Fortress toolchain - predicted vs realized

Calibration Report

Scored 2026-09-12 cc_scanner 8,407 picks logged - 7,103 matured fortress_fight 2673 graded - span 71d fortress_rebuild 333 resolved wall 0m 17s (17.0s)

Predicted vs realized, across every calibration ledger

10,109 graded predictions across cc_scanner, fortress_fight, and fortress_rebuild, scored against actual daily highs and closes. Bars and dots compare what the models claimed to what the market did.

cc_scanner - survival
86% 89%
predicted vs realized, 7,103 matured picks across all sigma bands.
fortress_fight - touch
38% 34%
raw factor 0.90 over 2673 expired contracts.
fortress_fight - breach
395 breaches
raw breach factor 0.82 (predicted rate vs realized) over 2673 graded contracts.
fortress_rebuild - OTM at expiry
83% 85%
283 of 333 resolved. Touch: 34% predicted, 35% realized.

fortress_fight: touch calibration curve

Each dot is a predicted-touch bucket of expired contracts (dot size = count, hover for exact rates). On the dashed line, prediction equals reality. Dots below the line = the engine over-warned; above = under-warned. Also on the books: 24 in progress (censored), 1 unexpired early touches (preview only).

perfectly calibrated 0-10% bucket - n=373 - predicted 4.0% - realized 2.9% 10-20% bucket - n=546 - predicted 16.3% - realized 13.7% 20-30% bucket - n=349 - predicted 24.2% - realized 26.1% 30-40% bucket - n=292 - predicted 34.5% - realized 30.5% 40-50% bucket - n=353 - predicted 44.2% - realized 39.1% 50-60% bucket - n=178 - predicted 55.2% - realized 43.8% 60-70% bucket - n=106 - predicted 65.0% - realized 59.4% 70-80% bucket - n=208 - predicted 75.8% - realized 67.3% 80-90% bucket - n=160 - predicted 84.5% - realized 78.8% 90-100% bucket - n=108 - predicted 96.5% - realized 89.8% 0 25 50 75 100 0 25 50 75 100 Predicted touch % Realized touch %

Correction-factor arming gates

Raw factors are written every grading pass; the engine only applies them when all four gates pass.

✓
Graded contracts ≥ 30
2673 / 30
✓
Breach events ≥ 10
395 / 10
✓
Expiry weeks ≥ 6 (one cycle teaches one regime)
10 / 6
✓
History span ≥ 60d (must see vol vary)
71d / 60d
✕
Challenger beats raw out of sample (champion check)
raw
MetricNPredictedRealizedRaw factor
Touch267337.8%34.0%0.90
Breach2673realized 395 breach event(s)0.82
IV haircutcorrect the input: touch model on iv × h0.92
Champion/challenger (rolling-origin walk-forward, 9 folds, 2360 pooled out-of-sample rows): flat Brier 0.1784 / ECE 0.0787; haircut Brier 0.1769 / ECE 0.0839; logit Brier 0.1755 / ECE 0.0745; raw Brier 0.1702 / ECE 0.0370; trend Brier 0.1762 / ECE 0.0889 (experimental, loses to raw) → winner raw. Admission needs Brier no worse than raw AND ECE better; ties go to raw; a correction only arms while it wins. Experimental shapes are scored but never armed.

fortress_fight: week by week

Each row is one expiry week of fully graded contracts. Ratio = realized ÷ predicted touch: below 1.0 the model over-warned that week, above 1.0 it under-warned. A correction that would have looked right on the first rows is judged by the later ones.

1.0 = calibrated2026-W28: n=313, predicted 34.3%, realized 24.0%, ratio 0.70W282026-W29: n=574, predicted 39.6%, realized 20.2%, ratio 0.51W292026-W30: n=530, predicted 41.8%, realized 38.7%, ratio 0.93W302026-W31: n=257, predicted 34.8%, realized 36.6%, ratio 1.05W312026-W32: n=368, predicted 43.4%, realized 50.3%, ratio 1.16W322026-W33: n=212, predicted 36.2%, realized 40.6%, ratio 1.12W332026-W34: n=120, predicted 27.3%, realized 43.3%, ratio 1.59W342026-W35: n=114, predicted 29.6%, realized 24.6%, ratio 0.83W352026-W36: n=102, predicted 33.8%, realized 38.2%, ratio 1.13W362026-W37: n=83, predicted 32.5%, realized 33.7%, ratio 1.04W37
WeekNPred touchReal touchRatioPred breachReal breach
2026-W2831334.3%24.0%0.7016.5%6.7%
2026-W2957439.6%20.2%0.5119.1%1.6%
2026-W3053041.8%38.7%0.9320.2%10.0%
2026-W3125734.8%36.6%1.0516.8%18.7%
2026-W3236843.4%50.3%1.1620.3%38.9%
2026-W3321236.2%40.6%1.1217.2%19.3%
2026-W3412027.3%43.3%1.5913.2%31.7%
2026-W3511429.6%24.6%0.8314.3%7.9%
2026-W3610233.8%38.2%1.1316.3%24.5%
2026-W378332.5%33.7%1.0415.8%9.6%

Touch by prior-20-day trend (ex-ante, experimental)

Rows bucket each graded contract by its ticker’s return over the 20 trading days BEFORE the forecast, the one trend fact knowable at pick time. If realized splits by bucket while predicted does not, the model is trend-blind. Walk-forward verdict on the trend challenger: loses to raw out of sample.

Prior 20d trendNPredicted touchRealized touchGap
down >10%138540%36%-4 pts
flat +/-10%115236%33%-3 pts
up >10%12633%21%-13 pts

Scoring-run history

One row per weekly grading pass (cumulative sample at that date). Watch the raw factors converge and the winner column stay on raw.

RunFight NTouch pred / realRaw touch factorRaw breach factorWinnerScanner NScanner surv pred / realRebuild NRebuild touch pred / real
2026-09-03248938% / 34%0.890.80raw663986% / 88%30035% / 35%
2026-09-05259038% / 34%0.890.82raw694086% / 89%31535% / 35%
2026-09-12267338% / 34%0.900.82raw710386% / 89%33334% / 35%

cc_scanner: sigma bands, then ticker by ticker

7,103 matured picks graded against daily highs (8,407 logged since 20260703). Survival = finishing OTM. Red touch bars: realized ran hotter than predicted.

Predicted Realized

Survival % by sigma band

0.0-0.5σ - n=574
64% → 81%
0.5-1.0σ - n=2153
79% → 84%
1.0-1.5σ - n=1959
88% → 91%
>1.5σ - n=2417
97% → 94%

Touch % by sigma band

0.0-0.5σ - n=574
57% → 68%
0.5-1.0σ - n=2153
37% → 30%
1.0-1.5σ - n=1959
23% → 14%
>1.5σ - n=2417
7% → 8%

Predicted vs realized touch by ticker (min 3 matured picks)

Sorted by how hot reality ran vs the model. Red connectors: realized above predicted (model too relaxed). Gray connectors: realized below predicted (model too scared).

Ticker
0% — predicted • realized — 100%
Gap (real - pred)
ETHA
+74 ▲ n=34
NEM
+48 ▲ n=203
BMNR
+26 ▲ n=267
SNDK
+23 ▲ n=33
IBIT
+20 ▲ n=12
DELL
+17 ▲ n=256
AMZN
+14 ▲ n=237
NVDA
+12 ▲ n=185
MDB
+9 ▲ n=240
META
+8 ▲ n=246
SPY
+8 ▲ n=215
IGV
+8 ▲ n=216
RIOT
+5 ▲ n=192
NOW
+1 ▲ n=259
GLD
-1 n=158
COPX
-2 n=205
SOFI
-3 n=152
GOOG
-3 n=704
RKLB
-4 n=27
MSTR
-7 n=4
HIMS
-8 n=263
CRWV
-10 n=54
COIN
-14 n=441
QCOM
-16 n=198
ENPH
-16 n=149
AMD
-18 n=28
APP
-18 n=188
SPCX
-19 n=153
GLXY
-19 n=155
MU
-19 n=517
IREN
-21 n=574
CLSK
-23 n=159
MARA
-23 n=144
INTC
-24 n=235

fortress_rebuild: CC-survival forecasts

Rebuild-local ledger (same calibration machinery as fight): 428 predictions since 2026-07-03, 333 resolved. Buckets group picks by the survival odds the model claimed.

Predicted survival Realized OTM rate
0-60% - n=15
54% → 87%
60-70% - n=21
66% → 86%
70-80% - n=68
76% → 78%
80-90% - n=132
85% → 82%
90-100% - n=97
94% → 94%
MetricValue
Predictions logged428
Resolved (expiry passed)333
Predicted touch (avg)34.4%
Realized touch35.2%
Touch grades from exact daily highs331 / 332

Diversification complements

Out-of-sample search for tickers that zig when the $1M model book zags — low correlation, positive behaviour on the book's worst days, and enough volatility to sell covered calls against. 49 nights logged (2026-07-24 → 2026-09-12).

No confident add yet — 50 days of history, needs ≥60 nights at 70% persistence and 60% out-of-sample delivery. Candidates are ranked below; a complement moves opposite the model book (corr ≤ 0.30) AND still pays sellable CC premium (vol ≥ 25%).
TickerGroupNightsPersistMed corrCrisis dayOOS deliverVerdict
BILLFintech (TECH)2596%-0.14+1.29%100%WATCH
HCCCoal (ENERGY)2596%+0.22-1.51%31%WATCH
UBERInternet (TECH)2596%-0.01-0.26%100%WATCH
VKTXGLP-1 (HEALTH)2596%+0.21-0.65%8%WATCH
ABNBInternet (TECH)2496%+0.06+0.37%100%WATCH
CAVARestaurants (CONS)2496%+0.02-0.08%15%WATCH
EXPETravel (CONS)2496%-0.18+0.90%100%WATCH
IHIMedTech (HEALTH)2496%-0.31+1.04%100%WATCH
LDOSDef Tech (INDL)2496%-0.05-0.12%100%WATCH
LLYGLP-1 (HEALTH)2496%-0.22+0.65%100%WATCH
NFLXInternet (TECH)2496%-0.10+0.94%100%WATCH
NVOGLP-1 (HEALTH)2496%-0.21+0.78%100%WATCH
PARRRefiners (ENERGY)2496%-0.11+1.66%85%WATCH
SNOWSoftware (TECH)2492%+0.15+0.78%0%WATCH
TOSTFintech (TECH)2496%-0.11+0.79%100%WATCH
VGMidstream (ENERGY)2496%-0.30+1.60%100%WATCH
XOPOil E&P (ENERGY)2496%-0.28+0.86%100%WATCH
ZSCyber (TECH)2496%+0.02+0.83%17%WATCH
CNRCoal (ENERGY)2396%+0.23-1.95%0%WATCH
DKNGCasinos (CONS)2396%-0.09+0.16%8%WATCH
FLUTCasinos (CONS)2396%-0.19+0.30%92%WATCH
OIHOil Svcs (ENERGY)2388%+0.23-0.78%77%WATCH
AMGNGLP-1 (HEALTH)2080%-0.14+0.85%100%WATCH
AMRCoal (ENERGY)2080%+0.24-2.31%82%WATCH
NRGPower (ENERGY)1083%+0.22-2.45%youngWATCH
XHBBuilders (CONS)1091%+0.23-0.41%youngWATCH
XYZFintech (TECH)788%+0.20+0.58%youngnoise
MDBSoftware (TECH)990%+0.21+0.21%youngnoise
WFRDOil Svcs (ENERGY)889%+0.21-1.80%youngnoise
BTUCoal (ENERGY)312%+0.22-2.69%0%noise
PLTRDef Tech (INDL)1352%+0.22-1.07%0%noise
FTNTCyber (TECH)1456%+0.23-0.20%29%noise
TXSteel (MATL)889%+0.23-0.63%youngnoise
SQMRare Earth (MATL)889%+0.24-1.67%youngnoise
“Crisis day” = mean candidate return on the model composite's worst days (average correlation lies in a crash; this is the all-weather test). “OOS deliver” = share of matured claims whose low correlation held in the 30 days AFTER the flag. 26 WATCH.
Generated by calibration_report.py (weekly cron, Sat 08:30 SGT). Sources: runs/cc_scanner/PORTFOLIO/ledger_calibration.json, runs/fight/calib_factors.json, runs/rebuild/calibration_stats.json. Graders grade only fully expired windows; duplicate nightly re-forecasts of the same contract are deduped to one bet. This page mirrors the graders' published numbers and derives nothing of its own.