Fortress toolchain - predicted vs realized

Calibration Report

Scored 2026-10-10 cc_scanner 10,471 picks logged - 8,949 matured fortress_fight 3084 graded - span 99d fortress_rebuild 474 resolved wall 0m 28s (28.1s)

Predicted vs realized, across every calibration ledger

12,507 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% 90%
predicted vs realized, 8,949 matured picks across all sigma bands.
fortress_fight - touch
37% 34%
raw factor 0.91 over 3084 expired contracts.
fortress_fight - breach
478 breaches
raw breach factor 0.86 (predicted rate vs realized) over 3084 graded contracts.
fortress_rebuild - OTM at expiry
82% 84%
400 of 474 resolved. Touch: 36% predicted, 37% 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: 22 in progress (censored), 0 unexpired early touches (preview only).

perfectly calibrated 0-10% bucket - n=435 - predicted 4.0% - realized 4.1% 10-20% bucket - n=644 - predicted 16.4% - realized 14.1% 20-30% bucket - n=387 - predicted 24.3% - realized 24.3% 30-40% bucket - n=366 - predicted 34.6% - realized 31.7% 40-50% bucket - n=409 - predicted 44.3% - realized 41.6% 50-60% bucket - n=196 - predicted 55.0% - realized 44.9% 60-70% bucket - n=117 - predicted 65.0% - realized 57.3% 70-80% bucket - n=224 - predicted 75.9% - realized 68.8% 80-90% bucket - n=197 - predicted 84.4% - realized 79.2% 90-100% bucket - n=109 - predicted 96.5% - realized 89.9% 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
3084 / 30
✓
Breach events ≥ 10
478 / 10
✓
Expiry weeks ≥ 6 (one cycle teaches one regime)
14 / 6
✓
History span ≥ 60d (must see vol vary)
99d / 60d
✕
Challenger beats raw out of sample (champion check)
raw
MetricNPredictedRealizedRaw factor
Touch308437.3%34.1%0.91
Breach3084realized 478 breach event(s)0.86
IV haircutcorrect the input: touch model on iv × h0.92
Champion/challenger (rolling-origin walk-forward, 13 folds, 2771 pooled out-of-sample rows): flat Brier 0.1759 / ECE 0.0634; haircut Brier 0.1752 / ECE 0.0694; logit Brier 0.1746 / ECE 0.0685; raw Brier 0.1697 / ECE 0.0274; trend Brier 0.1763 / ECE 0.0806 (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 24.9%, ratio 0.73W282026-W29: n=574, predicted 39.6%, realized 20.9%, ratio 0.53W292026-W30: n=530, predicted 41.8%, realized 39.4%, ratio 0.94W302026-W31: n=257, predicted 34.8%, realized 39.3%, ratio 1.13W312026-W32: n=368, predicted 43.4%, realized 50.5%, ratio 1.17W322026-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 34.9%, ratio 1.07W372026-W38: n=155, predicted 34.7%, realized 32.9%, ratio 0.95W382026-W39: n=86, predicted 34.3%, realized 46.5%, ratio 1.36W392026-W40: n=82, predicted 30.7%, realized 11.0%, ratio 0.36W402026-W41: n=88, predicted 36.3%, realized 27.3%, ratio 0.75W41
WeekNPred touchReal touchRatioPred breachReal breach
2026-W2831334.3%24.9%0.7316.5%7.7%
2026-W2957439.6%20.9%0.5319.1%2.6%
2026-W3053041.8%39.4%0.9420.2%10.9%
2026-W3125734.8%39.3%1.1316.8%21.8%
2026-W3236843.4%50.5%1.1720.3%39.4%
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%34.9%1.0715.8%14.5%
2026-W3815534.7%32.9%0.9517.0%16.1%
2026-W398634.3%46.5%1.3616.7%19.8%
2026-W408230.7%11.0%0.3614.9%4.9%
2026-W418836.3%27.3%0.7517.7%10.2%

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%141839%36%-4 pts
flat +/-10%142536%34%-2 pts
up >10%23132%25%-7 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%
2026-09-19282838% / 33%0.890.81raw762487% / 90%37636% / 38%
2026-09-26291438% / 34%0.910.84raw800386% / 90%40636% / 39%
2026-10-03299637% / 33%0.900.82raw857786% / 90%44836% / 38%
2026-10-10308437% / 34%0.910.86raw894986% / 90%47436% / 37%

cc_scanner: sigma bands, then ticker by ticker

8,949 matured picks graded against daily highs (10,471 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=724
64% → 82%
0.5-1.0σ - n=2803
78% → 86%
1.0-1.5σ - n=2396
88% → 92%
>1.5σ - n=3026
97% → 95%

Touch % by sigma band

0.0-0.5σ - n=724
56% → 68%
0.5-1.0σ - n=2803
36% → 32%
1.0-1.5σ - n=2396
23% → 18%
>1.5σ - n=3026
8% → 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
+43 ▲ n=225
BMNR
+25 ▲ n=279
AAPL
+23 ▲ n=33
IBIT
+20 ▲ n=12
META
+18 ▲ n=294
MSTR
+16 ▲ n=150
NVDA
+16 ▲ n=247
AMD
+14 ▲ n=75
DELL
+14 ▲ n=333
IGV
+11 ▲ n=287
AMZN
+10 ▲ n=273
SPY
+8 ▲ n=274
MDB
+8 ▲ n=252
SNDK
+6 ▲ n=97
RIOT
+5 ▲ n=192
NOW
+2 ▲ n=291
GOOG
0 n=909
SOFI
-3 n=152
COPX
-3 n=261
GLD
-3 n=217
RKLB
-4 n=27
HIMS
-6 n=332
INTC
-7 n=327
COIN
-11 n=555
CRWV
-12 n=90
MU
-14 n=723
QCOM
-16 n=207
ENPH
-16 n=151
SPCX
-17 n=200
APP
-18 n=194
GLXY
-19 n=159
IREN
-22 n=724
MARA
-23 n=150
CLSK
-23 n=163
HOOD
-33 n=30
CRCL
-40 n=30

fortress_rebuild: CC-survival forecasts

Rebuild-local ledger (same calibration machinery as fight): 580 predictions since 2026-07-03, 474 resolved. Buckets group picks by the survival odds the model claimed.

Predicted survival Realized OTM rate
0-60% - n=26
51% → 73%
60-70% - n=34
67% → 74%
70-80% - n=94
76% → 77%
80-90% - n=195
85% → 85%
90-100% - n=125
94% → 95%
MetricValue
Predictions logged580
Resolved (expiry passed)474
Predicted touch (avg)35.9%
Realized touch36.6%
Touch grades from exact daily highs459 / 473

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. 77 nights logged (2026-07-24 → 2026-10-10).

No confident add yet — 78 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
HCCCoal (ENERGY)4698%+0.23-1.51%46%WATCH
VKTXGLP-1 (HEALTH)4698%+0.20-1.04%4%WATCH
BILLFintech (TECH)4598%-0.24+1.98%71%WATCH
CAVARestaurants (CONS)4598%+0.11-0.61%13%WATCH
EXPETravel (CONS)4598%-0.21+1.99%79%WATCH
LDOSDef Tech (INDL)4598%-0.03+0.06%100%WATCH
LLYGLP-1 (HEALTH)4598%-0.31+0.74%96%WATCH
NFLXInternet (TECH)4598%-0.13+1.03%78%WATCH
NVOGLP-1 (HEALTH)4598%-0.20+0.97%67%WATCH
PARRRefiners (ENERGY)4598%+0.02+1.66%91%WATCH
TOSTFintech (TECH)4598%-0.14+1.22%78%WATCH
UBERInternet (TECH)4598%-0.04+0.27%100%WATCH
VGMidstream (ENERGY)4598%-0.17+1.10%100%WATCH
XOPOil E&P (ENERGY)4598%-0.18+0.76%100%WATCH
ABNBInternet (TECH)4498%+0.02+1.10%83%WATCH
DKNGCasinos (CONS)4498%-0.06-0.10%23%WATCH
FLUTCasinos (CONS)4498%-0.19+0.30%65%WATCH
SNOWSoftware (TECH)4494%+0.20+0.48%9%WATCH
ZSCyber (TECH)4498%+0.03+0.83%35%WATCH
OIHOil Svcs (ENERGY)4289%+0.24-0.78%77%WATCH
AMGNGLP-1 (HEALTH)4189%-0.14+0.87%100%WATCH
ZBRARobotics (INDL)3983%+0.23+0.51%53%WATCH
AMRCoal (ENERGY)3883%+0.24-1.96%50%WATCH
HONRobotics (INDL)3779%+0.25-0.89%31%WATCH
FTNTCyber (TECH)3576%+0.13-0.17%64%WATCH
IHIMedTech (HEALTH)3576%-0.33+1.24%100%WATCH
NRGPower (ENERGY)3094%+0.24-2.89%0%WATCH
MDBSoftware (TECH)3097%+0.18+0.13%12%WATCH
SQMRare Earth (MATL)2997%+0.26-1.67%0%WATCH
WFRDOil Svcs (ENERGY)2997%+0.22-1.16%14%WATCH
XYZFintech (TECH)2897%+0.19-0.02%0%WATCH
ROKRobotics (INDL)2377%+0.26-1.23%50%WATCH
TXSteel (MATL)2377%+0.23-0.51%14%WATCH
VIKTravel (CONS)2170%+0.25-0.57%0%WATCH
CIBRCyber (TECH)1895%+0.26-0.46%youngWATCH
PANWCyber (TECH)2247%+0.21-0.38%0%noise
CRWDCyber (TECH)2247%+0.22-0.61%0%noise
BTUCoal (ENERGY)36%+0.22-2.69%0%noise
CNRCoal (ENERGY)2351%+0.23-1.95%22%noise
PLTRDef Tech (INDL)2861%+0.23-0.40%23%noise
XHBBuilders (CONS)2269%+0.23-0.41%0%noise
JETSAirlines (CONS)1040%+0.24+0.04%0%noise
AVAVDef Tech (INDL)613%+0.26-2.29%33%noise
“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. 35 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.