | 0 | 0 | 0 | pass |
| AMZN, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| AMZN, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| AMZN: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 26 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| AMZN: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| AMZN: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| AMZN: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 25 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| AMZN: the entries followed never grow in number from one session to the next, and all 25 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| AMZN: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (2162 stretches) | 0.049584852402 | 0.049584852402 | 1e-12 | pass |
| AMZN: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| AMZN: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| AMZN: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 26 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| AMZN: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| AMZN: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| AMZN: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 25 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| AMZN: the entries followed never grow in number from one session to the next, and all 25 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| AMZN: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (2062 stretches) | 0.043324820366 | 0.043324820366 | 1e-12 | pass |
| AMZN: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| AMZN: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| NFLX, 91 days: the price distribution's mass (target 1.0, tolerance 0.001) | 1 | 0.999999773836 | 0.001 | pass |
| NFLX, 91 days: the price distribution's no negative (target 0.0, tolerance 0.005) | 0 | -7.037e-9 | 0.005 | pass |
| NFLX, 91 days: the price distribution's mean forward (target 0.0, tolerance 0.001) | 0 | -0.000018386573 | 0.001 | pass |
| NFLX, 91 days: the price distribution's atm iv (target 0.0, tolerance 0.005) | 0 | -0.001409031645 | 0.005 | pass |
| NFLX, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| NFLX, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| NFLX, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| NFLX, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| NFLX: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| NFLX: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| NFLX: the 30-day risk reversal is read only on days carrying both grid points (days dropped: 0; none invented) | 0 | 0 | 0 | pass |
| NFLX, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| NFLX, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| NFLX: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 19 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| NFLX: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| NFLX: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| NFLX: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 18 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| NFLX: the entries followed never grow in number from one session to the next, and all 18 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| NFLX: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (2033 stretches) | 0.062356774284 | 0.062356774284 | 1e-12 | pass |
| NFLX: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| NFLX: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| NFLX: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 11 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| NFLX: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| NFLX: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| NFLX: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 10 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| NFLX: the entries followed never grow in number from one session to the next, and all 10 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| NFLX: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (1447 stretches) | 0.061851137113 | 0.061851137113 | 1e-12 | pass |
| NFLX: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| NFLX: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| ORCL, 91 days: the price distribution's mass (target 1.0, tolerance 0.001) | 1 | 0.99999998839 | 0.001 | pass |
| ORCL, 91 days: the price distribution's no negative (target 0.0, tolerance 0.005) | 0 | 0 | 0.005 | pass |
| ORCL, 91 days: the price distribution's mean forward (target 0.0, tolerance 0.001) | 0 | 1.1401e-8 | 0.001 | pass |
| ORCL, 91 days: the price distribution's atm iv (target 0.0, tolerance 0.005) | 0 | 0.000592162117 | 0.005 | pass |
| ORCL, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| ORCL, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| ORCL, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| ORCL, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| ORCL: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| ORCL: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| ORCL: the 30-day risk reversal is read only on days carrying both grid points (days dropped: 0; none invented) | 0 | 0 | 0 | pass |
| ORCL, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| ORCL, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| ORCL: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 30 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| ORCL: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| ORCL: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| ORCL: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 29 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| ORCL: the entries followed never grow in number from one session to the next, and all 29 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| ORCL: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (2127 stretches) | 0.045155795893 | 0.045155795893 | 1e-12 | pass |
| ORCL: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| ORCL: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| ORCL: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 21 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| ORCL: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| ORCL: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| ORCL: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 20 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| ORCL: the entries followed never grow in number from one session to the next, and all 20 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| ORCL: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (2064 stretches) | 0.05014015156 | 0.05014015156 | 1e-12 | pass |
| ORCL: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| ORCL: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| CRM, 30 days: the price distribution's mass (target 1.0, tolerance 0.001) | 1 | 0.999999999445 | 0.001 | pass |
| CRM, 30 days: the price distribution's no negative (target 0.0, tolerance 0.005) | 0 | -5.933e-9 | 0.005 | pass |
| CRM, 30 days: the price distribution's mean forward (target 0.0, tolerance 0.001) | 0 | -3.8746e-8 | 0.001 | pass |
| CRM, 30 days: the price distribution's atm iv (target 0.0, tolerance 0.005) | 0 | 0.002449101143 | 0.005 | pass |
| CRM, 91 days: the price distribution's mass (target 1.0, tolerance 0.001) | 1 | 0.999999997043 | 0.001 | pass |
| CRM, 91 days: the price distribution's no negative (target 0.0, tolerance 0.005) | 0 | 0 | 0.005 | pass |
| CRM, 91 days: the price distribution's mean forward (target 0.0, tolerance 0.001) | 0 | -3.71e-9 | 0.001 | pass |
| CRM, 91 days: the price distribution's atm iv (target 0.0, tolerance 0.005) | 0 | -0.000373074248 | 0.005 | pass |
| CRM, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| CRM, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| CRM, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| CRM, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| CRM: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| CRM: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| CRM: the 30-day risk reversal is read only on days carrying both grid points (days dropped: 0; none invented) | 0 | 0 | 0 | pass |
| CRM, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| CRM, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| CRM: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 18 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| CRM: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| CRM: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| CRM: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 17 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| CRM: the entries followed never grow in number from one session to the next, and all 17 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| CRM: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (1778 stretches) | -0.001001510758 | -0.001001510758 | 1e-12 | pass |
| CRM: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| CRM: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| CRM: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 15 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| CRM: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| CRM: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| CRM: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 14 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| CRM: the entries followed never grow in number from one session to the next, and all 14 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| CRM: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (1414 stretches) | -0.008363093156 | -0.008363093156 | 1e-12 | pass |
| CRM: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| CRM: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| ADBE, 30 days: the price distribution's mass (target 1.0, tolerance 0.001) | 1 | 0.999993679136 | 0.001 | pass |
| ADBE, 30 days: the price distribution's no negative (target 0.0, tolerance 0.005) | 0 | 0 | 0.005 | pass |
| ADBE, 30 days: the price distribution's mean forward (target 0.0, tolerance 0.001) | 0 | 0.000006240436 | 0.001 | pass |
| ADBE, 30 days: the price distribution's atm iv (target 0.0, tolerance 0.005) | 0 | 0.003006678868 | 0.005 | pass |
| ADBE, 91 days: the price distribution's mass (target 1.0, tolerance 0.001) | 1 | 0.999889220534 | 0.001 | pass |
| ADBE, 91 days: the price distribution's no negative (target 0.0, tolerance 0.005) | 0 | 0 | 0.005 | pass |
| ADBE, 91 days: the price distribution's mean forward (target 0.0, tolerance 0.001) | 0 | 0.000109571993 | 0.001 | pass |
| ADBE, 91 days: the price distribution's atm iv (target 0.0, tolerance 0.005) | 0 | -0.001917827669 | 0.005 | pass |
| ADBE, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| ADBE, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| ADBE, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| ADBE, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| ADBE: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| ADBE: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| ADBE: the 30-day risk reversal is read only on days carrying both grid points (days dropped: 0; none invented) | 0 | 0 | 0 | pass |
| ADBE, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| ADBE, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| ADBE: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 23 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| ADBE: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| ADBE: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| ADBE: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 22 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| ADBE: the entries followed never grow in number from one session to the next, and all 22 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| ADBE: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (1567 stretches) | -0.027475427968 | -0.027475427968 | 1e-12 | pass |
| ADBE: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| ADBE: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| ADBE: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 28 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| ADBE: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| ADBE: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| ADBE: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 27 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| ADBE: the entries followed never grow in number from one session to the next, and all 27 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| ADBE: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (1620 stretches) | -0.024228911296 | -0.024228911296 | 1e-12 | pass |
| ADBE: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| ADBE: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| NOW, 30 days: the price distribution's mass (target 1.0, tolerance 0.001) | 1 | 0.999999805452 | 0.001 | pass |
| NOW, 30 days: the price distribution's no negative (target 0.0, tolerance 0.005) | 0 | 0 | 0.005 | pass |
| NOW, 30 days: the price distribution's mean forward (target 0.0, tolerance 0.001) | 0 | -4.408e-9 | 0.001 | pass |
| NOW, 30 days: the price distribution's atm iv (target 0.0, tolerance 0.005) | 0 | -0.001855614171 | 0.005 | pass |
| NOW, 91 days: the price distribution's mass (target 1.0, tolerance 0.001) | 1 | 0.999999544032 | 0.001 | pass |
| NOW, 91 days: the price distribution's no negative (target 0.0, tolerance 0.005) | 0 | 0 | 0.005 | pass |
| NOW, 91 days: the price distribution's mean forward (target 0.0, tolerance 0.001) | 0 | -0.000039765171 | 0.001 | pass |
| NOW, 91 days: the price distribution's atm iv (target 0.0, tolerance 0.005) | 0 | -0.001286146641 | 0.005 | pass |
| NOW, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| NOW, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| NOW, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| NOW, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| NOW: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| NOW: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| NOW: the 30-day risk reversal is read only on days carrying both grid points (days dropped: 0; none invented) | 0 | 0 | 0 | pass |
| NOW, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| NOW, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| NOW: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 19 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| NOW: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| NOW: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| NOW: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 18 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| NOW: the entries followed never grow in number from one session to the next, and all 18 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| NOW: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (2014 stretches) | 0.056082838023 | 0.056082838023 | 1e-12 | pass |
| NOW: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| NOW: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| NOW: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 22 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| NOW: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| NOW: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| NOW: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 21 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| NOW: the entries followed never grow in number from one session to the next, and all 21 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| NOW: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (2006 stretches) | 0.057033314187 | 0.057033314187 | 1e-12 | pass |
| NOW: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| NOW: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| SNOW, 30 days: the price distribution's mass (target 1.0, tolerance 0.001) | 1 | 0.999999959483 | 0.001 | pass |
| SNOW, 30 days: the price distribution's no negative (target 0.0, tolerance 0.005) | 0 | -2.0446e-8 | 0.005 | pass |
| SNOW, 30 days: the price distribution's mean forward (target 0.0, tolerance 0.001) | 0 | -0.000003098236 | 0.001 | pass |
| SNOW, 30 days: the price distribution's atm iv (target 0.0, tolerance 0.005) | 0 | -0.001885403461 | 0.005 | pass |
| SNOW, 91 days: the price distribution's mass (target 1.0, tolerance 0.001) | 1 | 0.999999410622 | 0.001 | pass |
| SNOW, 91 days: the price distribution's no negative (target 0.0, tolerance 0.005) | 0 | 0 | 0.005 | pass |
| SNOW, 91 days: the price distribution's mean forward (target 0.0, tolerance 0.001) | 0 | -0.000052210247 | 0.001 | pass |
| SNOW, 91 days: the price distribution's atm iv (target 0.0, tolerance 0.005) | 0 | -0.001291016037 | 0.005 | pass |
| SNOW, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| SNOW, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| SNOW, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| SNOW, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| SNOW: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| SNOW: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| SNOW: the 30-day risk reversal is read only on days carrying both grid points (days dropped: 0; none invented) | 0 | 0 | 0 | pass |
| SNOW, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| SNOW, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| SNOW: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 6 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| SNOW: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| SNOW: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| SNOW: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 5 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| SNOW: the entries followed never grow in number from one session to the next, and all 5 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| SNOW: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (550 stretches) | -0.02257560585 | -0.02257560585 | 1e-12 | pass |
| SNOW: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| SNOW: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| SNOW: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 11 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| SNOW: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| SNOW: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| SNOW: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 10 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| SNOW: the entries followed never grow in number from one session to the next, and all 10 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| SNOW: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (552 stretches) | -0.066288598032 | -0.066288598032 | 1e-12 | pass |
| SNOW: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| SNOW: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| PLTR, 30 days: the price distribution's mass (target 1.0, tolerance 0.001) | 1 | 0.999999850908 | 0.001 | pass |
| PLTR, 30 days: the price distribution's no negative (target 0.0, tolerance 0.005) | 0 | -0.000057378603 | 0.005 | pass |
| PLTR, 30 days: the price distribution's mean forward (target 0.0, tolerance 0.001) | 0 | -0.000011971873 | 0.001 | pass |
| PLTR, 30 days: the price distribution's atm iv (target 0.0, tolerance 0.005) | 0 | -0.001348202695 | 0.005 | pass |
| PLTR, 91 days: the price distribution's mass (target 1.0, tolerance 0.001) | 1 | 0.999997559009 | 0.001 | pass |
| PLTR, 91 days: the price distribution's no negative (target 0.0, tolerance 0.005) | 0 | 0 | 0.005 | pass |
| PLTR, 91 days: the price distribution's mean forward (target 0.0, tolerance 0.001) | 0 | 9.8527e-7 | 0.001 | pass |
| PLTR, 91 days: the price distribution's atm iv (target 0.0, tolerance 0.005) | 0 | -0.001062443423 | 0.005 | pass |
| PLTR, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| PLTR, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| PLTR, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| PLTR, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| PLTR: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| PLTR: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| PLTR: the 30-day risk reversal is read only on days carrying both grid points (days dropped: 0; none invented) | 0 | 0 | 0 | pass |
| PLTR, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| PLTR, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| PLTR: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 11 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| PLTR: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| PLTR: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| PLTR: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 10 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| PLTR: the entries followed never grow in number from one session to the next, and all 10 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| PLTR: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (695 stretches) | 0.26580202743 | 0.26580202743 | 1e-12 | pass |
| PLTR: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| PLTR: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| PLTR: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 12 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| PLTR: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| PLTR: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| PLTR: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 11 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| PLTR: the entries followed never grow in number from one session to the next, and all 11 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| PLTR: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (692 stretches) | 0.265897540396 | 0.265897540396 | 1e-12 | pass |
| PLTR: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| PLTR: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| CRWD, 91 days: the price distribution's mass (target 1.0, tolerance 0.001) | 1 | 0.999999992354 | 0.001 | pass |
| CRWD, 91 days: the price distribution's no negative (target 0.0, tolerance 0.005) | 0 | 0 | 0.005 | pass |
| CRWD, 91 days: the price distribution's mean forward (target 0.0, tolerance 0.001) | 0 | 3.28e-10 | 0.001 | pass |
| CRWD, 91 days: the price distribution's atm iv (target 0.0, tolerance 0.005) | 0 | -0.00037019289 | 0.005 | pass |
| CRWD, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| CRWD, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| CRWD, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| CRWD, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| CRWD: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| CRWD: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| CRWD: the 30-day risk reversal is read only on days carrying both grid points (days dropped: 0; none invented) | 0 | 0 | 0 | pass |
| CRWD, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| CRWD, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| CRWD: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 9 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| CRWD: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| CRWD: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| CRWD: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 8 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| CRWD: the entries followed never grow in number from one session to the next, and all 8 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| CRWD: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (477 stretches) | 0.031543324181 | 0.031543324181 | 1e-12 | pass |
| CRWD: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| CRWD: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| CRWD: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 12 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| CRWD: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| CRWD: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| CRWD: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 11 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| CRWD: the entries followed never grow in number from one session to the next, and all 11 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| CRWD: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (483 stretches) | 0.033280987409 | 0.033280987409 | 1e-12 | pass |
| CRWD: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| CRWD: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| PANW, 30 days: the price distribution's mass (target 1.0, tolerance 0.001) | 1 | 0.999999821021 | 0.001 | pass |
| PANW, 30 days: the price distribution's no negative (target 0.0, tolerance 0.005) | 0 | -1.5926e-8 | 0.005 | pass |
| PANW, 30 days: the price distribution's mean forward (target 0.0, tolerance 0.001) | 0 | -0.00001447037 | 0.001 | pass |
| PANW, 30 days: the price distribution's atm iv (target 0.0, tolerance 0.005) | 0 | 0.000189273843 | 0.005 | pass |
| PANW, 91 days: the price distribution's mass (target 1.0, tolerance 0.001) | 1 | 0.999999994002 | 0.001 | pass |
| PANW, 91 days: the price distribution's no negative (target 0.0, tolerance 0.005) | 0 | 0 | 0.005 | pass |
| PANW, 91 days: the price distribution's mean forward (target 0.0, tolerance 0.001) | 0 | -4.3818e-8 | 0.001 | pass |
| PANW, 91 days: the price distribution's atm iv (target 0.0, tolerance 0.005) | 0 | -0.001287113488 | 0.005 | pass |
| PANW, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| PANW, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| PANW, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| PANW, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| PANW: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| PANW: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| PANW: the 30-day risk reversal is read only on days carrying both grid points (days dropped: 0; none invented) | 0 | 0 | 0 | pass |
| PANW, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| PANW, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| PANW: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 31 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| PANW: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| PANW: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| PANW: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 30 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| PANW: the entries followed never grow in number from one session to the next, and all 30 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| PANW: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (2022 stretches) | 0.074757713118 | 0.074757713118 | 1e-12 | pass |
| PANW: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| PANW: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| PANW: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 22 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| PANW: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| PANW: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| PANW: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 21 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| PANW: the entries followed never grow in number from one session to the next, and all 21 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| PANW: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (1872 stretches) | 0.072718755631 | 0.072718755631 | 1e-12 | pass |
| PANW: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| PANW: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| NET, 30 days: the price distribution's mass (target 1.0, tolerance 0.001) | 1 | 0.99999999726 | 0.001 | pass |
| NET, 30 days: the price distribution's no negative (target 0.0, tolerance 0.005) | 0 | 0 | 0.005 | pass |
| NET, 30 days: the price distribution's mean forward (target 0.0, tolerance 0.001) | 0 | 2.544e-9 | 0.001 | pass |
| NET, 30 days: the price distribution's atm iv (target 0.0, tolerance 0.005) | 0 | 0.000109700812 | 0.005 | pass |
| NET, 91 days: the price distribution's mass (target 1.0, tolerance 0.001) | 1 | 0.999999994741 | 0.001 | pass |
| NET, 91 days: the price distribution's no negative (target 0.0, tolerance 0.005) | 0 | 0 | 0.005 | pass |
| NET, 91 days: the price distribution's mean forward (target 0.0, tolerance 0.001) | 0 | 5.171e-9 | 0.001 | pass |
| NET, 91 days: the price distribution's atm iv (target 0.0, tolerance 0.005) | 0 | -0.001140733075 | 0.005 | pass |
| NET, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| NET, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| NET, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| NET, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| NET: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| NET: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| NET: the 30-day risk reversal is read only on days carrying both grid points (days dropped: 0; none invented) | 0 | 0 | 0 | pass |
| NET, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| NET, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| NET: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 13 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| NET: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| NET: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| NET: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 12 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| NET: the entries followed never grow in number from one session to the next, and all 12 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| NET: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (669 stretches) | 0.127002226204 | 0.127002226204 | 1e-12 | pass |
| NET: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| NET: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| NET: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 12 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| NET: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| NET: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| NET: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 11 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| NET: the entries followed never grow in number from one session to the next, and all 11 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| NET: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (880 stretches) | 0.116902530402 | 0.116902530402 | 1e-12 | pass |
| NET: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| NET: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| SHOP, 30 days: the price distribution's mass (target 1.0, tolerance 0.001) | 1 | 0.999999905141 | 0.001 | pass |
| SHOP, 30 days: the price distribution's no negative (target 0.0, tolerance 0.005) | 0 | -6.022e-9 | 0.005 | pass |
| SHOP, 30 days: the price distribution's mean forward (target 0.0, tolerance 0.001) | 0 | -0.000007437396 | 0.001 | pass |
| SHOP, 30 days: the price distribution's atm iv (target 0.0, tolerance 0.005) | 0 | 0.000501325128 | 0.005 | pass |
| SHOP, 91 days: the price distribution's mass (target 1.0, tolerance 0.001) | 1 | 0.99999981403 | 0.001 | pass |
| SHOP, 91 days: the price distribution's no negative (target 0.0, tolerance 0.005) | 0 | -1.4645e-8 | 0.005 | pass |
| SHOP, 91 days: the price distribution's mean forward (target 0.0, tolerance 0.001) | 0 | -0.000015172019 | 0.001 | pass |
| SHOP, 91 days: the price distribution's atm iv (target 0.0, tolerance 0.005) | 0 | -0.004345622512 | 0.005 | pass |
| SHOP, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| SHOP, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| SHOP, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| SHOP, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| SHOP: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| SHOP: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| SHOP: the 30-day risk reversal is read only on days carrying both grid points (days dropped: 0; none invented) | 0 | 0 | 0 | pass |
| SHOP, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| SHOP, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| SHOP: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 22 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| SHOP: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| SHOP: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| SHOP: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 21 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| SHOP: the entries followed never grow in number from one session to the next, and all 21 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| SHOP: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (1772 stretches) | 0.113923421724 | 0.113923421724 | 1e-12 | pass |
| SHOP: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| SHOP: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| SHOP: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 26 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| SHOP: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| SHOP: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| SHOP: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 25 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| SHOP: the entries followed never grow in number from one session to the next, and all 25 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| SHOP: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (2056 stretches) | 0.101143386694 | 0.101143386694 | 1e-12 | pass |
| SHOP: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| SHOP: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| WIX, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| WIX, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| WIX, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| WIX, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| WIX: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| WIX: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| WIX: the 30-day risk reversal is read only on days carrying both grid points (days dropped: 0; none invented) | 0 | 0 | 0 | pass |
| WIX, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| WIX, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| WIX: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 16 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| WIX: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| WIX: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| WIX: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 15 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| WIX: the entries followed never grow in number from one session to the next, and all 15 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| WIX: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (1573 stretches) | -0.016480738768 | -0.016480738768 | 1e-12 | pass |
| WIX: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| WIX: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| WIX: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 8 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| WIX: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| WIX: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| WIX: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 7 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| WIX: the entries followed never grow in number from one session to the next, and all 7 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| WIX: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (725 stretches) | -0.013711920316 | -0.013711920316 | 1e-12 | pass |
| WIX: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| WIX: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| APP, 30 days: the price distribution's mass (target 1.0, tolerance 0.001) | 1 | 0.999997509494 | 0.001 | pass |
| APP, 30 days: the price distribution's no negative (target 0.0, tolerance 0.005) | 0 | -0.000170365233 | 0.005 | pass |
| APP, 30 days: the price distribution's mean forward (target 0.0, tolerance 0.001) | 0 | -0.000239626105 | 0.001 | pass |
| APP, 30 days: the price distribution's atm iv (target 0.0, tolerance 0.005) | 0 | -0.003912093905 | 0.005 | pass |
| APP, 91 days: the price distribution's mass (target 1.0, tolerance 0.001) | 1 | 0.999999981727 | 0.001 | pass |
| APP, 91 days: the price distribution's no negative (target 0.0, tolerance 0.005) | 0 | 0 | 0.005 | pass |
| APP, 91 days: the price distribution's mean forward (target 0.0, tolerance 0.001) | 0 | -0.000001208956 | 0.001 | pass |
| APP, 91 days: the price distribution's atm iv (target 0.0, tolerance 0.005) | 0 | -0.004947729997 | 0.005 | pass |
| APP, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| APP, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| APP, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| APP, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| APP: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| APP: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| APP: the 30-day risk reversal is read only on days carrying both grid points (days dropped: 0; none invented) | 0 | 0 | 0 | pass |
| APP, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| APP, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| APP: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 9 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| APP: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| APP: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| APP: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 8 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| APP: the entries followed never grow in number from one session to the next, and all 8 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| APP: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (587 stretches) | 0.151820801677 | 0.151820801677 | 1e-12 | pass |
| APP: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| APP: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| APP: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 3 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| APP: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| APP: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| APP: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 2 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| APP: the entries followed never grow in number from one session to the next, and all 2 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| APP: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (126 stretches) | 0.588051092311 | 0.588051092311 | 1e-12 | pass |
| APP: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| APP: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| TTD, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| TTD, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| TTD, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| TTD, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| TTD: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| TTD: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| TTD: the 30-day risk reversal is read only on days carrying both grid points (days dropped: 0; none invented) | 0 | 0 | 0 | pass |
| TTD, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| TTD, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| TTD: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 8 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| TTD: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| TTD: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| TTD: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 7 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| TTD: the entries followed never grow in number from one session to the next, and all 7 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| TTD: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (764 stretches) | 0.144159983776 | 0.144159983776 | 1e-12 | pass |
| TTD: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| TTD: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| TTD: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 8 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| TTD: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| TTD: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| TTD: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 7 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| TTD: the entries followed never grow in number from one session to the next, and all 7 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| TTD: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (1261 stretches) | -0.062107208872 | -0.062107208872 | 1e-12 | pass |
| TTD: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| TTD: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| RDDT, 91 days: the price distribution's mass (target 1.0, tolerance 0.001) | 1 | 0.9999998832 | 0.001 | pass |
| RDDT, 91 days: the price distribution's no negative (target 0.0, tolerance 0.005) | 0 | 0 | 0.005 | pass |
| RDDT, 91 days: the price distribution's mean forward (target 0.0, tolerance 0.001) | 0 | -0.000003109322 | 0.001 | pass |
| RDDT, 91 days: the price distribution's atm iv (target 0.0, tolerance 0.005) | 0 | 0.00282868231 | 0.005 | pass |
| RDDT, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| RDDT, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| RDDT, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| RDDT, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| RDDT: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| RDDT: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| RDDT: the 30-day risk reversal is read only on days carrying both grid points (days dropped: 0; none invented) | 0 | 0 | 0 | pass |
| RDDT, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| RDDT, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| RDDT: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 1 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| RDDT: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| RDDT: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 3 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| RDDT: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| RDDT: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| RDDT: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 2 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| RDDT: the entries followed never grow in number from one session to the next, and all 2 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| RDDT: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (7 stretches) | -0.377371616494 | -0.377371616494 | 1e-12 | pass |
| RDDT: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| RDDT: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| SNAP, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| SNAP, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| SNAP, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| SNAP, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| SNAP: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| SNAP: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| SNAP: the 30-day risk reversal is read only on days carrying both grid points (days dropped: 0; none invented) | 0 | 0 | 0 | pass |
| SNAP, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| SNAP, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| SNAP: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 25 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| SNAP: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| SNAP: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| SNAP: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 24 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| SNAP: the entries followed never grow in number from one session to the next, and all 24 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| SNAP: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (1564 stretches) | -0.038219859722 | -0.038219859722 | 1e-12 | pass |
| SNAP: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| SNAP: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| SNAP: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 20 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| SNAP: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| SNAP: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| SNAP: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 19 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| SNAP: the entries followed never grow in number from one session to the next, and all 19 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| SNAP: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (1388 stretches) | -0.028140176212 | -0.028140176212 | 1e-12 | pass |
| SNAP: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| SNAP: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| DASH, 30 days: the price distribution's mass (target 1.0, tolerance 0.001) | 1 | 0.999999972302 | 0.001 | pass |
| DASH, 30 days: the price distribution's no negative (target 0.0, tolerance 0.005) | 0 | -9.24e-9 | 0.005 | pass |
| DASH, 30 days: the price distribution's mean forward (target 0.0, tolerance 0.001) | 0 | -0.000002084081 | 0.001 | pass |
| DASH, 30 days: the price distribution's atm iv (target 0.0, tolerance 0.005) | 0 | -0.002381985087 | 0.005 | pass |
| DASH, 91 days: the price distribution's mass (target 1.0, tolerance 0.001) | 1 | 0.999999155585 | 0.001 | pass |
| DASH, 91 days: the price distribution's no negative (target 0.0, tolerance 0.005) | 0 | -1.266e-8 | 0.005 | pass |
| DASH, 91 days: the price distribution's mean forward (target 0.0, tolerance 0.001) | 0 | -0.000075465735 | 0.001 | pass |
| DASH, 91 days: the price distribution's atm iv (target 0.0, tolerance 0.005) | 0 | -0.001448987363 | 0.005 | pass |
| DASH, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| DASH, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| DASH, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| DASH, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| DASH: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| DASH: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| DASH: the 30-day risk reversal is read only on days carrying both grid points (days dropped: 0; none invented) | 0 | 0 | 0 | pass |
| DASH, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| DASH, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| DASH: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 9 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| DASH: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| DASH: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| DASH: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 8 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| DASH: the entries followed never grow in number from one session to the next, and all 8 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| DASH: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (438 stretches) | 0.196251557295 | 0.196251557295 | 1e-12 | pass |
| DASH: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| DASH: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| DASH: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 19 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| DASH: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| DASH: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| DASH: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 18 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| DASH: the entries followed never grow in number from one session to the next, and all 18 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| DASH: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (696 stretches) | 0.10342970218 | 0.10342970218 | 1e-12 | pass |
| DASH: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| DASH: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| UBER, 30 days: the price distribution's mass (target 1.0, tolerance 0.001) | 1 | 0.999999836672 | 0.001 | pass |
| UBER, 30 days: the price distribution's no negative (target 0.0, tolerance 0.005) | 0 | -4.056e-9 | 0.005 | pass |
| UBER, 30 days: the price distribution's mean forward (target 0.0, tolerance 0.001) | 0 | -0.000013106981 | 0.001 | pass |
| UBER, 30 days: the price distribution's atm iv (target 0.0, tolerance 0.005) | 0 | -0.004978925988 | 0.005 | pass |
| UBER, 91 days: the price distribution's mass (target 1.0, tolerance 0.001) | 1 | 0.999999972011 | 0.001 | pass |
| UBER, 91 days: the price distribution's no negative (target 0.0, tolerance 0.005) | 0 | 0 | 0.005 | pass |
| UBER, 91 days: the price distribution's mean forward (target 0.0, tolerance 0.001) | 0 | 2.7588e-8 | 0.001 | pass |
| UBER, 91 days: the price distribution's atm iv (target 0.0, tolerance 0.005) | 0 | 0.000201482435 | 0.005 | pass |
| UBER, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| UBER, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| UBER, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| UBER, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| UBER: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| UBER: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| UBER: the 30-day risk reversal is read only on days carrying both grid points (days dropped: 0; none invented) | 0 | 0 | 0 | pass |
| UBER, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| UBER, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| UBER: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 9 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| UBER: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| UBER: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| UBER: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 8 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| UBER: the entries followed never grow in number from one session to the next, and all 8 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| UBER: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (523 stretches) | 0.070359444991 | 0.070359444991 | 1e-12 | pass |
| UBER: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| UBER: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| UBER: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 17 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| UBER: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| UBER: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| UBER: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 16 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| UBER: the entries followed never grow in number from one session to the next, and all 16 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| UBER: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (926 stretches) | 0.057664921992 | 0.057664921992 | 1e-12 | pass |
| UBER: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| UBER: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| ABNB, 91 days: the price distribution's mass (target 1.0, tolerance 0.001) | 1 | 0.99995964422 | 0.001 | pass |
| ABNB, 91 days: the price distribution's no negative (target 0.0, tolerance 0.005) | 0 | 0 | 0.005 | pass |
| ABNB, 91 days: the price distribution's mean forward (target 0.0, tolerance 0.001) | 0 | 0.000039913673 | 0.001 | pass |
| ABNB, 91 days: the price distribution's atm iv (target 0.0, tolerance 0.005) | 0 | -0.002500805462 | 0.005 | pass |
| ABNB, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| ABNB, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| ABNB, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| ABNB, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| ABNB: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| ABNB: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| ABNB: the 30-day risk reversal is read only on days carrying both grid points (days dropped: 0; none invented) | 0 | 0 | 0 | pass |
| ABNB, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| ABNB, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| ABNB: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 3 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| ABNB: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| ABNB: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| ABNB: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 2 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| ABNB: the entries followed never grow in number from one session to the next, and all 2 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| ABNB: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (14 stretches) | 0.008171014352 | 0.008171014352 | 1e-12 | pass |
| ABNB: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| ABNB: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| ABNB: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 4 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| ABNB: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| ABNB: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| ABNB: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 3 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| ABNB: the entries followed never grow in number from one session to the next, and all 3 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| ABNB: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (89 stretches) | 0.140030025931 | 0.140030025931 | 1e-12 | pass |
| ABNB: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| ABNB: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| SE, 91 days: the price distribution's mass (target 1.0, tolerance 0.001) | 1 | 0.999998970825 | 0.001 | pass |
| SE, 91 days: the price distribution's no negative (target 0.0, tolerance 0.005) | 0 | -0.000008268029 | 0.005 | pass |
| SE, 91 days: the price distribution's mean forward (target 0.0, tolerance 0.001) | 0 | -0.000091795723 | 0.001 | pass |
| SE, 91 days: the price distribution's atm iv (target 0.0, tolerance 0.005) | 0 | -0.00163175402 | 0.005 | pass |
| SE, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| SE, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| SE, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| SE, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| SE: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| SE: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| SE: the 30-day risk reversal is read only on days carrying both grid points (days dropped: 0; none invented) | 0 | 0 | 0 | pass |
| SE, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| SE, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| SE: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 5 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| SE: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| SE: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| SE: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 4 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| SE: the entries followed never grow in number from one session to the next, and all 4 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| SE: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (1090 stretches) | -0.003486573702 | -0.003486573702 | 1e-12 | pass |
| SE: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| SE: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| SE: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 8 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| SE: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| SE: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| SE: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 7 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| SE: the entries followed never grow in number from one session to the next, and all 7 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| SE: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (561 stretches) | -0.229494949495 | -0.229494949495 | 1e-12 | pass |
| SE: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| SE: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| BABA, 30 days: the price distribution's mass (target 1.0, tolerance 0.001) | 1 | 0.999999953191 | 0.001 | pass |
| BABA, 30 days: the price distribution's no negative (target 0.0, tolerance 0.005) | 0 | -8.57e-9 | 0.005 | pass |
| BABA, 30 days: the price distribution's mean forward (target 0.0, tolerance 0.001) | 0 | -0.000003634318 | 0.001 | pass |
| BABA, 30 days: the price distribution's atm iv (target 0.0, tolerance 0.005) | 0 | -0.004326974985 | 0.005 | pass |
| BABA, 91 days: the price distribution's mass (target 1.0, tolerance 0.001) | 1 | 0.999999944841 | 0.001 | pass |
| BABA, 91 days: the price distribution's no negative (target 0.0, tolerance 0.005) | 0 | 0 | 0.005 | pass |
| BABA, 91 days: the price distribution's mean forward (target 0.0, tolerance 0.001) | 0 | -0.000004472127 | 0.001 | pass |
| BABA, 91 days: the price distribution's atm iv (target 0.0, tolerance 0.005) | 0 | -0.004064763206 | 0.005 | pass |
| BABA, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| BABA, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| BABA, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| BABA, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| BABA: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| BABA: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| BABA: the 30-day risk reversal is read only on days carrying both grid points (days dropped: 0; none invented) | 0 | 0 | 0 | pass |
| BABA, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| BABA, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| BABA: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 18 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| BABA: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| BABA: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| BABA: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 17 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| BABA: the entries followed never grow in number from one session to the next, and all 17 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| BABA: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (2162 stretches) | -0.031507578177 | -0.031507578177 | 1e-12 | pass |
| BABA: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| BABA: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| BABA: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 22 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| BABA: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| BABA: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| BABA: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 21 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| BABA: the entries followed never grow in number from one session to the next, and all 21 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| BABA: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (2106 stretches) | -0.030077893004 | -0.030077893004 | 1e-12 | pass |
| BABA: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| BABA: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| IBM, 30 days: the price distribution's mass (target 1.0, tolerance 0.001) | 1 | 0.999999972485 | 0.001 | pass |
| IBM, 30 days: the price distribution's no negative (target 0.0, tolerance 0.005) | 0 | 0 | 0.005 | pass |
| IBM, 30 days: the price distribution's mean forward (target 0.0, tolerance 0.001) | 0 | 2.6978e-8 | 0.001 | pass |
| IBM, 30 days: the price distribution's atm iv (target 0.0, tolerance 0.005) | 0 | -0.000786909999 | 0.005 | pass |
| IBM, 91 days: the price distribution's mass (target 1.0, tolerance 0.001) | 1 | 0.999999944142 | 0.001 | pass |
| IBM, 91 days: the price distribution's no negative (target 0.0, tolerance 0.005) | 0 | -7.577e-9 | 0.005 | pass |
| IBM, 91 days: the price distribution's mean forward (target 0.0, tolerance 0.001) | 0 | -0.000004353642 | 0.001 | pass |
| IBM, 91 days: the price distribution's atm iv (target 0.0, tolerance 0.005) | 0 | -0.000506498031 | 0.005 | pass |
| IBM, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| IBM, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| IBM, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| IBM, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| IBM: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| IBM: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| IBM: the 30-day risk reversal is read only on days carrying both grid points (days dropped: 0; none invented) | 0 | 0 | 0 | pass |
| IBM, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| IBM, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| IBM: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 40 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| IBM: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| IBM: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| IBM: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 39 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| IBM: the entries followed never grow in number from one session to the next, and all 39 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| IBM: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (1904 stretches) | 0.036957161188 | 0.036957161188 | 1e-12 | pass |
| IBM: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| IBM: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| IBM: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 45 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| IBM: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| IBM: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| IBM: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 44 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| IBM: the entries followed never grow in number from one session to the next, and all 44 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| IBM: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (1905 stretches) | 0.037010126174 | 0.037010126174 | 1e-12 | pass |
| IBM: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| IBM: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| CSCO, 91 days: the price distribution's mass (target 1.0, tolerance 0.001) | 1 | 0.999999997056 | 0.001 | pass |
| CSCO, 91 days: the price distribution's no negative (target 0.0, tolerance 0.005) | 0 | 0 | 0.005 | pass |
| CSCO, 91 days: the price distribution's mean forward (target 0.0, tolerance 0.001) | 0 | 2.749e-9 | 0.001 | pass |
| CSCO, 91 days: the price distribution's atm iv (target 0.0, tolerance 0.005) | 0 | -0.003763141173 | 0.005 | pass |
| CSCO, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| CSCO, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| CSCO, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| CSCO, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| CSCO: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| CSCO: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| CSCO: the 30-day risk reversal is read only on days carrying both grid points (days dropped: 0; none invented) | 0 | 0 | 0 | pass |
| CSCO, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| CSCO, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| CSCO: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 18 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| CSCO: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| CSCO: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| CSCO: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 17 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| CSCO: the entries followed never grow in number from one session to the next, and all 17 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| CSCO: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (1989 stretches) | 0.03468975605 | 0.03468975605 | 1e-12 | pass |
| CSCO: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| CSCO: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| CSCO: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 28 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| CSCO: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| CSCO: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| CSCO: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 27 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| CSCO: the entries followed never grow in number from one session to the next, and all 27 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| CSCO: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (2139 stretches) | 0.033124732284 | 0.033124732284 | 1e-12 | pass |
| CSCO: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| CSCO: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| DELL, 30 days: the price distribution's mass (target 1.0, tolerance 0.001) | 1 | 0.999999924922 | 0.001 | pass |
| DELL, 30 days: the price distribution's no negative (target 0.0, tolerance 0.005) | 0 | 0 | 0.005 | pass |
| DELL, 30 days: the price distribution's mean forward (target 0.0, tolerance 0.001) | 0 | 7.3997e-8 | 0.001 | pass |
| DELL, 30 days: the price distribution's atm iv (target 0.0, tolerance 0.005) | 0 | -0.004990401474 | 0.005 | pass |
| DELL, 91 days: the price distribution's mass (target 1.0, tolerance 0.001) | 1 | 0.99999236614 | 0.001 | pass |
| DELL, 91 days: the price distribution's no negative (target 0.0, tolerance 0.005) | 0 | -1.80134e-7 | 0.005 | pass |
| DELL, 91 days: the price distribution's mean forward (target 0.0, tolerance 0.001) | 0 | -0.000868525606 | 0.001 | pass |
| DELL, 91 days: the price distribution's atm iv (target 0.0, tolerance 0.005) | 0 | -0.000468553256 | 0.005 | pass |
| DELL, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| DELL, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| DELL, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| DELL, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| DELL: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| DELL: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| DELL: the 30-day risk reversal is read only on days carrying both grid points (days dropped: 0; none invented) | 0 | 0 | 0 | pass |
| DELL, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| DELL, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| DELL: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 5 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| DELL: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| DELL: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| DELL: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 4 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| DELL: the entries followed never grow in number from one session to the next, and all 4 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| DELL: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (203 stretches) | -0.076657051365 | -0.076657051365 | 1e-12 | pass |
| DELL: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| DELL: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| DELL: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 9 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| DELL: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| DELL: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| DELL: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 8 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| DELL: the entries followed never grow in number from one session to the next, and all 8 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| DELL: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (520 stretches) | 0.150483359915 | 0.150483359915 | 1e-12 | pass |
| DELL: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| DELL: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| HPE, 91 days: the price distribution's mass (target 1.0, tolerance 0.001) | 1 | 0.999999950575 | 0.001 | pass |
| HPE, 91 days: the price distribution's no negative (target 0.0, tolerance 0.005) | 0 | 0 | 0.005 | pass |
| HPE, 91 days: the price distribution's mean forward (target 0.0, tolerance 0.001) | 0 | -8.025e-9 | 0.001 | pass |
| HPE, 91 days: the price distribution's atm iv (target 0.0, tolerance 0.005) | 0 | 0.002693211551 | 0.005 | pass |
| HPE, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| HPE, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| HPE, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| HPE, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| HPE: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| HPE: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| HPE: the 30-day risk reversal is read only on days carrying both grid points (days dropped: 0; none invented) | 0 | 0 | 0 | pass |
| HPE, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| HPE, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| HPE: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 15 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| HPE: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| HPE: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| HPE: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 14 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| HPE: the entries followed never grow in number from one session to the next, and all 14 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| HPE: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (1824 stretches) | 0.022702500097 | 0.022702500097 | 1e-12 | pass |
| HPE: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| HPE: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| HPE: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 23 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| HPE: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| HPE: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| HPE: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 22 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| HPE: the entries followed never grow in number from one session to the next, and all 22 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| HPE: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (1800 stretches) | 0.046957635854 | 0.046957635854 | 1e-12 | pass |
| HPE: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| HPE: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| NOK, 30 days: the price distribution's mass (target 1.0, tolerance 0.001) | 1 | 0.999989403199 | 0.001 | pass |
| NOK, 30 days: the price distribution's no negative (target 0.0, tolerance 0.005) | 0 | 0 | 0.005 | pass |
| NOK, 30 days: the price distribution's mean forward (target 0.0, tolerance 0.001) | 0 | 0.000010463311 | 0.001 | pass |
| NOK, 30 days: the price distribution's atm iv (target 0.0, tolerance 0.005) | 0 | -0.00377152965 | 0.005 | pass |
| NOK, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| NOK, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| NOK, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| NOK, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| NOK: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| NOK: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| NOK: the 30-day risk reversal is read only on days carrying both grid points (days dropped: 0; none invented) | 0 | 0 | 0 | pass |
| NOK, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| NOK, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| NOK: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 19 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| NOK: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| NOK: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| NOK: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 18 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| NOK: the entries followed never grow in number from one session to the next, and all 18 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| NOK: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (2012 stretches) | 0.014928128587 | 0.014928128587 | 1e-12 | pass |
| NOK: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| NOK: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| NOK: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 13 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| NOK: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| NOK: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| NOK: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 12 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| NOK: the entries followed never grow in number from one session to the next, and all 12 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| NOK: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (2162 stretches) | 0.013971538339 | 0.013971538339 | 1e-12 | pass |
| NOK: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| NOK: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| BB, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| BB, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| BB, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| BB, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| BB: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| BB: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| BB: the 30-day risk reversal is read only on days carrying both grid points (days dropped: 0; none invented) | 0 | 0 | 0 | pass |
| BB, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| BB, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| BB: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 7 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| BB: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| BB: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| BB: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 6 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| BB: the entries followed never grow in number from one session to the next, and all 6 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| BB: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (1374 stretches) | -0.048091978385 | -0.048091978385 | 1e-12 | pass |
| BB: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| BB: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| BB: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 2 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| BB: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| BB: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| BB: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 1 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| BB: the entries followed never grow in number from one session to the next, and all 1 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| BB: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| BB: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| NBIS, 30 days: the price distribution's mass (target 1.0, tolerance 0.001) | 1 | 0.99999397646 | 0.001 | pass |
| NBIS, 30 days: the price distribution's no negative (target 0.0, tolerance 0.005) | 0 | 0 | 0.005 | pass |
| NBIS, 30 days: the price distribution's mean forward (target 0.0, tolerance 0.001) | 0 | -0.000643175463 | 0.001 | pass |
| NBIS, 30 days: the price distribution's atm iv (target 0.0, tolerance 0.005) | 0 | -0.000409630517 | 0.005 | pass |
| NBIS, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| NBIS, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| NBIS: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| NBIS: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| NBIS: the 30-day risk reversal is read only on days carrying both grid points (days dropped: 0; none invented) | 0 | 0 | 0 | pass |
| NBIS, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| NBIS, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| CRWV, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| CRWV, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| CRWV: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| CRWV: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| CRWV: the 30-day risk reversal is read only on days carrying both grid points (days dropped: 0; none invented) | 0 | 0 | 0 | pass |
| CRWV, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| CRWV, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| IREN, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| IREN, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| IREN, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| IREN, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| IREN: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| IREN: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| IREN: the 30-day risk reversal is read only on days carrying both grid points (days dropped: 0; none invented) | 0 | 0 | 0 | pass |
| IREN, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| IREN, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| IREN: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 5 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| IREN: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| IREN: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| IREN: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 4 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| IREN: the entries followed never grow in number from one session to the next, and all 4 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| IREN: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (382 stretches) | 0.044993720915 | 0.044993720915 | 1e-12 | pass |
| IREN: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| IREN: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| IREN: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 10 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| IREN: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| IREN: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| IREN: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 9 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| IREN: the entries followed never grow in number from one session to the next, and all 9 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| IREN: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (443 stretches) | -0.042890915231 | -0.042890915231 | 1e-12 | pass |
| IREN: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| IREN: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| CEG, 30 days: the price distribution's mass (target 1.0, tolerance 0.001) | 1 | 0.999999999511 | 0.001 | pass |
| CEG, 30 days: the price distribution's no negative (target 0.0, tolerance 0.005) | 0 | -8.667e-9 | 0.005 | pass |
| CEG, 30 days: the price distribution's mean forward (target 0.0, tolerance 0.001) | 0 | -3.3783e-8 | 0.001 | pass |
| CEG, 30 days: the price distribution's atm iv (target 0.0, tolerance 0.005) | 0 | -0.002009209958 | 0.005 | pass |
| CEG, 91 days: the price distribution's mass (target 1.0, tolerance 0.001) | 1 | 0.999687772981 | 0.001 | pass |
| CEG, 91 days: the price distribution's no negative (target 0.0, tolerance 0.005) | 0 | 0 | 0.005 | pass |
| CEG, 91 days: the price distribution's mean forward (target 0.0, tolerance 0.001) | 0 | 0.000309097636 | 0.001 | pass |
| CEG, 91 days: the price distribution's atm iv (target 0.0, tolerance 0.005) | 0 | 0.000100428988 | 0.005 | pass |
| CEG, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| CEG, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| CEG, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| CEG, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| CEG: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| CEG: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| CEG: the 30-day risk reversal is read only on days carrying both grid points (days dropped: 0; none invented) | 0 | 0 | 0 | pass |
| CEG, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| CEG, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| CEG: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 8 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| CEG: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| CEG: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| CEG: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 7 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| CEG: the entries followed never grow in number from one session to the next, and all 7 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| CEG: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (473 stretches) | -0.001201607566 | -0.001201607566 | 1e-12 | pass |
| CEG: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| CEG: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| CEG: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 5 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| CEG: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| CEG: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| CEG: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 4 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| CEG: the entries followed never grow in number from one session to the next, and all 4 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| CEG: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (453 stretches) | -0.005171771041 | -0.005171771041 | 1e-12 | pass |
| CEG: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| CEG: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| VST, 30 days: the price distribution's mass (target 1.0, tolerance 0.001) | 1 | 0.999999944834 | 0.001 | pass |
| VST, 30 days: the price distribution's no negative (target 0.0, tolerance 0.005) | 0 | -7.057e-9 | 0.005 | pass |
| VST, 30 days: the price distribution's mean forward (target 0.0, tolerance 0.001) | 0 | -0.000004245349 | 0.001 | pass |
| VST, 30 days: the price distribution's atm iv (target 0.0, tolerance 0.005) | 0 | -0.002501851734 | 0.005 | pass |
| VST, 91 days: the price distribution's mass (target 1.0, tolerance 0.001) | 1 | 0.999999999835 | 0.001 | pass |
| VST, 91 days: the price distribution's no negative (target 0.0, tolerance 0.005) | 0 | -1.704e-9 | 0.005 | pass |
| VST, 91 days: the price distribution's mean forward (target 0.0, tolerance 0.001) | 0 | -3.252e-9 | 0.001 | pass |
| VST, 91 days: the price distribution's atm iv (target 0.0, tolerance 0.005) | 0 | 0.002545083689 | 0.005 | pass |
| VST, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| VST, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| VST, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| VST, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| VST: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| VST: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| VST: the 30-day risk reversal is read only on days carrying both grid points (days dropped: 0; none invented) | 0 | 0 | 0 | pass |
| VST, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| VST, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| VST: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 9 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| VST: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| VST: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| VST: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 8 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| VST: the entries followed never grow in number from one session to the next, and all 8 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| VST: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (1100 stretches) | 0.11927893534 | 0.11927893534 | 1e-12 | pass |
| VST: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| VST: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| VST: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 14 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| VST: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| VST: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| VST: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 13 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| VST: the entries followed never grow in number from one session to the next, and all 13 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| VST: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (1043 stretches) | 0.064037062695 | 0.064037062695 | 1e-12 | pass |
| VST: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| VST: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| GEV, 30 days: the price distribution's mass (target 1.0, tolerance 0.001) | 1 | 0.999999992792 | 0.001 | pass |
| GEV, 30 days: the price distribution's no negative (target 0.0, tolerance 0.005) | 0 | -0.000063728648 | 0.005 | pass |
| GEV, 30 days: the price distribution's mean forward (target 0.0, tolerance 0.001) | 0 | -5.25163e-7 | 0.001 | pass |
| GEV, 30 days: the price distribution's atm iv (target 0.0, tolerance 0.005) | 0 | -0.000938432831 | 0.005 | pass |
| GEV, 91 days: the price distribution's mass (target 1.0, tolerance 0.001) | 1 | 0.999999995573 | 0.001 | pass |
| GEV, 91 days: the price distribution's no negative (target 0.0, tolerance 0.005) | 0 | 0 | 0.005 | pass |
| GEV, 91 days: the price distribution's mean forward (target 0.0, tolerance 0.001) | 0 | 4.309e-9 | 0.001 | pass |
| GEV, 91 days: the price distribution's atm iv (target 0.0, tolerance 0.005) | 0 | 0.001675285243 | 0.005 | pass |
| GEV, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| GEV, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| GEV, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| GEV, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| GEV: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| GEV: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| GEV: the 30-day risk reversal is read only on days carrying both grid points (days dropped: 0; none invented) | 0 | 0 | 0 | pass |
| GEV, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| GEV, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| GEV: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 1 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| GEV: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| GEV: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 3 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| GEV: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| GEV: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| GEV: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 2 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| GEV: the entries followed never grow in number from one session to the next, and all 2 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| GEV: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (8 stretches) | -0.219518433344 | -0.219518433344 | 1e-12 | pass |
| GEV: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| GEV: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| BE, 30 days: the price distribution's mass (target 1.0, tolerance 0.001) | 1 | 0.999999994712 | 0.001 | pass |
| BE, 30 days: the price distribution's no negative (target 0.0, tolerance 0.005) | 0 | 0 | 0.005 | pass |
| BE, 30 days: the price distribution's mean forward (target 0.0, tolerance 0.001) | 0 | 5.187e-9 | 0.001 | pass |
| BE, 30 days: the price distribution's atm iv (target 0.0, tolerance 0.005) | 0 | 0.001661524282 | 0.005 | pass |
| BE, 91 days: the price distribution's mass (target 1.0, tolerance 0.001) | 1 | 0.999924323261 | 0.001 | pass |
| BE, 91 days: the price distribution's no negative (target 0.0, tolerance 0.005) | 0 | 0 | 0.005 | pass |
| BE, 91 days: the price distribution's mean forward (target 0.0, tolerance 0.001) | 0 | 0.000072418635 | 0.001 | pass |
| BE, 91 days: the price distribution's atm iv (target 0.0, tolerance 0.005) | 0 | -0.004609645198 | 0.005 | pass |
| BE, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| BE, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| BE, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| BE, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| BE: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| BE: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| BE: the 30-day risk reversal is read only on days carrying both grid points (days dropped: 0; none invented) | 0 | 0 | 0 | pass |
| BE, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| BE, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| BE: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 6 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| BE: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| BE: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| BE: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 5 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| BE: the entries followed never grow in number from one session to the next, and all 5 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| BE: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (983 stretches) | 0.003026634383 | 0.003026634383 | 1e-12 | pass |
| BE: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| BE: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| BE: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 18 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| BE: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| BE: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| BE: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 17 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| BE: the entries followed never grow in number from one session to the next, and all 17 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| BE: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (1119 stretches) | 0.01871657754 | 0.01871657754 | 1e-12 | pass |
| BE: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| BE: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| OXY, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| OXY, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| OXY, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| OXY, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| OXY: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| OXY: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| OXY: the 30-day risk reversal is read only on days carrying both grid points (days dropped: 0; none invented) | 0 | 0 | 0 | pass |
| OXY, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| OXY, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| OXY: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 11 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| OXY: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| OXY: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| OXY: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 10 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| OXY: the entries followed never grow in number from one session to the next, and all 10 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| OXY: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (1619 stretches) | 0.017304073871 | 0.017304073871 | 1e-12 | pass |
| OXY: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| OXY: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| OXY: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 13 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| OXY: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| OXY: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| OXY: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 12 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| OXY: the entries followed never grow in number from one session to the next, and all 12 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| OXY: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (1161 stretches) | -0.008524115432 | -0.008524115432 | 1e-12 | pass |
| OXY: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| OXY: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| XOM, 91 days: the price distribution's mass (target 1.0, tolerance 0.001) | 1 | 0.999999999051 | 0.001 | pass |
| XOM, 91 days: the price distribution's no negative (target 0.0, tolerance 0.005) | 0 | -6.163e-9 | 0.005 | pass |
| XOM, 91 days: the price distribution's mean forward (target 0.0, tolerance 0.001) | 0 | -6.7684e-8 | 0.001 | pass |
| XOM, 91 days: the price distribution's atm iv (target 0.0, tolerance 0.005) | 0 | 0.000586187295 | 0.005 | pass |
| XOM, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| XOM, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| XOM, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| XOM, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| XOM: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| XOM: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| XOM: the 30-day risk reversal is read only on days carrying both grid points (days dropped: 0; none invented) | 0 | 0 | 0 | pass |
| XOM, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| XOM, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| XOM: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 26 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| XOM: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| XOM: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| XOM: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 25 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| XOM: the entries followed never grow in number from one session to the next, and all 25 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| XOM: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (2141 stretches) | 0.032277428986 | 0.032277428986 | 1e-12 | pass |
| XOM: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| XOM: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| XOM: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 18 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| XOM: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| XOM: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| XOM: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 17 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| XOM: the entries followed never grow in number from one session to the next, and all 17 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| XOM: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (1523 stretches) | 0.054836398466 | 0.054836398466 | 1e-12 | pass |
| XOM: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| XOM: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| CVX, 91 days: the price distribution's mass (target 1.0, tolerance 0.001) | 1 | 0.999999998208 | 0.001 | pass |
| CVX, 91 days: the price distribution's no negative (target 0.0, tolerance 0.005) | 0 | 0 | 0.005 | pass |
| CVX, 91 days: the price distribution's mean forward (target 0.0, tolerance 0.001) | 0 | 9.92e-10 | 0.001 | pass |
| CVX, 91 days: the price distribution's atm iv (target 0.0, tolerance 0.005) | 0 | -0.001063817149 | 0.005 | pass |
| CVX, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| CVX, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| CVX, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| CVX, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| CVX: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| CVX: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| CVX: the 30-day risk reversal is read only on days carrying both grid points (days dropped: 0; none invented) | 0 | 0 | 0 | pass |
| CVX, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| CVX, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| CVX: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 25 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| CVX: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| CVX: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| CVX: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 24 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| CVX: the entries followed never grow in number from one session to the next, and all 24 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| CVX: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (1873 stretches) | 0.022793493135 | 0.022793493135 | 1e-12 | pass |
| CVX: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| CVX: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| CVX: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 21 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| CVX: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| CVX: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| CVX: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 20 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| CVX: the entries followed never grow in number from one session to the next, and all 20 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| CVX: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (2157 stretches) | 0.020326738806 | 0.020326738806 | 1e-12 | pass |
| CVX: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| CVX: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| EQT, 91 days: the price distribution's mass (target 1.0, tolerance 0.001) | 1 | 0.9999998779 | 0.001 | pass |
| EQT, 91 days: the price distribution's no negative (target 0.0, tolerance 0.005) | 0 | 0 | 0.005 | pass |
| EQT, 91 days: the price distribution's mean forward (target 0.0, tolerance 0.001) | 0 | -0.000002650454 | 0.001 | pass |
| EQT, 91 days: the price distribution's atm iv (target 0.0, tolerance 0.005) | 0 | -0.002354569538 | 0.005 | pass |
| EQT, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| EQT, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| EQT, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| EQT, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| EQT: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| EQT: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| EQT: the 30-day risk reversal is read only on days carrying both grid points (days dropped: 0; none invented) | 0 | 0 | 0 | pass |
| EQT, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| EQT, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| EQT: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 28 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| EQT: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| EQT: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| EQT: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 27 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| EQT: the entries followed never grow in number from one session to the next, and all 27 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| EQT: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (2162 stretches) | 0.017555965028 | 0.017555965028 | 1e-12 | pass |
| EQT: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| EQT: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| EQT: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 24 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| EQT: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| EQT: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| EQT: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 23 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| EQT: the entries followed never grow in number from one session to the next, and all 23 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| EQT: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (2082 stretches) | 0.015918541335 | 0.015918541335 | 1e-12 | pass |
| EQT: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| EQT: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| LNG, 91 days: the price distribution's mass (target 1.0, tolerance 0.001) | 1 | 0.999999989998 | 0.001 | pass |
| LNG, 91 days: the price distribution's no negative (target 0.0, tolerance 0.005) | 0 | 0 | 0.005 | pass |
| LNG, 91 days: the price distribution's mean forward (target 0.0, tolerance 0.001) | 0 | 9.756e-9 | 0.001 | pass |
| LNG, 91 days: the price distribution's atm iv (target 0.0, tolerance 0.005) | 0 | 0.001129360939 | 0.005 | pass |
| LNG, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| LNG, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| LNG, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| LNG, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| LNG: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| LNG: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| LNG: the 30-day risk reversal is read only on days carrying both grid points (days dropped: 0; none invented) | 0 | 0 | 0 | pass |
| LNG, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| LNG, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| LNG: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 31 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| LNG: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| LNG: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| LNG: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 30 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| LNG: the entries followed never grow in number from one session to the next, and all 30 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| LNG: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (2141 stretches) | 0.048100242757 | 0.048100242757 | 1e-12 | pass |
| LNG: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| LNG: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| LNG: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 21 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| LNG: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| LNG: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| LNG: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 20 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| LNG: the entries followed never grow in number from one session to the next, and all 20 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| LNG: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (2162 stretches) | 0.048230357443 | 0.048230357443 | 1e-12 | pass |
| LNG: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| LNG: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| APO, 91 days: the price distribution's mass (target 1.0, tolerance 0.001) | 1 | 0.999999908069 | 0.001 | pass |
| APO, 91 days: the price distribution's no negative (target 0.0, tolerance 0.005) | 0 | -7.36e-9 | 0.005 | pass |
| APO, 91 days: the price distribution's mean forward (target 0.0, tolerance 0.001) | 0 | -0.000007452692 | 0.001 | pass |
| APO, 91 days: the price distribution's atm iv (target 0.0, tolerance 0.005) | 0 | -0.002715125595 | 0.005 | pass |
| APO, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| APO, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| APO, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| APO, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| APO: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| APO: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| APO: the 30-day risk reversal is read only on days carrying both grid points (days dropped: 0; none invented) | 0 | 0 | 0 | pass |
| APO, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| APO, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| APO: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 10 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| APO: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| APO: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| APO: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 9 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| APO: the entries followed never grow in number from one session to the next, and all 9 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| APO: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (1461 stretches) | 0.062172856661 | 0.062172856661 | 1e-12 | pass |
| APO: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| APO: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| APO: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 12 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| APO: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| APO: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| APO: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 11 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| APO: the entries followed never grow in number from one session to the next, and all 11 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| APO: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (1464 stretches) | 0.061974615916 | 0.061974615916 | 1e-12 | pass |
| APO: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| APO: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| BX, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| BX, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| BX, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| BX, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| BX: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| BX: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| BX: the 30-day risk reversal is read only on days carrying both grid points (days dropped: 0; none invented) | 0 | 0 | 0 | pass |
| BX, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| BX, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| BX: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 21 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| BX: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| BX: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| BX: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 20 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| BX: the entries followed never grow in number from one session to the next, and all 20 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| BX: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (1821 stretches) | 0.077208898568 | 0.077208898568 | 1e-12 | pass |
| BX: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| BX: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| BX: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 22 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| BX: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| BX: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| BX: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 21 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| BX: the entries followed never grow in number from one session to the next, and all 21 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| BX: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (1458 stretches) | 0.084947106648 | 0.084947106648 | 1e-12 | pass |
| BX: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| BX: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| KKR, 30 days: the price distribution's mass (target 1.0, tolerance 0.001) | 1 | 0.999871957221 | 0.001 | pass |
| KKR, 30 days: the price distribution's no negative (target 0.0, tolerance 0.005) | 0 | 0 | 0.005 | pass |
| KKR, 30 days: the price distribution's mean forward (target 0.0, tolerance 0.001) | 0 | -0.000528034677 | 0.001 | pass |
| KKR, 30 days: the price distribution's atm iv (target 0.0, tolerance 0.005) | 0 | -0.002630880184 | 0.005 | pass |
| KKR, 91 days: the price distribution's mass (target 1.0, tolerance 0.001) | 1 | 0.999999995072 | 0.001 | pass |
| KKR, 91 days: the price distribution's no negative (target 0.0, tolerance 0.005) | 0 | 0 | 0.005 | pass |
| KKR, 91 days: the price distribution's mean forward (target 0.0, tolerance 0.001) | 0 | 4.746e-9 | 0.001 | pass |
| KKR, 91 days: the price distribution's atm iv (target 0.0, tolerance 0.005) | 0 | -0.001742442575 | 0.005 | pass |
| KKR, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| KKR, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| KKR, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| KKR, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| KKR: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| KKR: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| KKR: the 30-day risk reversal is read only on days carrying both grid points (days dropped: 0; none invented) | 0 | 0 | 0 | pass |
| KKR, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| KKR, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| KKR: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 19 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| KKR: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| KKR: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| KKR: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 18 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| KKR: the entries followed never grow in number from one session to the next, and all 18 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| KKR: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (1856 stretches) | 0.11296184069 | 0.11296184069 | 1e-12 | pass |
| KKR: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| KKR: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| KKR: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 35 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| KKR: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| KKR: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| KKR: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 34 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| KKR: the entries followed never grow in number from one session to the next, and all 34 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| KKR: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (2139 stretches) | 0.085135496801 | 0.085135496801 | 1e-12 | pass |
| KKR: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| KKR: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| OWL, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| OWL, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| OWL, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| OWL, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| OWL: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| OWL: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| OWL: the 30-day risk reversal is read only on days carrying both grid points (days dropped: 0; none invented) | 0 | 0 | 0 | pass |
| OWL, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| OWL, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| OWL: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 5 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| OWL: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| OWL: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| OWL: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 4 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| OWL: the entries followed never grow in number from one session to the next, and all 4 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| OWL: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (460 stretches) | -0.075342917678 | -0.075342917678 | 1e-12 | pass |
| OWL: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| OWL: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| OWL: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 6 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| OWL: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| OWL: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| OWL: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 5 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| OWL: the entries followed never grow in number from one session to the next, and all 5 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| OWL: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (446 stretches) | -0.086348730635 | -0.086348730635 | 1e-12 | pass |
| OWL: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| OWL: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| CME, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| CME, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| CME, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| CME, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| CME: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| CME: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| CME: the 30-day risk reversal is read only on days carrying both grid points (days dropped: 0; none invented) | 0 | 0 | 0 | pass |
| CME, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| CME, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| CME: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 23 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| CME: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| CME: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| CME: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 22 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| CME: the entries followed never grow in number from one session to the next, and all 22 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| CME: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (2154 stretches) | 0.038784240943 | 0.038784240943 | 1e-12 | pass |
| CME: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| CME: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| CME: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 28 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| CME: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| CME: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| CME: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 27 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| CME: the entries followed never grow in number from one session to the next, and all 27 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| CME: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (2155 stretches) | 0.038939406546 | 0.038939406546 | 1e-12 | pass |
| CME: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| CME: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| CBOE, 30 days: the price distribution's mass (target 1.0, tolerance 0.001) | 1 | 0.99999999952 | 0.001 | pass |
| CBOE, 30 days: the price distribution's no negative (target 0.0, tolerance 0.005) | 0 | -4.768e-9 | 0.005 | pass |
| CBOE, 30 days: the price distribution's mean forward (target 0.0, tolerance 0.001) | 0 | -2.8827e-8 | 0.001 | pass |
| CBOE, 30 days: the price distribution's atm iv (target 0.0, tolerance 0.005) | 0 | -0.001153759271 | 0.005 | pass |
| CBOE, 91 days: the price distribution's mass (target 1.0, tolerance 0.001) | 1 | 0.99999999791 | 0.001 | pass |
| CBOE, 91 days: the price distribution's no negative (target 0.0, tolerance 0.005) | 0 | 0 | 0.005 | pass |
| CBOE, 91 days: the price distribution's mean forward (target 0.0, tolerance 0.001) | 0 | 1.526e-9 | 0.001 | pass |
| CBOE, 91 days: the price distribution's atm iv (target 0.0, tolerance 0.005) | 0 | -0.002129019343 | 0.005 | pass |
| CBOE, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| CBOE, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| CBOE, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| CBOE, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| CBOE: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| CBOE: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| CBOE: the 30-day risk reversal is read only on days carrying both grid points (days dropped: 0; none invented) | 0 | 0 | 0 | pass |
| CBOE, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| CBOE, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| CBOE: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 23 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| CBOE: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| CBOE: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| CBOE: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 22 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| CBOE: the entries followed never grow in number from one session to the next, and all 22 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| CBOE: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (1989 stretches) | 0.050189312616 | 0.050189312616 | 1e-12 | pass |
| CBOE: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| CBOE: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| CBOE: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 36 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| CBOE: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| CBOE: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| CBOE: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 35 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| CBOE: the entries followed never grow in number from one session to the next, and all 35 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| CBOE: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (1995 stretches) | 0.049768912415 | 0.049768912415 | 1e-12 | pass |
| CBOE: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| CBOE: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| HOOD, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| HOOD, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| HOOD, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| HOOD, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| HOOD: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| HOOD: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| HOOD: the 30-day risk reversal is read only on days carrying both grid points (days dropped: 0; none invented) | 0 | 0 | 0 | pass |
| HOOD, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| HOOD, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| HOOD: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 8 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| HOOD: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| HOOD: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| HOOD: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 7 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| HOOD: the entries followed never grow in number from one session to the next, and all 7 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| HOOD: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (537 stretches) | 0.188542771723 | 0.188542771723 | 1e-12 | pass |
| HOOD: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| HOOD: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| HOOD: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 9 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| HOOD: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| HOOD: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| HOOD: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 8 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| HOOD: the entries followed never grow in number from one session to the next, and all 8 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| HOOD: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (425 stretches) | 0.206310679612 | 0.206310679612 | 1e-12 | pass |
| HOOD: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| HOOD: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| SOFI, 91 days: the price distribution's mass (target 1.0, tolerance 0.001) | 1 | 0.999996148942 | 0.001 | pass |
| SOFI, 91 days: the price distribution's no negative (target 0.0, tolerance 0.005) | 0 | 0 | 0.005 | pass |
| SOFI, 91 days: the price distribution's mean forward (target 0.0, tolerance 0.001) | 0 | -0.000444628009 | 0.001 | pass |
| SOFI, 91 days: the price distribution's atm iv (target 0.0, tolerance 0.005) | 0 | 0.001887499627 | 0.005 | pass |
| SOFI, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| SOFI, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| SOFI, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| SOFI, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| SOFI: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| SOFI: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| SOFI: the 30-day risk reversal is read only on days carrying both grid points (days dropped: 0; none invented) | 0 | 0 | 0 | pass |
| SOFI, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| SOFI, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| SOFI: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 8 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| SOFI: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| SOFI: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| SOFI: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 7 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| SOFI: the entries followed never grow in number from one session to the next, and all 7 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| SOFI: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (494 stretches) | 0.013988312998 | 0.013988312998 | 1e-12 | pass |
| SOFI: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| SOFI: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| SOFI: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 3 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| SOFI: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| SOFI: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| SOFI: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 2 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| SOFI: the entries followed never grow in number from one session to the next, and all 2 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| SOFI: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| SOFI: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| AFRM, 91 days: the price distribution's mass (target 1.0, tolerance 0.001) | 1 | 0.999999981529 | 0.001 | pass |
| AFRM, 91 days: the price distribution's no negative (target 0.0, tolerance 0.005) | 0 | -1.8883e-8 | 0.005 | pass |
| AFRM, 91 days: the price distribution's mean forward (target 0.0, tolerance 0.001) | 0 | -0.000001350566 | 0.001 | pass |
| AFRM, 91 days: the price distribution's atm iv (target 0.0, tolerance 0.005) | 0 | -0.000818784623 | 0.005 | pass |
| AFRM, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| AFRM, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| AFRM, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| AFRM, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| AFRM: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| AFRM: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| AFRM: the 30-day risk reversal is read only on days carrying both grid points (days dropped: 0; none invented) | 0 | 0 | 0 | pass |
| AFRM, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| AFRM, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| AFRM: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 13 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| AFRM: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| AFRM: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| AFRM: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 12 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| AFRM: the entries followed never grow in number from one session to the next, and all 12 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| AFRM: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (648 stretches) | 0.031984633262 | 0.031984633262 | 1e-12 | pass |
| AFRM: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| AFRM: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| AFRM: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 16 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| AFRM: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| AFRM: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| AFRM: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 15 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| AFRM: the entries followed never grow in number from one session to the next, and all 15 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| AFRM: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (643 stretches) | 0.051983962395 | 0.051983962395 | 1e-12 | pass |
| AFRM: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| AFRM: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| PYPL, 91 days: the price distribution's mass (target 1.0, tolerance 0.001) | 1 | 0.999884135525 | 0.001 | pass |
| PYPL, 91 days: the price distribution's no negative (target 0.0, tolerance 0.005) | 0 | 0 | 0.005 | pass |
| PYPL, 91 days: the price distribution's mean forward (target 0.0, tolerance 0.001) | 0 | 0.000114593876 | 0.001 | pass |
| PYPL, 91 days: the price distribution's atm iv (target 0.0, tolerance 0.005) | 0 | 0.004192293295 | 0.005 | pass |
| PYPL, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| PYPL, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| PYPL, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| PYPL, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| PYPL: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| PYPL: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| PYPL: the 30-day risk reversal is read only on days carrying both grid points (days dropped: 0; none invented) | 0 | 0 | 0 | pass |
| PYPL, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| PYPL, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| PYPL: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 6 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| PYPL: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| PYPL: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| PYPL: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 5 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| PYPL: the entries followed never grow in number from one session to the next, and all 5 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| PYPL: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (1528 stretches) | 0.021549708205 | 0.021549708205 | 1e-12 | pass |
| PYPL: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| PYPL: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| PYPL: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 18 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| PYPL: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| PYPL: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| PYPL: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 17 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| PYPL: the entries followed never grow in number from one session to the next, and all 17 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| PYPL: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (2026 stretches) | 0.001496304156 | 0.001496304156 | 1e-12 | pass |
| PYPL: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| PYPL: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| LLY, 30 days: the price distribution's mass (target 1.0, tolerance 0.001) | 1 | 0.99999999939 | 0.001 | pass |
| LLY, 30 days: the price distribution's no negative (target 0.0, tolerance 0.005) | 0 | -1.23e-10 | 0.005 | pass |
| LLY, 30 days: the price distribution's mean forward (target 0.0, tolerance 0.001) | 0 | -4.201e-9 | 0.001 | pass |
| LLY, 30 days: the price distribution's atm iv (target 0.0, tolerance 0.005) | 0 | -0.003561007968 | 0.005 | pass |
| LLY, 91 days: the price distribution's mass (target 1.0, tolerance 0.001) | 1 | 0.999999999448 | 0.001 | pass |
| LLY, 91 days: the price distribution's no negative (target 0.0, tolerance 0.005) | 0 | 0 | 0.005 | pass |
| LLY, 91 days: the price distribution's mean forward (target 0.0, tolerance 0.001) | 0 | -2.7766e-8 | 0.001 | pass |
| LLY, 91 days: the price distribution's atm iv (target 0.0, tolerance 0.005) | 0 | 0.000503307747 | 0.005 | pass |
| LLY, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| LLY, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| LLY, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| LLY, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| LLY: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| LLY: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| LLY: the 30-day risk reversal is read only on days carrying both grid points (days dropped: 0; none invented) | 0 | 0 | 0 | pass |
| LLY, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| LLY, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| LLY: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 13 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| LLY: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| LLY: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| LLY: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 12 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| LLY: the entries followed never grow in number from one session to the next, and all 12 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| LLY: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (2052 stretches) | 0.079479989708 | 0.079479989708 | 1e-12 | pass |
| LLY: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| LLY: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| LLY: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 14 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| LLY: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| LLY: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| LLY: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 13 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| LLY: the entries followed never grow in number from one session to the next, and all 13 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| LLY: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (1927 stretches) | 0.073500668208 | 0.073500668208 | 1e-12 | pass |
| LLY: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| LLY: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| NVO, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| NVO, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| NVO, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| NVO, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| NVO: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| NVO: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| NVO: the 30-day risk reversal is read only on days carrying both grid points (days dropped: 0; none invented) | 0 | 0 | 0 | pass |
| NVO, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| NVO, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| NVO: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 15 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| NVO: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| NVO: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| NVO: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 14 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| NVO: the entries followed never grow in number from one session to the next, and all 14 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| NVO: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (2033 stretches) | 0.047888161808 | 0.047888161808 | 1e-12 | pass |
| NVO: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| NVO: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| NVO: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 39 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| NVO: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| NVO: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| NVO: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 38 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| NVO: the entries followed never grow in number from one session to the next, and all 38 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| NVO: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (2035 stretches) | 0.047861168481 | 0.047861168481 | 1e-12 | pass |
| NVO: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| NVO: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| JNJ, 91 days: the price distribution's mass (target 1.0, tolerance 0.001) | 1 | 0.999999998491 | 0.001 | pass |
| JNJ, 91 days: the price distribution's no negative (target 0.0, tolerance 0.005) | 0 | 0 | 0.005 | pass |
| JNJ, 91 days: the price distribution's mean forward (target 0.0, tolerance 0.001) | 0 | 1.083e-9 | 0.001 | pass |
| JNJ, 91 days: the price distribution's atm iv (target 0.0, tolerance 0.005) | 0 | 0.000586617653 | 0.005 | pass |
| JNJ, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| JNJ, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| JNJ, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| JNJ, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| JNJ: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| JNJ: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| JNJ: the 30-day risk reversal is read only on days carrying both grid points (days dropped: 0; none invented) | 0 | 0 | 0 | pass |
| JNJ, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| JNJ, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| JNJ: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 26 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| JNJ: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| JNJ: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| JNJ: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 25 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| JNJ: the entries followed never grow in number from one session to the next, and all 25 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| JNJ: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (2005 stretches) | 0.033199285687 | 0.033199285687 | 1e-12 | pass |
| JNJ: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| JNJ: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| JNJ: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 32 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| JNJ: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| JNJ: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| JNJ: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 31 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| JNJ: the entries followed never grow in number from one session to the next, and all 31 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| JNJ: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (2154 stretches) | 0.033803079316 | 0.033803079316 | 1e-12 | pass |
| JNJ: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| JNJ: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| PFE, 91 days: the price distribution's mass (target 1.0, tolerance 0.001) | 1 | 0.999999057712 | 0.001 | pass |
| PFE, 91 days: the price distribution's no negative (target 0.0, tolerance 0.005) | 0 | 0 | 0.005 | pass |
| PFE, 91 days: the price distribution's mean forward (target 0.0, tolerance 0.001) | 0 | 9.2967e-7 | 0.001 | pass |
| PFE, 91 days: the price distribution's atm iv (target 0.0, tolerance 0.005) | 0 | -0.002702534552 | 0.005 | pass |
| PFE, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| PFE, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| PFE, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| PFE, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| PFE: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| PFE: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| PFE: the 30-day risk reversal is read only on days carrying both grid points (days dropped: 0; none invented) | 0 | 0 | 0 | pass |
| PFE, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| PFE, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| PFE: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 21 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| PFE: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| PFE: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| PFE: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 20 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| PFE: the entries followed never grow in number from one session to the next, and all 20 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| PFE: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (1656 stretches) | 0.000319628805 | 0.000319628805 | 1e-12 | pass |
| PFE: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| PFE: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| PFE: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 24 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| PFE: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| PFE: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| PFE: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 23 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| PFE: the entries followed never grow in number from one session to the next, and all 23 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| PFE: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (1469 stretches) | 0.001978791842 | 0.001978791842 | 1e-12 | pass |
| PFE: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| PFE: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| CVS, 30 days: the price distribution's mass (target 1.0, tolerance 0.001) | 1 | 0.999595815997 | 0.001 | pass |
| CVS, 30 days: the price distribution's no negative (target 0.0, tolerance 0.005) | 0 | -0.003177371679 | 0.005 | pass |
| CVS, 30 days: the price distribution's mean forward (target 0.0, tolerance 0.001) | 0 | 0.000341446531 | 0.001 | pass |
| CVS, 30 days: the price distribution's atm iv (target 0.0, tolerance 0.005) | 0 | -0.000716186669 | 0.005 | pass |
| CVS, 91 days: the price distribution's mass (target 1.0, tolerance 0.001) | 1 | 0.999999915674 | 0.001 | pass |
| CVS, 91 days: the price distribution's no negative (target 0.0, tolerance 0.005) | 0 | -0.000010399235 | 0.005 | pass |
| CVS, 91 days: the price distribution's mean forward (target 0.0, tolerance 0.001) | 0 | -0.000006691813 | 0.001 | pass |
| CVS, 91 days: the price distribution's atm iv (target 0.0, tolerance 0.005) | 0 | 0.001462133807 | 0.005 | pass |
| CVS, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| CVS, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| CVS, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| CVS, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| CVS: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| CVS: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| CVS: the 30-day risk reversal is read only on days carrying both grid points (days dropped: 0; none invented) | 0 | 0 | 0 | pass |
| CVS, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| CVS, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| CVS: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 23 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| CVS: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| CVS: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| CVS: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 22 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| CVS: the entries followed never grow in number from one session to the next, and all 22 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| CVS: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (1994 stretches) | 0.023566177635 | 0.023566177635 | 1e-12 | pass |
| CVS: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| CVS: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| CVS: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 28 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| CVS: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| CVS: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| CVS: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 27 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| CVS: the entries followed never grow in number from one session to the next, and all 27 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| CVS: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (2140 stretches) | 0.023788212372 | 0.023788212372 | 1e-12 | pass |
| CVS: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| CVS: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| HIMS, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| HIMS, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| HIMS, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| HIMS, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| HIMS: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| HIMS: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| HIMS: the 30-day risk reversal is read only on days carrying both grid points (days dropped: 0; none invented) | 0 | 0 | 0 | pass |
| HIMS, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| HIMS, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| HIMS: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 9 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| HIMS: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| HIMS: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| HIMS: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 8 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| HIMS: the entries followed never grow in number from one session to the next, and all 8 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| HIMS: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (635 stretches) | 0.099236641221 | 0.099236641221 | 1e-12 | pass |
| HIMS: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| HIMS: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| HIMS: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 10 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| HIMS: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| HIMS: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| HIMS: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 9 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| HIMS: the entries followed never grow in number from one session to the next, and all 9 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| HIMS: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (582 stretches) | 0.038581804601 | 0.038581804601 | 1e-12 | pass |
| HIMS: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| HIMS: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| CAT, 30 days: the price distribution's mass (target 1.0, tolerance 0.001) | 1 | 0.999999967878 | 0.001 | pass |
| CAT, 30 days: the price distribution's no negative (target 0.0, tolerance 0.005) | 0 | -0.00025740346 | 0.005 | pass |
| CAT, 30 days: the price distribution's mean forward (target 0.0, tolerance 0.001) | 0 | -0.000002431022 | 0.001 | pass |
| CAT, 30 days: the price distribution's atm iv (target 0.0, tolerance 0.005) | 0 | -0.002674039777 | 0.005 | pass |
| CAT, 91 days: the price distribution's mass (target 1.0, tolerance 0.001) | 1 | 0.999999997125 | 0.001 | pass |
| CAT, 91 days: the price distribution's no negative (target 0.0, tolerance 0.005) | 0 | 0 | 0.005 | pass |
| CAT, 91 days: the price distribution's mean forward (target 0.0, tolerance 0.001) | 0 | 2.697e-9 | 0.001 | pass |
| CAT, 91 days: the price distribution's atm iv (target 0.0, tolerance 0.005) | 0 | 0.001582834484 | 0.005 | pass |
| CAT, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| CAT, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| CAT, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| CAT, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| CAT: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| CAT: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| CAT: the 30-day risk reversal is read only on days carrying both grid points (days dropped: 0; none invented) | 0 | 0 | 0 | pass |
| CAT, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| CAT, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| CAT: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 21 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| CAT: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| CAT: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| CAT: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 20 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| CAT: the entries followed never grow in number from one session to the next, and all 20 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| CAT: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (2123 stretches) | 0.056564459648 | 0.056564459648 | 1e-12 | pass |
| CAT: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| CAT: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| CAT: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 29 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| CAT: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| CAT: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| CAT: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 28 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| CAT: the entries followed never grow in number from one session to the next, and all 28 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| CAT: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (2124 stretches) | 0.056207166201 | 0.056207166201 | 1e-12 | pass |
| CAT: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| CAT: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| LMT, 91 days: the price distribution's mass (target 1.0, tolerance 0.001) | 1 | 0.999999999734 | 0.001 | pass |
| LMT, 91 days: the price distribution's no negative (target 0.0, tolerance 0.005) | 0 | -3.576e-9 | 0.005 | pass |
| LMT, 91 days: the price distribution's mean forward (target 0.0, tolerance 0.001) | 0 | -8.082e-9 | 0.001 | pass |
| LMT, 91 days: the price distribution's atm iv (target 0.0, tolerance 0.005) | 0 | 0.002279036353 | 0.005 | pass |
| LMT, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| LMT, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| LMT, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| LMT, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| LMT: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| LMT: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| LMT: the 30-day risk reversal is read only on days carrying both grid points (days dropped: 0; none invented) | 0 | 0 | 0 | pass |
| LMT, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| LMT, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| LMT: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 23 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| LMT: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| LMT: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| LMT: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 22 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| LMT: the entries followed never grow in number from one session to the next, and all 22 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| LMT: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (1980 stretches) | 0.015713877302 | 0.015713877302 | 1e-12 | pass |
| LMT: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| LMT: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| LMT: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 29 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| LMT: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| LMT: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| LMT: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 28 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| LMT: the entries followed never grow in number from one session to the next, and all 28 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| LMT: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (2149 stretches) | 0.012122624094 | 0.012122624094 | 1e-12 | pass |
| LMT: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| LMT: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| AVAV, 91 days: the price distribution's mass (target 1.0, tolerance 0.001) | 1 | 0.999996883767 | 0.001 | pass |
| AVAV, 91 days: the price distribution's no negative (target 0.0, tolerance 0.005) | 0 | -1.6588e-8 | 0.005 | pass |
| AVAV, 91 days: the price distribution's mean forward (target 0.0, tolerance 0.001) | 0 | -0.000287152191 | 0.001 | pass |
| AVAV, 91 days: the price distribution's atm iv (target 0.0, tolerance 0.005) | 0 | 0.002390926556 | 0.005 | pass |
| AVAV, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| AVAV, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| AVAV, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| AVAV, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| AVAV: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| AVAV: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| AVAV: the 30-day risk reversal is read only on days carrying both grid points (days dropped: 0; none invented) | 0 | 0 | 0 | pass |
| AVAV, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| AVAV, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| AVAV: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 20 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| AVAV: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| AVAV: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| AVAV: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 19 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| AVAV: the entries followed never grow in number from one session to the next, and all 19 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| AVAV: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (2162 stretches) | 0.022897795417 | 0.022897795417 | 1e-12 | pass |
| AVAV: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| AVAV: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| AVAV: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 21 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| AVAV: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| AVAV: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| AVAV: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 20 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| AVAV: the entries followed never grow in number from one session to the next, and all 20 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| AVAV: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (2156 stretches) | 0.021680487613 | 0.021680487613 | 1e-12 | pass |
| AVAV: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| AVAV: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| ONDS, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| ONDS, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| ONDS, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| ONDS, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| ONDS: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| ONDS: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| ONDS: the 30-day risk reversal is read only on days carrying both grid points (days dropped: 2; none invented) | 0 | 0 | 0 | pass |
| ONDS, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| ONDS, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| ONDS: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 10 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| ONDS: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| ONDS: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| ONDS: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 9 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| ONDS: the entries followed never grow in number from one session to the next, and all 9 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| ONDS: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (524 stretches) | 0.201304816457 | 0.201304816457 | 1e-12 | pass |
| ONDS: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| ONDS: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| ONDS: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 9 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| ONDS: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| ONDS: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| ONDS: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 8 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| ONDS: the entries followed never grow in number from one session to the next, and all 8 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| ONDS: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (472 stretches) | 0.303640408392 | 0.303640408392 | 1e-12 | pass |
| ONDS: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| ONDS: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| TSLA, 30 days: the price distribution's mass (target 1.0, tolerance 0.001) | 1 | 0.999999999499 | 0.001 | pass |
| TSLA, 30 days: the price distribution's no negative (target 0.0, tolerance 0.005) | 0 | -1.225e-9 | 0.005 | pass |
| TSLA, 30 days: the price distribution's mean forward (target 0.0, tolerance 0.001) | 0 | -1.7915e-8 | 0.001 | pass |
| TSLA, 30 days: the price distribution's atm iv (target 0.0, tolerance 0.005) | 0 | -0.003475267272 | 0.005 | pass |
| TSLA, 91 days: the price distribution's mass (target 1.0, tolerance 0.001) | 1 | 0.999906950766 | 0.001 | pass |
| TSLA, 91 days: the price distribution's no negative (target 0.0, tolerance 0.005) | 0 | 0 | 0.005 | pass |
| TSLA, 91 days: the price distribution's mean forward (target 0.0, tolerance 0.001) | 0 | 0.000092115098 | 0.001 | pass |
| TSLA, 91 days: the price distribution's atm iv (target 0.0, tolerance 0.005) | 0 | -0.000505273913 | 0.005 | pass |
| TSLA, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| TSLA, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| TSLA, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| TSLA, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| TSLA: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| TSLA: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| TSLA: the 30-day risk reversal is read only on days carrying both grid points (days dropped: 0; none invented) | 0 | 0 | 0 | pass |
| TSLA, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| TSLA, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| TSLA: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 22 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| TSLA: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| TSLA: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| TSLA: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 21 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| TSLA: the entries followed never grow in number from one session to the next, and all 21 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| TSLA: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (1939 stretches) | 0.070255600969 | 0.070255600969 | 1e-12 | pass |
| TSLA: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| TSLA: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| TSLA: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 25 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| TSLA: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| TSLA: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| TSLA: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 24 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| TSLA: the entries followed never grow in number from one session to the next, and all 24 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| TSLA: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (2050 stretches) | 0.054955865747 | 0.054955865747 | 1e-12 | pass |
| TSLA: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| TSLA: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| MCD, 30 days: the price distribution's mass (target 1.0, tolerance 0.001) | 1 | 0.999999998804 | 0.001 | pass |
| MCD, 30 days: the price distribution's no negative (target 0.0, tolerance 0.005) | 0 | -4.73e-10 | 0.005 | pass |
| MCD, 30 days: the price distribution's mean forward (target 0.0, tolerance 0.001) | 0 | -1.5974e-8 | 0.001 | pass |
| MCD, 30 days: the price distribution's atm iv (target 0.0, tolerance 0.005) | 0 | -0.003116675751 | 0.005 | pass |
| MCD, 91 days: the price distribution's mass (target 1.0, tolerance 0.001) | 1 | 0.999999998556 | 0.001 | pass |
| MCD, 91 days: the price distribution's no negative (target 0.0, tolerance 0.005) | 0 | 0 | 0.005 | pass |
| MCD, 91 days: the price distribution's mean forward (target 0.0, tolerance 0.001) | 0 | 1.075e-9 | 0.001 | pass |
| MCD, 91 days: the price distribution's atm iv (target 0.0, tolerance 0.005) | 0 | -0.002013188621 | 0.005 | pass |
| MCD, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| MCD, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| MCD, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| MCD, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| MCD: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| MCD: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| MCD: the 30-day risk reversal is read only on days carrying both grid points (days dropped: 0; none invented) | 0 | 0 | 0 | pass |
| MCD, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| MCD, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| MCD: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 9 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| MCD: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| MCD: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| MCD: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 8 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| MCD: the entries followed never grow in number from one session to the next, and all 8 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| MCD: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (1717 stretches) | 0.040940435057 | 0.040940435057 | 1e-12 | pass |
| MCD: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| MCD: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| MCD: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 22 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| MCD: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| MCD: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| MCD: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 21 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| MCD: the entries followed never grow in number from one session to the next, and all 21 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| MCD: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (1823 stretches) | 0.035964473689 | 0.035964473689 | 1e-12 | pass |
| MCD: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| MCD: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| NKE, 30 days: the price distribution's mass (target 1.0, tolerance 0.001) | 1 | 0.999999996139 | 0.001 | pass |
| NKE, 30 days: the price distribution's no negative (target 0.0, tolerance 0.005) | 0 | 0 | 0.005 | pass |
| NKE, 30 days: the price distribution's mean forward (target 0.0, tolerance 0.001) | 0 | 2.838e-9 | 0.001 | pass |
| NKE, 30 days: the price distribution's atm iv (target 0.0, tolerance 0.005) | 0 | 0.000591622999 | 0.005 | pass |
| NKE, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| NKE, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| NKE, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| NKE, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| NKE: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| NKE: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| NKE: the 30-day risk reversal is read only on days carrying both grid points (days dropped: 0; none invented) | 0 | 0 | 0 | pass |
| NKE, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| NKE, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| NKE: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 21 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| NKE: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| NKE: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| NKE: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 20 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| NKE: the entries followed never grow in number from one session to the next, and all 20 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| NKE: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (2162 stretches) | -0.019815792009 | -0.019815792009 | 1e-12 | pass |
| NKE: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| NKE: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| NKE: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 23 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| NKE: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| NKE: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| NKE: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 22 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| NKE: the entries followed never grow in number from one session to the next, and all 22 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| NKE: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (1902 stretches) | -0.021966303968 | -0.021966303968 | 1e-12 | pass |
| NKE: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| NKE: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| LULU, 91 days: the price distribution's mass (target 1.0, tolerance 0.001) | 1 | 0.999999993311 | 0.001 | pass |
| LULU, 91 days: the price distribution's no negative (target 0.0, tolerance 0.005) | 0 | -9.206e-9 | 0.005 | pass |
| LULU, 91 days: the price distribution's mean forward (target 0.0, tolerance 0.001) | 0 | -4.88707e-7 | 0.001 | pass |
| LULU, 91 days: the price distribution's atm iv (target 0.0, tolerance 0.005) | 0 | 0.00344338971 | 0.005 | pass |
| LULU, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| LULU, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| LULU, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| LULU, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| LULU: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| LULU: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| LULU: the 30-day risk reversal is read only on days carrying both grid points (days dropped: 0; none invented) | 0 | 0 | 0 | pass |
| LULU, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| LULU, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| LULU: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 16 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| LULU: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| LULU: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| LULU: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 15 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| LULU: the entries followed never grow in number from one session to the next, and all 15 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| LULU: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (1891 stretches) | 0.014845841839 | 0.014845841839 | 1e-12 | pass |
| LULU: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| LULU: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| LULU: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 13 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| LULU: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| LULU: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| LULU: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 12 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| LULU: the entries followed never grow in number from one session to the next, and all 12 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| LULU: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (1836 stretches) | 0.007309887172 | 0.007309887172 | 1e-12 | pass |
| LULU: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| LULU: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| ONON, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| ONON, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| ONON, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| ONON, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| ONON: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| ONON: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| ONON: the 30-day risk reversal is read only on days carrying both grid points (days dropped: 0; none invented) | 0 | 0 | 0 | pass |
| ONON, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| ONON, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| ONON: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 9 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| ONON: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| ONON: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| ONON: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 8 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| ONON: the entries followed never grow in number from one session to the next, and all 8 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| ONON: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (494 stretches) | -0.064509521051 | -0.064509521051 | 1e-12 | pass |
| ONON: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| ONON: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| ONON: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 9 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| ONON: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| ONON: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| ONON: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 8 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| ONON: the entries followed never grow in number from one session to the next, and all 8 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| ONON: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (494 stretches) | -0.064509521051 | -0.064509521051 | 1e-12 | pass |
| ONON: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| ONON: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| TGT, 30 days: the price distribution's mass (target 1.0, tolerance 0.001) | 1 | 0.999980872696 | 0.001 | pass |
| TGT, 30 days: the price distribution's no negative (target 0.0, tolerance 0.005) | 0 | -0.000210363614 | 0.005 | pass |
| TGT, 30 days: the price distribution's mean forward (target 0.0, tolerance 0.001) | 0 | 0.000018894998 | 0.001 | pass |
| TGT, 30 days: the price distribution's atm iv (target 0.0, tolerance 0.005) | 0 | -0.004906659883 | 0.005 | pass |
| TGT, 91 days: the price distribution's mass (target 1.0, tolerance 0.001) | 1 | 0.999848467183 | 0.001 | pass |
| TGT, 91 days: the price distribution's no negative (target 0.0, tolerance 0.005) | 0 | 0 | 0.005 | pass |
| TGT, 91 days: the price distribution's mean forward (target 0.0, tolerance 0.001) | 0 | 0.000149903496 | 0.001 | pass |
| TGT, 91 days: the price distribution's atm iv (target 0.0, tolerance 0.005) | 0 | -0.003503583616 | 0.005 | pass |
| TGT, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| TGT, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| TGT, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| TGT, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| TGT: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| TGT: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| TGT: the 30-day risk reversal is read only on days carrying both grid points (days dropped: 0; none invented) | 0 | 0 | 0 | pass |
| TGT, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| TGT, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| TGT: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 6 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| TGT: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| TGT: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| TGT: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 5 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| TGT: the entries followed never grow in number from one session to the next, and all 5 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| TGT: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (668 stretches) | 0.045160952586 | 0.045160952586 | 1e-12 | pass |
| TGT: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| TGT: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| TGT: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 9 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| TGT: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| TGT: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| TGT: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 8 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| TGT: the entries followed never grow in number from one session to the next, and all 8 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| TGT: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (625 stretches) | 0.031001306302 | 0.031001306302 | 1e-12 | pass |
| TGT: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| TGT: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| PM, 91 days: the price distribution's mass (target 1.0, tolerance 0.001) | 1 | 0.999999998195 | 0.001 | pass |
| PM, 91 days: the price distribution's no negative (target 0.0, tolerance 0.005) | 0 | 0 | 0.005 | pass |
| PM, 91 days: the price distribution's mean forward (target 0.0, tolerance 0.001) | 0 | 1.507e-9 | 0.001 | pass |
| PM, 91 days: the price distribution's atm iv (target 0.0, tolerance 0.005) | 0 | 0.003259078633 | 0.005 | pass |
| PM, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| PM, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| PM, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| PM, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| PM: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| PM: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| PM: the 30-day risk reversal is read only on days carrying both grid points (days dropped: 0; none invented) | 0 | 0 | 0 | pass |
| PM, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| PM, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| PM: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 13 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| PM: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| PM: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| PM: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 12 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| PM: the entries followed never grow in number from one session to the next, and all 12 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| PM: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (2162 stretches) | 0.034696886961 | 0.034696886961 | 1e-12 | pass |
| PM: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| PM: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| PM: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 29 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| PM: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| PM: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| PM: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 28 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| PM: the entries followed never grow in number from one session to the next, and all 28 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| PM: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (2142 stretches) | 0.035573674538 | 0.035573674538 | 1e-12 | pass |
| PM: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| PM: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| DKNG, 30 days: the price distribution's mass (target 1.0, tolerance 0.001) | 1 | 0.999933973744 | 0.001 | pass |
| DKNG, 30 days: the price distribution's no negative (target 0.0, tolerance 0.005) | 0 | 0 | 0.005 | pass |
| DKNG, 30 days: the price distribution's mean forward (target 0.0, tolerance 0.001) | 0 | -0.000362559191 | 0.001 | pass |
| DKNG, 30 days: the price distribution's atm iv (target 0.0, tolerance 0.005) | 0 | -0.000362414569 | 0.005 | pass |
| DKNG, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| DKNG, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| DKNG, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| DKNG, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| DKNG: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| DKNG: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| DKNG: the 30-day risk reversal is read only on days carrying both grid points (days dropped: 0; none invented) | 0 | 0 | 0 | pass |
| DKNG, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| DKNG, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| DKNG: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 11 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| DKNG: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| DKNG: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| DKNG: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 10 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| DKNG: the entries followed never grow in number from one session to the next, and all 10 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| DKNG: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (593 stretches) | -0.026708562451 | -0.026708562451 | 1e-12 | pass |
| DKNG: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| DKNG: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| DKNG: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 13 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| DKNG: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| DKNG: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| DKNG: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 12 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| DKNG: the entries followed never grow in number from one session to the next, and all 12 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| DKNG: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (751 stretches) | -0.011506276151 | -0.011506276151 | 1e-12 | pass |
| DKNG: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| DKNG: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| UA, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| UA, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| UA, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| UA, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| UA: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| UA: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| UA: the 30-day risk reversal is read only on days carrying both grid points (days dropped: 0; none invented) | 0 | 0 | 0 | pass |
| UA, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| UA, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| UA: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 13 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| UA: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| UA: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| UA: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 12 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| UA: the entries followed never grow in number from one session to the next, and all 12 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| UA: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (1093 stretches) | -0.078751857355 | -0.078751857355 | 1e-12 | pass |
| UA: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| UA: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| UA: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 9 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| UA: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| UA: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| UA: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 8 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| UA: the entries followed never grow in number from one session to the next, and all 8 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| UA: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (934 stretches) | -0.068217735736 | -0.068217735736 | 1e-12 | pass |
| UA: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| UA: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| WEN, Step PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| WEN, Step PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| WEN, Orthogonal PCs: the paths re-drawn from the seed equal the published bands (max abs difference) | 0 | 0 | 1e-9 | pass |
| WEN, Orthogonal PCs: the record's origins are all on or before the last session used (origins after it) | 0 | 0 | 0 | pass |
| WEN: the engine's fan keeps its percentiles in order on every target date (dates out of order) | 0 | 0 | 0 | pass |
| WEN: the engine's mean ± 2σ lines sit either side of its mean on every target date (dates out of order) | 0 | 0 | 0 | pass |
| WEN: the 30-day risk reversal is read only on days carrying both grid points (days dropped: 0; none invented) | 0 | 0 | 0 | pass |
| WEN, step: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| WEN, normal: every dated step of the paths is on the shared axis (steps off it) | 0 | 0 | 0 | pass |
| WEN: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 23 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| WEN: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| WEN: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| WEN: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 22 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| WEN: the entries followed never grow in number from one session to the next, and all 22 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| WEN: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (2136 stretches) | -0.006201368387 | -0.006201368387 | 1e-12 | pass |
| WEN: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| WEN: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| WEN: the volatility at each episode's entry, the 63 sessions' log returns before it annualised, by pandas and by numpy (max abs; 0 of 29 entries have too few closes before them and are left out of the volatility units, none invented) | 0 | 0 | 1e-12 | pass |
| WEN: the current episode's path equals its closes from the entry close (max abs) | 0 | 0 | 1e-9 | pass |
| WEN: every episode path starts at zero on its entry day (max abs) | 0 | 0 | 1e-9 | pass |
| WEN: every past episode's path equals its closes from the entry close for as long as it ran (max abs; all 28 entries are on the closes) | 0 | 0 | 1e-9 | pass |
| WEN: the entries followed never grow in number from one session to the next, and all 28 start on the entry day (rises, and entries missing on day 0) | 0 | 0 | 0 | pass |
| WEN: the median return of any 65 sessions between the first and the last past entry, by numpy and by pandas (2162 stretches) | -0.006501200844 | -0.006501200844 | 1e-12 | pass |
| WEN: every past episode is in the regime on its entry day (the share still in the regime on day 0, %) | 100 | 100 | 1e-9 | pass |
| WEN: on every session the entries still in the regime and those whose regime had ended make up the entries followed (max abs) | 0 | 0 | 0 | pass |
| Every ticker of the universe is either built or listed with its reason (176 tickers; unaccounted) | 0 | 0 | 0 | pass |
| SPY: the engine's own test finds no change of adjustment basis of 25% or more in its closes (days found; a smaller change is printed with its date; a lead asset the test was not run on fails) | 0 | 0 | 0 | pass |
| SPY: the session folder's component scores are the engine's own published scores (largest difference, 2,479 sessions) | 0 | 0 | 1e-9 | pass |
| Every fan the page draws carries the session folder's closes on its past target dates (175 fans drawn; 0 withheld for it and named; drawn fans that differ) | 0 | 0 | 0 | pass |
| SPY: the regime read from the sign of the three components is the engine's regime on every session inside its episodes (4,454 sessions on 2 bases; sessions that differ) | 0 | 0 | 0 | pass |
| The fork table's rows add to each regime's sessions and to its sessions on which the regime ended, on every basis (max abs) | 0 | 0 | 0 | pass |
| The regimes' sessions add to the sessions of any regime, on every basis (max abs) | 0 | 0 | 0 | pass |
| Every takeaway and tile keeps the house grammar and carries no process word (5 takeaways, 4 tiles; faults) | 0 | 0 | 0 | pass |
| The options source line carries the attribution string the session folder states | marketdata.app | marketdata.app | 0 | pass |
| Decisions: every desk decision this page's specification names is in the decisions file (D-2026-11-10-OL1, D-2026-11-10-OL2, D-2026-11-10-OL3, D-2026-11-10-OL4, D-2026-11-11-OL5, D-2026-11-11-OL6, D-2026-11-11-OL7, D-2026-11-11-OL8, D-2026-10-03-21; ids missing) | 0 | 0 | 0 | pass |
- Options engine session folder: the latest session's end-of-day option chains, the option-based forecasts' component scores and trend slopes, and adjusted closes — marketdata.app