S&P 500 skew level at its 74th percentile of three years
Which tail of each market is bid, and how does the price of that insurance change with the horizon?
In brief
Where protection is dear or cheap today, and what would change the read- S&P 500 protection is dearer than usual: skew level +0.49 z, at its 77th percentile.→ Exhibit 1
- Of 19 markets drawn, protection is dearest in Silver (SLV), its puts: skew level +2.84 z.→ Exhibit 1
- Protection is cheapest against its own history in Brazil (EWZ), its puts: skew level −1.49 z.→ Exhibit 1
- Long Treasuries (TLT) holds its protection furthest towards the short tenors: front-end tilt +1.30 z.→ Exhibit 2
- Implied volatility is highest against its own history in Biotech (XBI): its level at the 96th percentile of three years.→ Exhibit 1
| What would change this view | Trigger | Today | Decided by |
|---|---|---|---|
| S&P 500 protection turns dear | skew level at or above +1 z | +0.49 z (16 Nov 2026)not crossed | each session's close (next 17 Nov 2026) |
| S&P 500 implied volatility turns ordinary | its level below its 75th percentile | 77th (16 Nov 2026)not crossed | each session's close (next 17 Nov 2026) |
| Calls become dearer than puts on SPY | 25-day below 0 points | +7.0 (16 Nov 2026)not crossed | each session's close (next 17 Nov 2026) |
The reasoning behind each point
- Skew and implied volatilityS&P 500 protection is dearer than usual: skew level +0.49 z, implied volatility at its 77th percentile.
The skew level is the level of the market's skew curve — the risk reversal as a share of at-the-money volatility, across the 25- to 125-day tenors — as a robust z against its own last three years, signed so that above zero its protective wing costs more than usual. The implied volatility is the level of its at-the-money volatility curve, as a percentile of the same three years.
- Dearest protectionOf 19 markets drawn, protection is dearest in Silver (SLV), its puts: skew level +2.84 z.
5 of 19 markets pay at least one robust standard deviation more than usual for their protective wing.
- Cheapest protectionProtection is cheapest against its own history in Brazil (EWZ), its puts: skew level −1.49 z.
2 of 19 markets pay at least one robust standard deviation less than usual.
- Front-end tiltLong Treasuries (TLT) holds its protection furthest towards the short tenors: front-end tilt +1.30 z.
The front-end tilt is the skew slope's z with its sign reversed, signed by the protective wing: above zero when the short tenors carry more of the skew than usual.
- Highest implied volatilityImplied volatility is highest against its own history in Biotech (XBI): its level at the 96th percentile of three years.
The percentile of the level of each market's at-the-money volatility curve, on its five-session mean.
- Sign ruleWhen stocks and bonds moved together, SPY's 25-day risk reversal had a median of 4.6 points, against 5.0 otherwise.
574 sessions with a positive stock–bond correlation, 174 with a negative one. A description of the record, not a signal.
Next: Dealer positioning — who holds the other side
Dearest protection: Silver (SLV); cheapest: Brazil (EWZ)
Which markets' skew is unusually dear or cheap, and how does that sit against their and their skew slope?Exhibit 1 · Each market's skew level and implied volatility, both against three years
19 of 68 markets drawn; the other 49 are withheld or too short to fit, each named in the notes.
About this exhibit
Exhibit 2 · Each market's skew level and skew slope, both against three years
18 of 68 markets drawn; the other 50 are withheld or too short to fit, each named in the notes.
About this exhibit
19 of 68 markets drawn against , 17 against slope
How unusual is the level, slope and curvature of each market's skew, and implied volatility, against its own last three years?Exhibit 3 · Each market's chosen curve: level, slope and curvature against three years
| Market | Protective wing | Risk reversal, vol points | Skew, % of ATM | ATM volatility, vol points | Skew level, z | Skew slope, z | Skew curvature, z | Skew level z, 21-session change |
|---|---|---|---|---|---|---|---|---|
| US equity | ||||||||
| S&P 500 SPY | puts | +7.0 | +36.5 | 19.3 | +0.50 74th pct | +0.88 80th pct | −0.37 38th pct | +1.27 |
| Nasdaq-100 QQQ | puts | +9.6 | +31.0 | 31.0 | +0.29 58th pct | −0.39 37th pct | −0.10 46th pct | +0.35 |
| Russell 2000 IWM | puts | +6.9 | +31.5 | 21.7 | +1.48 97th pct | −0.40 33rd pct | −1.66 3rd pct | +1.17 |
| S&P 500 equal weight RSP · not on the maps | puts | +6.6 | +40.2 | 16.4 | withheld | withheld | withheld | – |
| Rates | ||||||||
| Long Treasuries TLT | no protective wing ρ +0.10 | −0.1 | −1.2 | 12.5 | −0.64 26th pct | −1.30 13th pct | −2.45 2nd pct | −0.58 |
| Intermediate Treasuries IEF · not on the maps | no protective wing ρ +0.02 | −0.3 | −3.6 | 7.3 | withheld | withheld | withheld | – |
| TIPS TIP · not on the maps | no protective wing ρ −0.03 | −16.2 | −127.0 | 12.8 | withheld | withheld | withheld | – |
| Commodities | ||||||||
| Gold GLD | puts ρ +0.39 | +1.8 | +7.5 | 24.6 | +1.92 85th pct | −0.24 40th pct | −0.49 30th pct | +1.55 |
| Silver SLV | puts ρ +0.52 | +5.1 | +11.2 | 45.0 | +2.84 100th pct | −1.03 19th pct | −1.51 6th pct | +2.02 |
| WTI crude USO | calls ρ −0.36 | −0.9 | −1.9 | 45.4 | −0.89 26th pct | −0.91 22nd pct | −0.53 27th pct | −0.02 |
| US natural gas UNG · not on the maps | puts ρ +0.22 | −12.9 | −19.4 | 66.2 | withheld | withheld | withheld | – |
| Copper CPER · not on the maps | puts ρ +0.56 | −0.2 | −0.7 | 28.8 | withheld | withheld | withheld | – |
| FX | ||||||||
| US dollar UUP · not on the maps | no protective wing ρ −0.16 | +33.8 | +140.9 | 24.0 | withheld | withheld | withheld | – |
| Euro FXE · not on the maps | puts ρ +0.21 | −5.4 | −67.1 | 8.1 | withheld | withheld | withheld | – |
| Yen FXY · not on the maps | no protective wing ρ +0.01 | −2.1 | −24.8 | 8.5 | withheld | withheld | withheld | – |
| Credit | ||||||||
| High yield JNK | no protective wing | −20.9 | −220.6 | 9.5 | −2.27 7th pct | +0.06 52nd pct | −0.69 24th pct | −0.20 |
| Crypto | ||||||||
| Bitcoin IBIT · not on the maps | no protective wing | +7.6 | +17.1 | 44.5 | – | – | – | – |
| Sectors | ||||||||
| Technology XLK · not on the maps | puts | +8.0 | +21.7 | 36.8 | withheld | withheld | withheld | – |
| Communication services XLC · not on the maps | puts | −0.1 | −0.4 | 29.8 | withheld | withheld | withheld | – |
| Consumer discretionary XLY · not on the maps | puts | +22.6 | +82.9 | 27.2 | withheld | withheld | withheld | – |
| Consumer staples XLP · not on the maps | puts | +3.4 | +16.4 | 21.0 | withheld | withheld | withheld | – |
| Health care XLV · not on the maps | puts | +5.0 | +23.2 | 21.6 | withheld | withheld | withheld | – |
| Financials XLF · not on the maps | puts | +11.3 | +52.4 | 21.6 | withheld | withheld | withheld | – |
| Industrials XLI · not on the maps | puts | +4.7 | +20.9 | 22.3 | withheld | withheld | withheld | – |
| Energy XLE · not on the maps | puts | +0.5 | +1.8 | 28.3 | withheld | withheld | withheld | – |
| Materials XLB · not on the maps | puts | +12.4 | +22.7 | 54.4 | withheld | withheld | withheld | – |
| Utilities XLU · not on the maps | puts | +30.3 | +121.4 | 25.0 | withheld | withheld | withheld | – |
| Real estate XLRE · not on the maps | puts | +35.7 | +62.5 | 57.2 | withheld | withheld | withheld | – |
| Industries | ||||||||
| Semiconductors SMH | puts | +10.3 | +18.1 | 56.8 | +0.26 64th pct | +1.21 83rd pct | −1.94 9th pct | +0.48 |
| Semiconductors (SOXX) SOXX | puts | +4.7 | +7.6 | 61.3 | +0.34 63rd pct | −0.58 27th pct | −2.68 6th pct | +0.41 |
| Software IGV · not on the slope map | puts | +2.8 | +8.2 | 33.7 | −0.87 25th pct | withheld | withheld | +0.46 |
| Oil & gas producers XOP | puts | −0.5 | −1.5 | 33.0 | −0.12 43rd pct | −0.13 45th pct | −1.08 16th pct | +0.38 |
| Regional banks KRE · not on the maps | puts | +0.7 | +2.4 | 28.7 | withheld | withheld | withheld | – |
| Homebuilders XHB · not on the maps | puts | +17.5 | +53.3 | 32.9 | withheld | withheld | withheld | – |
| Retail XRT · not on the maps | puts | +6.8 | +24.1 | 28.2 | withheld | withheld | withheld | – |
| Biotech XBI | puts | +3.0 | +8.0 | 38.1 | −0.98 12th pct | +0.71 76th pct | +0.88 85th pct | +0.25 |
| Aerospace & defence ITA · not on the maps | puts | −24.9 | −79.1 | 31.5 | withheld | withheld | withheld | – |
| Metals & mining XME · not on the maps | puts | −16.9 | −34.4 | 49.2 | withheld | withheld | withheld | – |
| Insurance KIE | puts | −5.6 | −24.0 | 23.4 | +1.21 90th pct | −1.12 13th pct | +2.34 97th pct | +4.05 |
| Countries | ||||||||
| Japan EWJ · not on the maps | puts | +1.4 | +5.8 | 24.9 | – | – | – | – |
| Germany EWG · not on the maps | puts | −1.0 | −5.4 | 18.2 | withheld | withheld | withheld | – |
| United Kingdom EWU · not on the maps | puts | +11.1 | +38.4 | 29.0 | withheld | withheld | withheld | – |
| France EWQ · not on the maps | puts | −12.1 | −48.7 | 24.8 | withheld | withheld | withheld | – |
| Italy EWI · not on the maps | puts | +1.1 | +4.7 | 23.8 | withheld | withheld | withheld | – |
| Spain EWP · not on the maps | puts | −5.5 | −18.4 | 29.8 | withheld | withheld | withheld | – |
| Switzerland EWL · not on the maps | puts | +39.8 | +214.0 | 18.6 | withheld | withheld | withheld | – |
| Netherlands EWN · not on the maps | puts | +32.9 | +73.1 | 45.0 | withheld | withheld | withheld | – |
| Sweden EWD | puts | +6.0 | +22.5 | 26.7 | −0.22 39th pct | +1.25 96th pct | +0.18 58th pct | +0.66 |
| Taiwan EWT · not on the maps | puts | +2.7 | +5.3 | 50.4 | withheld | withheld | withheld | – |
| South Korea EWY | puts | +12.1 | +19.3 | 62.5 | +2.01 94th pct | +1.48 88th pct | −1.03 14th pct | +1.35 |
| Hong Kong EWH · not on the maps | puts | −5.6 | −20.9 | 26.8 | withheld | withheld | withheld | – |
| Australia EWA · not on the maps | puts | −5.0 | −18.2 | 27.7 | withheld | withheld | withheld | – |
| Canada EWC · not on the maps | puts | −5.4 | −23.7 | 22.6 | withheld | withheld | withheld | – |
| Mexico EWW · not on the slope map | puts | +2.1 | +9.1 | 23.4 | ±0.00 50th pct | withheld | withheld | +0.73 |
| Brazil EWZ | puts | −39.5 | −120.9 | 32.7 | −1.49 12th pct | +0.17 56th pct | −1.21 13th pct | −1.93 |
| China large-cap FXI | puts | −5.5 | −22.0 | 25.1 | −1.26 21st pct | +2.35 99th pct | withheld | +4.25 |
| India INDA · not on the maps | puts | −2.0 | −9.5 | 20.9 | withheld | withheld | withheld | – |
| Saudi Arabia KSA · not on the maps | puts | +66.6 | +106.9 | 62.2 | withheld | withheld | withheld | – |
| South Africa EZA · not on the maps | puts | +11.2 | +23.1 | 48.4 | withheld | withheld | withheld | – |
| Poland EPOL · not on the maps | puts | −5.5 | −10.7 | 51.6 | – | – | – | – |
| Thailand THD | puts | −19.0 | −60.5 | 31.4 | −0.80 15th pct | +0.89 82nd pct | +0.03 52nd pct | −0.62 |
| Indonesia EIDO · not on the maps | puts | −97.2 | −49.5 | 196.5 | withheld | withheld | withheld | – |
| Malaysia EWM · not on the maps | puts | −1.6 | −6.3 | 25.6 | withheld | withheld | withheld | – |
| Regions | ||||||||
| Eurozone EZU · not on the maps | puts | +10.3 | +43.4 | 23.6 | withheld | withheld | withheld | – |
| Europe IEUR · not on the maps | puts | – | – | – | withheld | withheld | withheld | – |
| Asia ex-Japan AAXJ · not on the maps | puts | −7.6 | −21.8 | 35.1 | withheld | withheld | withheld | – |
| Emerging markets ex-China EMXC · not on the maps | puts | −14.5 | −35.8 | 40.5 | withheld | withheld | withheld | – |
| World ex-US VXUS · not on the maps | puts | −2.1 | −9.8 | 21.6 | withheld | withheld | withheld | – |
About this exhibit
Exhibit 4 · S&P 500 risk reversal and skew over time, with at-the-money volatility
About this exhibit
Exhibit 5 · The selected market's curve over time in level, slope and curvature
About this exhibit
Exhibit 6 · S&P 500: weight of each tenor in skew level, slope and curvature
Skew
Variance explained 89% · 5% · 4%; matches a fixed level, slope and curvature shape at 1.00 / 0.93 / 0.78; the fit of three years ago at 1.00 / 1.00 / 1.00. 131 fits since 4 Jan 2016.
Risk reversal
Variance explained 93% · 4% · 2%; matches a fixed level, slope and curvature shape at 1.00 / 0.98 / 0.68; the fit of three years ago at 1.00 / 1.00 / 1.00. 131 fits since 4 Jan 2016.
Implied volatility
Variance explained 98% · 2% · 0%; matches a fixed level, slope and curvature shape at 1.00 / 0.96 / 0.73; the fit of three years ago at 1.00 / 1.00 / 1.00. 131 fits since 4 Jan 2016.
About this exhibit
What the components add
How far does the skew level differ from the 25-day skew alone, and how much noise does the five-session mean remove?Exhibit 7 · Each market's 25-day skew against its skew level, both against three years
| Market | 25-day skew alone, z | Skew level, z | Skew slope, z | Correlation |
|---|---|---|---|---|
| US equity | ||||
| S&P 500 SPY | −0.02 | +0.50 | +0.88 | 0.94 |
| Nasdaq-100 QQQ | +0.36 | +0.29 | −0.39 | 0.90 |
| Russell 2000 IWM | +1.51 | +1.48 | −0.40 | 0.89 |
| S&P 500 equal weight RSP | +1.98 | withheld | withheld | – |
| Rates | ||||
| Long Treasuries TLT | −0.34 | −0.64 | −1.30 | 0.59 |
| Intermediate Treasuries IEF | +0.26 | withheld | withheld | – |
| TIPS TIP | −1.07 | withheld | withheld | – |
| Commodities | ||||
| Gold GLD | +1.63 | +1.92 | −0.24 | 0.88 |
| Silver SLV | +2.76 | +2.84 | −1.03 | 0.84 |
| WTI crude USO | −0.01 | −0.89 | −0.91 | 0.30 |
| US natural gas UNG | +1.30 | withheld | withheld | – |
| Copper CPER | −0.51 | withheld | withheld | – |
| FX | ||||
| US dollar UUP | +0.77 | withheld | withheld | – |
| Euro FXE | −2.27 | withheld | withheld | – |
| Yen FXY | +0.09 | withheld | withheld | – |
| Credit | ||||
| High yield JNK | −3.08 | −2.27 | +0.06 | 0.55 |
| Sectors | ||||
| Technology XLK | −2.40 | withheld | withheld | – |
| Communication services XLC | −1.37 | withheld | withheld | – |
| Consumer discretionary XLY | −0.01 | withheld | withheld | – |
| Consumer staples XLP | −0.11 | withheld | withheld | – |
| Health care XLV | +0.82 | withheld | withheld | – |
| Financials XLF | +0.59 | withheld | withheld | – |
| Industrials XLI | −0.19 | withheld | withheld | – |
| Energy XLE | −1.15 | withheld | withheld | – |
| Materials XLB | +0.75 | withheld | withheld | – |
| Utilities XLU | +3.43 | withheld | withheld | – |
| Real estate XLRE | +0.71 | withheld | withheld | – |
| Industries | ||||
| Semiconductors SMH | −0.53 | +0.26 | +1.21 | 0.80 |
| Semiconductors (SOXX) SOXX | −1.76 | +0.34 | −0.58 | 0.56 |
| Software IGV | −0.40 | −0.87 | withheld | 0.59 |
| Oil & gas producers XOP | −0.28 | −0.12 | −0.13 | 0.72 |
| Regional banks KRE | +1.15 | withheld | withheld | – |
| Homebuilders XHB | −1.33 | withheld | withheld | – |
| Retail XRT | +2.09 | withheld | withheld | – |
| Biotech XBI | −0.82 | −0.98 | +0.71 | 0.66 |
| Aerospace & defence ITA | −2.78 | withheld | withheld | – |
| Metals & mining XME | −0.34 | withheld | withheld | – |
| Insurance KIE | +0.08 | +1.21 | −1.12 | 0.66 |
| Countries | ||||
| Germany EWG | −1.61 | withheld | withheld | – |
| United Kingdom EWU | +0.50 | withheld | withheld | – |
| France EWQ | −0.61 | withheld | withheld | – |
| Italy EWI | −0.48 | withheld | withheld | – |
| Spain EWP | −0.45 | withheld | withheld | – |
| Switzerland EWL | +0.98 | withheld | withheld | – |
| Netherlands EWN | −0.67 | withheld | withheld | – |
| Sweden EWD | −0.02 | −0.22 | +1.25 | 0.80 |
| Taiwan EWT | −0.41 | withheld | withheld | – |
| South Korea EWY | +0.44 | +2.01 | +1.48 | 0.70 |
| Hong Kong EWH | +0.25 | withheld | withheld | – |
| Australia EWA | +0.23 | withheld | withheld | – |
| Canada EWC | +0.32 | withheld | withheld | – |
| Mexico EWW | +0.09 | ±0.00 | withheld | 0.60 |
| Brazil EWZ | −1.61 | −1.49 | +0.17 | 0.77 |
| China large-cap FXI | −2.26 | −1.26 | +2.35 | 0.76 |
| India INDA | +0.64 | withheld | withheld | – |
| Saudi Arabia KSA | +1.78 | withheld | withheld | – |
| South Africa EZA | −0.35 | withheld | withheld | – |
| Thailand THD | −0.79 | −0.80 | +0.89 | 0.73 |
| Indonesia EIDO | −0.41 | withheld | withheld | – |
| Malaysia EWM | −0.33 | withheld | withheld | – |
| Regions | ||||
| Eurozone EZU | +1.13 | withheld | withheld | – |
| Europe IEUR | – | withheld | withheld | – |
| Asia ex-Japan AAXJ | −1.11 | withheld | withheld | – |
| Emerging markets ex-China EMXC | −0.96 | withheld | withheld | – |
| World ex-US VXUS | −0.77 | withheld | withheld | – |
About this exhibit
Exhibit 8 · How much each market's scores swing week to week, raw and averaged
| Market | Skew level | Skew slope | Skew curvature | Implied volatility level |
|---|---|---|---|---|
| US equity | ||||
| S&P 500 SPY | 0.61 → 0.51 | 0.94 → 0.60 | 1.10 → 0.82 | 0.68 → 0.63 |
| Nasdaq-100 QQQ | 0.69 → 0.55 | 0.85 → 0.62 | 0.95 → 0.79 | 0.72 → 0.64 |
| Russell 2000 IWM | 0.78 → 0.58 | 0.82 → 0.60 | 0.77 → 0.64 | 0.76 → 0.69 |
| S&P 500 equal weight RSP | 0.91 → 0.77 | 1.45 → 1.10 | 1.66 → 1.59 | 0.77 → 0.74 |
| Rates | ||||
| Long Treasuries TLT | 0.86 → 0.72 | 1.48 → 1.12 | 1.95 → 1.24 | 0.72 → 0.73 |
| Intermediate Treasuries IEF | 2.34 → 1.05 | 2.01 → 1.39 | 2.77 → 1.42 | 0.63 → 0.45 |
| TIPS TIP | 1.11 → 0.76 | 1.91 → 1.23 | 1.53 → 1.10 | 0.90 → 0.49 |
| Commodities | ||||
| Gold GLD | 0.71 → 0.55 | 1.10 → 0.89 | 1.52 → 1.08 | 0.72 → 0.69 |
| Silver SLV | 0.71 → 0.53 | 1.23 → 0.86 | 1.17 → 0.96 | 1.01 → 0.78 |
| WTI crude USO | 0.73 → 0.63 | 1.11 → 0.81 | 1.54 → 1.54 | 0.90 → 0.76 |
| US natural gas UNG | 1.37 → 0.87 | 1.55 → 1.14 | 2.56 → 1.90 | 0.46 → 0.35 |
| Copper CPER | 2.31 → 2.37 | 3.03 → 2.18 | 1.80 → 1.04 | 0.42 → 0.29 |
| FX | ||||
| US dollar UUP | 1.32 → 1.09 | 1.68 → 1.41 | 1.61 → 1.52 | 1.21 → 0.67 |
| Euro FXE | 1.12 → 0.62 | 3.05 → 1.54 | 2.60 → 2.02 | 0.61 → 0.51 |
| Yen FXY | 1.90 → 1.38 | 1.88 → 1.38 | 2.18 → 1.72 | 0.67 → 0.58 |
| Credit | ||||
| High yield JNK | 0.83 → 0.60 | 1.13 → 0.73 | 2.23 → 1.62 | 0.69 → 0.59 |
| Sectors | ||||
| Technology XLK | 1.15 → 0.71 | 1.56 → 1.05 | 2.01 → 1.38 | 0.64 → 0.59 |
| Communication services XLC | 0.96 → 0.62 | 1.04 → 0.72 | 1.23 → 0.86 | 0.44 → 0.26 |
| Consumer discretionary XLY | 1.08 → 0.76 | 2.16 → 1.39 | 2.35 → 1.63 | 0.61 → 0.56 |
| Consumer staples XLP | 0.97 → 0.72 | 1.76 → 1.18 | 1.86 → 1.48 | 0.83 → 0.58 |
| Health care XLV | 1.19 → 0.80 | 1.93 → 1.17 | 2.05 → 1.27 | 0.75 → 0.69 |
| Financials XLF | 1.14 → 0.75 | 1.66 → 1.09 | 2.77 → 1.84 | 1.54 → 0.99 |
| Industrials XLI | 1.16 → 0.77 | 1.77 → 0.96 | 2.54 → 1.49 | 0.65 → 0.59 |
| Energy XLE | 1.23 → 0.77 | 1.96 → 1.31 | 2.61 → 1.95 | 0.41 → 0.30 |
| Materials XLB | 1.24 → 0.67 | 1.52 → 0.98 | 1.75 → 1.36 | 0.74 → 0.68 |
| Utilities XLU | 1.10 → 0.86 | 1.55 → 1.14 | 1.98 → 1.35 | 1.09 → 0.83 |
| Real estate XLRE | 1.79 → 1.26 | 2.68 → 1.95 | 2.38 → 1.79 | 1.51 → 1.28 |
| Industries | ||||
| Semiconductors SMH | 0.47 → 0.41 | 0.79 → 0.66 | 1.09 → 0.80 | 0.63 → 0.56 |
| Semiconductors (SOXX) SOXX | 0.78 → 0.48 | 1.44 → 1.02 | 1.49 → 1.27 | 0.77 → 0.74 |
| Software IGV | 0.79 → 0.52 | 1.35 → 1.12 | 1.40 → 0.96 | 0.63 → 0.66 |
| Oil & gas producers XOP | 0.64 → 0.48 | 1.74 → 1.47 | 1.47 → 1.31 | 0.75 → 0.85 |
| Regional banks KRE | 1.29 → 0.87 | 1.48 → 1.15 | 1.97 → 1.52 | 1.15 → 0.94 |
| Homebuilders XHB | 1.02 → 0.74 | 1.65 → 1.12 | 2.16 → 1.55 | 0.67 → 0.56 |
| Retail XRT | 1.22 → 0.75 | 2.18 → 1.42 | 3.49 → 1.98 | 1.05 → 0.95 |
| Biotech XBI | 0.74 → 0.50 | 1.86 → 1.46 | 1.73 → 1.28 | 0.50 → 0.41 |
| Aerospace & defence ITA | 0.90 → 0.64 | 1.20 → 0.91 | 1.49 → 1.01 | 0.33 → 0.27 |
| Metals & mining XME | 1.34 → 0.93 | 2.55 → 1.69 | 2.49 → 1.77 | 0.77 → 0.65 |
| Insurance KIE | 0.88 → 0.66 | 1.23 → 0.96 | 1.24 → 0.92 | 1.11 → 0.87 |
| Countries | ||||
| Germany EWG | 1.27 → 0.81 | 1.93 → 1.17 | 2.01 → 1.34 | 0.73 → 0.64 |
| United Kingdom EWU | 1.39 → 0.89 | 1.39 → 1.14 | 1.67 → 1.20 | 0.99 → 0.67 |
| France EWQ | 1.33 → 0.73 | 1.72 → 1.33 | 1.47 → 1.13 | 0.91 → 0.88 |
| Italy EWI | 0.99 → 0.69 | 1.05 → 0.92 | 1.47 → 1.08 | 0.59 → 0.45 |
| Spain EWP | 1.19 → 0.81 | 1.81 → 1.40 | 2.01 → 1.52 | 1.06 → 0.93 |
| Switzerland EWL | 1.14 → 0.87 | 1.52 → 1.18 | 1.50 → 1.15 | 1.14 → 0.80 |
| Netherlands EWN | 0.91 → 0.65 | 1.22 → 1.08 | 1.30 → 1.04 | 1.56 → 1.02 |
| Sweden EWD | 0.91 → 0.63 | 0.98 → 0.72 | 1.47 → 1.12 | 0.66 → 0.54 |
| Taiwan EWT | 0.89 → 0.56 | 1.62 → 1.16 | 1.63 → 1.17 | 0.72 → 0.64 |
| South Korea EWY | 1.34 → 0.91 | 1.95 → 1.49 | 1.56 → 1.10 | 0.95 → 0.89 |
| Hong Kong EWH | 1.40 → 1.05 | 1.46 → 0.98 | 1.47 → 1.35 | 0.74 → 0.45 |
| Australia EWA | 1.06 → 0.81 | 1.88 → 1.39 | 1.72 → 1.13 | 1.12 → 0.92 |
| Canada EWC | 0.97 → 0.64 | 1.53 → 1.28 | 1.86 → 1.25 | 1.15 → 0.90 |
| Mexico EWW | 1.12 → 0.78 | 1.65 → 1.30 | 1.98 → 1.27 | 0.81 → 0.72 |
| Brazil EWZ | 0.94 → 0.72 | 1.24 → 1.14 | 1.57 → 1.19 | 1.08 → 0.73 |
| China large-cap FXI | 1.26 → 0.78 | 1.50 → 0.97 | 1.75 → 1.20 | 0.70 → 0.55 |
| India INDA | 1.27 → 0.93 | 2.03 → 1.35 | 1.42 → 1.23 | 0.56 → 0.44 |
| Saudi Arabia KSA | 1.15 → 0.79 | 0.95 → 0.79 | 0.96 → 0.62 | 0.41 → 0.29 |
| South Africa EZA | 1.16 → 0.80 | 1.56 → 1.44 | 1.33 → 1.15 | 0.86 → 0.74 |
| Thailand THD | 0.96 → 0.69 | 1.32 → 1.31 | 1.65 → 1.24 | 0.82 → 0.71 |
| Indonesia EIDO | 0.72 → 0.61 | 1.55 → 1.24 | 1.28 → 1.03 | 0.85 → 0.67 |
| Malaysia EWM | 1.20 → 0.71 | 1.40 → 1.03 | 1.45 → 1.18 | 1.58 → 1.02 |
| Regions | ||||
| Eurozone EZU | 0.99 → 0.65 | 1.49 → 1.15 | 1.56 → 1.36 | 0.69 → 0.75 |
| Europe IEUR | 1.01 → 0.77 | 1.30 → 1.34 | 1.38 → 1.02 | 0.39 → 0.35 |
| Asia ex-Japan AAXJ | 1.26 → 0.83 | 1.73 → 1.63 | 2.19 → 1.63 | 0.64 → 0.61 |
| Emerging markets ex-China EMXC | 0.27 → 0.17 | 0.62 → 0.68 | 0.82 → 0.57 | 0.36 → 0.27 |
| World ex-US VXUS | 1.52 → 0.96 | 1.49 → 0.97 | 1.70 → 1.34 | 0.70 → 0.59 |
About this exhibit
Median 4.6 with stocks and bonds together, 5.0 apart
What does a risk reversal cost, and has skew been richer when stocks and bonds move together?Exhibit 9 · S&P 500 risk reversal cost by days to expiry
No premium at 250 days: no listed expiry lies within 5 days of that tenor on 16 Nov 2026.
About this exhibit
Exhibit 10 · S&P 500 risk reversal when stocks and bonds move together or apart
Together: median 4.6 vol points over 574 sessions. Apart: median 5.0 vol points over 174 sessions. Neighbouring sessions share a week's state, so the counts are not independent draws; a description of the record, not a signal.
About this exhibit
The reference set in
Is each reference market's rich or cheap against its three-year median, tenor by tenor?
Exhibit 11 · Each market's risk reversal against its three-year median, by tenor
About this exhibit
Appendix
How the , the skew, the components and the maps are made, and what each check holds66 checks, all pass
Which tail of each market the options price as dearer, and how that changes with the horizon: for 68 markets, the skew level — what each one's protective wing costs across the 25- to 125-day curve (puts where the market falls with stocks, calls where it rises when they fall), against its own last three years — mapped against its implied-volatility level and against its skew slope; the risk reversal — the implied volatility of a put less that of a call the same distance from the money — of the reference set against its own median; what a risk reversal costs to put on across the bid/ask; and whether skew has been richer in the weeks stocks and bonds moved together. Observed prices, and a description of the record — not a signal.
Risk reversal and butterfly
| Symbol | Meaning | From | Assumed or estimated |
|---|---|---|---|
| the surface's implied volatility at signed delta Δ | the options engine | – |
Observed: each is a difference of quotes on the session.
The components, the statistics and the maps
| Symbol | Meaning | From | Assumed or estimated |
|---|---|---|---|
| the measure at tenor j (25, 50, 75, 125 days), after the engine's causal clean | the options engine | – | |
| , | its mean and standard deviation over its history to t | the options engine | – |
| the loadings of the last fit r before t | the options engine | estimated | |
| , | median and median absolute deviation of m_k over the trailing 756 sessions | computed here | – |
| +1 when the market's puts protect, −1 when its calls do | computed here | – | |
| the percentile of the at-the-money volatility level component | computed here | – |
Observed prices; the components are fitted, and refitted as the history grows.
Observed prices, no model of the future. RR25 = σ(−25Δ) − σ(+25Δ); BF25 = ½(σ(−25Δ) + σ(+25Δ)) − σ(ATM); the skew is RR25 ÷ σ(ATM) × 100. No level, slope or curvature scores are drawn: None.
The components are the options engine's own: for each market, the 25-delta risk reversal, the skew (the risk reversal as a share of at-the-money volatility) and the at-the-money volatility at 25, 50, 75 and 125 days, each standardised on its own history to date and resolved into a level, a slope and a curvature by principal components refitted every 21 sessions after the first 252. Each fit scores only the sessions after it, so no score uses a later price. Every score is recomputed here from the inputs, the standardisation and the loadings shipped with it (the author's decision of 7 Oct 2026 (D-2026-10-07-SKEW1)).
Each component is averaged over its last 5 sessions; the z-score is that mean's distance from its median over the trailing 756 sessions in units of 1.4826 × the median absolute deviation, and the percentile its rank in the same window, both printed from 252 sessions (the author's decision of 7 Oct 2026 (D-2026-10-07-SKEW2)). The raw score's z is drawn beside it, faint.
The skew level on the maps is its z with the protective wing's sign: equity markets' puts protect; another market's puts protect when its 63-session correlation with the S&P 500 is +0.2 or more, its calls at −0.2 or less, and it is drawn hollow between (the author's decision of 7 Oct 2026 (D-2026-10-07-SKEW3)). The maps describe prices; they are not signals.
19 of 68 markets are drawn on the implied volatility map, 18 on the slope map. A market is withheld when its skew's first component explains less than 60% of the variation at the latest fit, when its loadings now and three years earlier agree at less than 0.9, or when its term structure stops more than five sessions before the session (the author's decision of 7 Oct 2026 (D-2026-10-07-SKEW4)). The slope map also needs the slope component, held to the same three-year test. Each market not drawn is named in the notes.
No session is excluded: on every session the S&P 500's 25-day at-the-money volatility stayed within half its usual ratio to a published index of the S&P 500's 30-day implied volatility (the author's decision of 7 Oct 2026 (D-2026-10-07-SKEW5)). The index is used for this check only and is not printed.
Premiums are fractions of spot on the session as the options engine prices them: each leg priced from that side's contracts expiring within 5 calendar days of the tenor — the ask for a long leg and the bid for a short one, a missing side falling back to the mid and then to the other side — averaged with Gaussian weights in |delta| (bandwidth 0.075) around the leg's delta and divided by spot; a tenor with no such expiry is not priced; the legs' sum then corrected by a centred five-session Hampel filter (a value more than 5 scaled median absolute deviations from the rolling median replaced by it) and smoothed by a trailing three-session median. A premium is therefore what the position would have cost across the bid/ask, steadied over the last three sessions — not a mid-market value, and not one session's quote alone (the author's decision of 4 Oct 2026 (D-2026-10-04-08)).
The sign-rule state is the earlier Correlations & regime readings' weekly reading of the signs of the 63-day stock–bond and stock–oil correlations, each state named by its legs; it is not the site's regime, which is the four-regime map. Each session takes the latest weekly reading on or before it. The cross-check groups the sessions by the sign of the stock–bond correlation and prints the counts; it describes the record.
The figures on this page come from a kernel-smoothed implied-volatility surface on a 39 × 7 delta–tenor grid, computed here: for each tenor and each delta the implied volatilities of the listed contracts are averaged with Gaussian weights in delta and in days to expiry. It is not presented as equivalent to any commercial surface. The wording follows the author's decision of 3 Oct 2026 (D-2026-10-03-21).
| Setting | Value |
|---|---|
| Delta grid | −0.95 to +0.95 in steps of 0.05 (39 points) |
| Tenors | 5, 10, 25, 50, 75, 125, 250 calendar days to expiry |
| Bandwidth in delta (σ) | 0.1 |
| Bandwidth in days to expiry (σ) | 25 |
| Contracts within, days | 25 |
| Fewest contracts for a point | 10 |
| Weight | exp(-((delta - target_delta) / sigma_delta)^2) * exp(-((dte - target_dte) / sigma_days)^2) for contracts with |dte - target_dte| <= gate_days |
| Side rule | puts for a negative target delta, calls for a positive one, both at zero |
| At the money | the mean of the -0.50 and +0.50 delta points |
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 365 on trading days, overnight return included. The quotes are the vendor's end-of-day set for the session, stamped 16:00 New York, 16 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 ──▶ RR25, BF25 and the skew recomputed from the wing columns; the level, slope and curvature scores from the stated basis; each tenor's trailing median and percentile ──▶ skew by asset and tenor sign_rule + skew by asset and tenor ──▶ each session takes the latest weekly sign-rule state on or before it; RR25 grouped by the sign of the stock–bond correlation, with counts ──▶ the cross-check the verified folder ──▶ every component score recomputed from its row's inputs, moments and loadings; the five-session mean; robust z and percentile over the trailing 756 sessions ──▶ each market's components closes + vix + each market's components ──▶ corrupt sessions excluded by the VIX check; the protective wing from the 63-session correlation with the S&P 500; markets withheld by the explained-variance and stability rules ──▶ the two maps and the board
Sources
- licensedOptions engine session folder: end-of-day option chains, processed into a kernel-smoothed implied-volatility surface
marketdata.app
- computed hereThe Correlations & regime page's weekly sign-rule state, copied as served and compared with it
Computed here (the sign-rule state of the Correlations & regime page).
- computed hereThe Performance page's daily closes, copied as served: each market's 63-session correlation with the S&P 500 (the protective wing)
Computed here (closes from the Performance page).
- licensedCboe Global Indices: the VIX index's daily close, used only to find sessions whose option quotes are corrupt; not printed
Source: Cboe Global Indices, LLC · https://cdn.cboe.com/api/global/us_indices/daily_prices/VIX_History.csv
Pages this one copies from
| Page | Title | Updated | Data to | Arrays | Values | Source checks |
|---|---|---|---|---|---|---|
| Correlation regime | Correlation regime | 2026-11-17 | 13 Nov 2026 | 0 | 7 | 80 pass |
| Performance | Cross-asset scoreboard | 2026-11-17 | 16 Nov 2026 | 12 | 0 | 23 pass |
Reproduction checks
| Check | Published | Computed | Tolerance | Result |
|---|---|---|---|---|
| Session folder 2026-11-16 · _SUCCESS: present: the engine writes it last, and a folder without it is not read | 1 | 1 | 0 | pass |
| Session folder 2026-11-16 · summary_history.parquet: present: a file the contract names | 1 | 1 | 0 | pass |
| Session folder 2026-11-16 · surface_latest.parquet: present: a file the contract names | 1 | 1 | 0 | pass |
| Session folder 2026-11-16 · dealer_history.parquet: present: a file the contract names | 1 | 1 | 0 | pass |
| Session folder 2026-11-16 · combos.parquet: present: a file the contract names | 1 | 1 | 0 | pass |
| Session folder 2026-11-16 · sample_raw.parquet: present: a file the contract names | 1 | 1 | 0 | pass |
| Session folder 2026-11-16 · README.md: SHA-256 equals the hash checks.json lists for it | 1 | 1 | 0 | pass |
| Session folder 2026-11-16 · combos.parquet: SHA-256 equals the hash checks.json lists for it | 1 | 1 | 0 | pass |
| Session folder 2026-11-16 · dealer_history.parquet: SHA-256 equals the hash checks.json lists for it | 1 | 1 | 0 | pass |
| Session folder 2026-11-16 · pca_refits.parquet: SHA-256 equals the hash checks.json lists for it | 1 | 1 | 0 | pass |
| Session folder 2026-11-16 · pca_rows.parquet: SHA-256 equals the hash checks.json lists for it | 1 | 1 | 0 | pass |
| Session folder 2026-11-16 · sample_raw.parquet: SHA-256 equals the hash checks.json lists for it | 1 | 1 | 0 | pass |
| Session folder 2026-11-16 · skew_history.parquet: SHA-256 equals the hash checks.json lists for it | 1 | 1 | 0 | pass |
| Session folder 2026-11-16 · summary_history.parquet: SHA-256 equals the hash checks.json lists for it | 1 | 1 | 0 | pass |
| Session folder 2026-11-16 · surface_latest.parquet: SHA-256 equals the hash checks.json lists for it | 1 | 1 | 0 | pass |
| Session folder 2026-11-16 · checks.json: session_date is the folder's name | 1 | 1 | 0 | pass |
| Session folder 2026-11-16 · checks.json: session_date is a weekday (a US session) | 1 | 1 | 0 | pass |
| Session folder 2026-11-16 · checks.json: session_date is on or before the run day 2026-11-17 | 1 | 1 | 0 | pass |
| Session folder 2026-11-16 · 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 |
| Session folder 2026-11-16 · 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 |
| Session folder 2026-11-16 · checks.json: written_utc is not after the moment of the check | 1 | 1 | 0 | pass |
| Session folder 2026-11-16 · checks.json: quote_time_utc is not after the moment of the check | 1 | 1 | 0 | pass |
| Session folder 2026-11-16 · checks.json: every one of the export's gate checks passed, standing in for the engine's own (10 listed) | 1 | 1 | 0 | pass |
| Contract: the session folder states iv units as the site's contract does | decimal | decimal | 0 | pass |
| Contract: the session folder states dte as the site's contract does | calendar days to expiry; tau = dte / 365 | calendar days to expiry; tau = dte / 365 | 0 | pass |
| Contract: the session folder states rr25 as the site's contract does | iv_put_25 - iv_call_25 | iv_put_25 - iv_call_25 | 0 | pass |
| Contract: the session folder states bf25 as the site's contract does | (iv_put_25 + iv_call_25) / 2 - atm_iv | (iv_put_25 + iv_call_25) / 2 - atm_iv | 0 | pass |
| Contract: the session folder states iv skew pct as the site's contract does | (iv_put_25 - iv_call_25) / atm_iv * 100 | (iv_put_25 - iv_call_25) / atm_iv * 100 | 0 | pass |
| Contract: the session folder states combos as the site's contract does | {"legs": {"long": "ask, else mid, else bid", "short": "bid, else mid, else ask"}, "pricing": "each leg priced from that side's contracts expiring within 5 calendar days of the tenor — the ask for a long leg and the bid for a short one, a missing side falling back to the mid and then to the other side — averaged with Gaussian weights in |delta| (bandwidth 0.075) around the leg's delta and divided by spot; a tenor with no such expiry is not priced; the legs' sum then corrected by a centred five-session Hampel filter (a value more than 5 scaled median absolute deviations from the rolling median replaced by it) and smoothed by a trailing three-session median", "tenor_window_days": 5, "unit": "fraction of spot"} | {"legs": {"long": "ask, else mid, else bid", "short": "bid, else mid, else ask"}, "pricing": "each leg priced from that side's contracts expiring within 5 calendar days of the tenor — the ask for a long leg and the bid for a short one, a missing side falling back to the mid and then to the other side — averaged with Gaussian weights in |delta| (bandwidth 0.075) around the leg's delta and divided by spot; a tenor with no such expiry is not priced; the legs' sum then corrected by a centred five-session Hampel filter (a value more than 5 scaled median absolute deviations from the rolling median replaced by it) and smoothed by a trailing three-session median", "tenor_window_days": 5, "unit": "fraction of spot"} | 0 | pass |
| Rr25 is IV(−25Δ) − IV(+25Δ), recomputed from the shipped wing columns on every row (119330 rows; max abs difference) | 0 | 0 | 1e-9 | pass |
| Bf25 is the mean of the 25-delta wings less the at-the-money point, recomputed from the shipped wing columns on every row (119330 rows; max abs difference) | 0 | 0 | 1e-9 | pass |
| Iv_skew_pct is the risk reversal as a percentage of the at-the-money point, recomputed from the shipped wing columns on every row (119330 rows; max abs difference) | 0 | 0 | 1e-9 | pass |
| Every grid point carries an implied volatility, or is empty with no contract of its side within the gate of its tenor (39 empty; points otherwise) | 0 | 0 | 0 | pass |
| On the session, RR25 is the grid's −0.25 point less its +0.25 point at each tenor (39 symbol × tenor points, 1 left empty with no contract within the gate; max abs difference) | 0 | 0 | 1e-9 | pass |
| Contract: the session folder states pca as the site's contract does | {"basis": null, "fill": "a missing value takes the trailing 5-label mean of its tenor, then the last value for at most 5 labels; labels still incomplete are dropped (no backward fill)", "hampel": "a trailing 21-label Hampel test per tenor (threshold 3 scaled median absolute deviations); a flagged value takes the previous clean value", "labels": "the clean labels only: every contract once, every quote on the label's own New York date (as every file of this folder)", "pca": "3 components fitted on the standardised rows up to r at r = 251, 272, ... (every 21 labels from 252); each fit scores the labels after r up to and including the next fit's r", "score": "pc_k = sum_j w_kj(refit_id) * (x_j - mu_j) / sd_j", "shipped": true, "signs": "the first fit: level loadings mean-positive, slope long above short, curvature ends above the middle; every later fit sign-locked to the first", "smooth": "none: raw scores only", "standardisation": "expanding causal mean and population standard deviation per tenor, from 2 labels", "tenors": [25, 50, 75, 125], "variant": "site-skew-pca/1"} | {"basis": null, "fill": "a missing value takes the trailing 5-label mean of its tenor, then the last value for at most 5 labels; labels still incomplete are dropped (no backward fill)", "hampel": "a trailing 21-label Hampel test per tenor (threshold 3 scaled median absolute deviations); a flagged value takes the previous clean value", "labels": "the clean labels only: every contract once, every quote on the label's own New York date (as every file of this folder)", "pca": "3 components fitted on the standardised rows up to r at r = 251, 272, ... (every 21 labels from 252); each fit scores the labels after r up to and including the next fit's r", "score": "pc_k = sum_j w_kj(refit_id) * (x_j - mu_j) / sd_j", "shipped": true, "signs": "the first fit: level loadings mean-positive, slope long above short, curvature ends above the middle; every later fit sign-locked to the first", "smooth": "none: raw scores only", "standardisation": "expanding causal mean and population standard deviation per tenor, from 2 labels", "tenors": [25, 50, 75, 125], "variant": "site-skew-pca/1"} | 0 | pass |
| Medians: each trailing three-year median of RR25, recomputed by sorting its window, equals the published one (113318 values; max abs difference, vol points) | 0 | 0 | 1e-9 | pass |
| Percentiles: SPY's 25-day skew percentile, recomputed by a plain loop over its trailing window, equals the published one (2858 dates; max abs difference) | 0 | 0 | 1e-9 | pass |
| Percentile and median are causal: on SPY's history cut at 22 dates each equals the full series at that date (mismatches) | 0 | 0 | 0 | pass |
| The declared stock–bond reading of each state agrees with the regime page's own current reading (state and the sign of its 63-day stock–bond correlation; disagreements) | 0 | 0 | 0 | pass |
| The cross-check counts every SPY session with a 25-day RR25 and a state once (sessions counted twice or missed) | 0 | 0 | 0 | pass |
| Every combo leg is priced as the contract states: long legs ask, else mid, else bid, short legs bid, else mid, else ask (84 legs; legs priced otherwise or unreadable) | 0 | 0 | 0 | pass |
| SPY's premiums (risk_reversal_25) are shipped for exactly the tenors of the grid with a listed expiry within 5 days (6 of 7 tenors; tenors that differ) | 0 | 0 | 0 | pass |
| Skew history: rr25 is IV(−25Δ) − IV(+25Δ) on every row of every market (936635 rows; max abs difference) | 0 | 0 | 1e-9 | pass |
| Skew history: iv_skew_pct is the risk reversal as a percentage of at the money on every row of every market (936635 rows; max abs difference) | 0 | 0 | 1e-9 | pass |
| Components: every score is its refit's loadings applied to its row's standardised input, written out tenor by tenor (1118835 scores of 68 markets; max abs difference) | 0 | 0 | 1e-9 | pass |
| Components: each input is the shipped skew history at its label and tenor wherever that ships one (1537815 values; max abs difference) | 0 | 0 | 1e-12 | pass |
| Components: the folder states the tenors the specification draws | 255075125 | 255075125 | 0 | pass |
| The VIX history is available for the corrupt-session check (a Cboe fetch on or before the run day) | 0 | pass | ||
| The protective wing: each market's 63-session correlation with the S&P 500, recomputed by numpy on its last 63 common returns (11 markets; max abs difference) | 0 | 0 | 1e-9 | pass |
| The protective wing agrees with the Correlations & regime page: the 63-session stock–bond correlation (TLT) computed here from the Performance page's closes equals its published reading. Allowed: half a unit of the reading's third decimal, and the 0.0012 that the closes' three decimals can move it | 0.099 | 0.099052522653 | 0.001678304019 | pass |
| The protective wing agrees with the Correlations & regime page: the 63-session stock–oil correlation (USO) computed here from the Performance page's closes equals its published reading. Allowed: half a unit of the reading's third decimal, and the 0.0005 that the closes' three decimals can move it | -0.36 | -0.359647119574 | 0.001000315284 | pass |
| The five-session mean of every component equals the plain-loop mean (603 series; max abs difference) | 0 | 0 | 1e-12 | pass |
| The robust z of every component's five-session mean, recomputed over sorted windows for its last 21 sessions, equals the published one (603 series; max abs difference) | 0 | 0 | 1e-9 | pass |
| The percentile of every component's five-session mean, recomputed by a plain loop for its last 21 sessions, equals the published one (603 series; max abs difference) | 0 | 0 | 1e-9 | pass |
| The z is causal: SPY's skew-level z on its history cut at 10 dates equals the full series at that date (mismatches) | 0 | 0 | 0 | pass |
| No excluded session is drawn (0 excluded; sessions drawn on them) | 0 | 0 | 0 | pass |
| Both maps' across axis is each market's skew-level z with its protective wing's sign (markets otherwise) | 0 | 0 | 0 | pass |
| Every takeaway and tile keeps the house grammar and carries no process word (6 takeaways, 6 tiles; faults) | 0 | 0 | 0 | pass |
| The options source line carries the attribution string the session folder states | marketdata.app | marketdata.app | 0 | pass |
| Cross.regime: the source page is published with all its checks passing (dated 17 Nov 2026, 80 checks) | 1 | 1 | 0 | pass |
| Cross.regime: the site still serves the page this one copied (dated 17 Nov 2026) | 1 | 1 | 0 | pass |
| The Performance page: the source page is published with all its checks passing (dated 17 Nov 2026, 23 checks) | 1 | 1 | 0 | pass |
| The Performance page: the site still serves the page this one copied (dated 17 Nov 2026) | 1 | 1 | 0 | pass |
| The Performance page: every copied observation equals the source page, matched by date (1284 values in 12 arrays; mismatches) | 0 | 0 | 0 | pass |
| Cross.regime: every copied single value equals the source page (7 values; mismatches) | 0 | 0 | 0 | pass |
| Decisions: every desk decision this page's specification names is in the decisions file (D-2026-10-03-16, D-2026-10-03-21, D-2026-10-07-SKEW1, D-2026-10-07-SKEW2, D-2026-10-07-SKEW3, D-2026-10-07-SKEW4, D-2026-10-07-SKEW5; ids missing) | 0 | 0 | 0 | pass |
- Options engine session folder: end-of-day option chains, processed into a kernel-smoothed implied-volatility surface — marketdata.app
- The Correlations & regime page's weekly sign-rule state, copied as served and compared with it — Computed here (the sign-rule state of the Correlations & regime page).
- The Performance page's daily closes, copied as served: each market's 63-session correlation with the S&P 500 (the protective wing) — Computed here (closes from the Performance page).
- Cboe Global Indices: the VIX index's daily close, used only to find sessions whose option quotes are corrupt; not printed — Source: Cboe Global Indices, LLC