Back to home
Public learning proof
Learning Proof
A first-class proof chain showing whether the V2 engine is actually improving over time, whether the aggregate still matches ledger truth, and whether calibration and challenger workflows are healthy enough to support public claims.
Generated 09/05/2026, 16:31:54
HEALTHY
Predictions Analysed
63K+
63,354 exact surfaced forecasts
Settled Forecasts
57,092
Exact settled surfaced V2 ledger
Model Brier
0.2035
Lower is better
Alpha
-0.0060
Market Brier minus model Brier
ROI
-0.92%
Flat-stake surfaced ROI
Calibration Freshness
0.1h
SLA: 24h
Aggregate Reconciliation
Healthy
Last reconciled 09/05/2026, 16:32:26
Max drift 0.000000
Calibration SLA
Within SLA
Last updated 09/05/2026, 16:24:57
Challenger Activity
Active
Last activity 04/05/2026, 06:13:58
Pending promotions: 3 · recent decisions: 25
Rolling Walk-Forward Evidence
Shareable audit of persisted pre-match V2 forecasts. Rows are grouped by forecast-created week, and a cohort is only evaluated once that market has at least 100 prior surfaced settled forecasts. Forecast rows are not the same as executed bets: many markets can come from one fixture. This is deliberately separate from the tiny scheduled weekly comparison jobs.
Generated 09/05/2026, 16:32:33
Scope: 365d, surfaced-only, engine opportunities-engine-v2, cohort field createdAt, settled-outcome field settledAt. This validates stored forecasts against the market Brier baseline; it is not a claim that every row was executable at unlimited stake size.
Evaluated Holdout
56,581
2,690 fixtures, 57,092 eligible rows
Cohorts
5
1 warm-up cohorts skipped
Alpha vs Market
-0.0060
Market Brier minus model Brier
Flat ROI
-0.99%
-560.43u over holdout rows
CLV Check
+14.32%
34.7% hit, 16,186 samples
| Market | Eligible | Holdout | Fixtures | Cohorts | Alpha | ROI | Model Brier | Market Brier | Avg CLV | Range |
|---|---|---|---|---|---|---|---|---|---|---|
| All Markets | 57,092 | 56,581 | 2,690 | 5 | -0.0060 | -0.99% | 0.2034 | 0.1974 | +14.32% | 27/03/2026 to 09/05/2026 |
| Asian Handicap | 13,873 | 10,569 | 1,887 | 4 | -0.0227 | -7.75% | 0.2265 | 0.2037 | n/a | 27/03/2026 to 09/05/2026 |
| Goals Over/Under | 10,259 | 10,127 | 2,513 | 5 | -0.0062 | -1.31% | 0.1904 | 0.1842 | -1.20% | 27/03/2026 to 09/05/2026 |
| Match Winner | 6,166 | 4,966 | 1,884 | 4 | -0.0051 | -2.60% | 0.2002 | 0.1951 | -1.65% | 27/03/2026 to 09/05/2026 |
| Half Goals | 6,418 | 4,352 | 1,762 | 4 | +0.0727 | +14.29% | 0.1933 | 0.2660 | +129.52% | 27/03/2026 to 09/05/2026 |
| Double Chance | 5,020 | 3,713 | 1,612 | 4 | -0.0107 | -4.26% | 0.2298 | 0.2191 | -1.05% | 27/03/2026 to 08/05/2026 |
| Corners | 5,594 | 3,601 | 1,038 | 4 | -0.0135 | -6.55% | 0.2161 | 0.2026 | n/a | 28/03/2026 to 08/05/2026 |
| Correct Score | 3,154 | 2,345 | 1,114 | 4 | -0.0017 | -22.80% | 0.0538 | 0.0521 | n/a | 29/03/2026 to 08/05/2026 |
| BTTS | 2,511 | 1,734 | 1,080 | 4 | -0.0040 | -0.74% | 0.2446 | 0.2405 | -1.39% | 28/03/2026 to 09/05/2026 |
| Clean Sheet | 1,710 | 1,333 | 1,203 | 4 | +0.0006 | +9.75% | 0.1720 | 0.1726 | n/a | 28/03/2026 to 09/05/2026 |
| Even/Odd Goals | 1,754 | 544 | 498 | 4 | -0.0037 | -2.59% | 0.2541 | 0.2504 | -0.12% | 27/03/2026 to 09/05/2026 |
| HT/FT | 563 | 264 | 220 | 3 | -0.0032 | -17.96% | 0.1120 | 0.1089 | n/a | 03/04/2026 to 05/05/2026 |
| Exact Goals | 68 | 0 | 0 | 0 | n/a | n/a | n/a | n/a | n/a | 06/05/2026 to 08/05/2026 |
Win to Nil evidence quarantined
Old Win to Nil rows were recomputed with strict settlement rules, then excluded from headline proof because the original odds feed mapped stale market IDs into Win to Nil. These rows remain visible here as an audit trail, but they do not count toward public ROI, Brier, alpha, CLV, or market-family proof.
Quarantined Rows
6,356
6,801 historical Win to Nil rows
Active Proof Rows
0
0 settled rows currently counting
Removed ROI
+846.47%
Audit-only; +57568.60u is excluded
Strict Audit ROI
+1007.87%
5,957 recomputable rows, still excluded
Dominant Bad Pattern
31.27
Win to Nil - Home / No / key=no / side=home
Reason: win_to_nil_market_id_contamination_2026_04_14. Last quarantine write: 05/05/2026, 18:13:05. Latest captured row: 11/04/2026, 17:43:06. Latest active row: n/a. Generated 09/05/2026, 20:58:24.
| Match | Stored Bet | Book | Captured | Score | Stored | Strict | Unit |
|---|---|---|---|---|---|---|---|
Rayo Vallecano vs Elche La Liga | Win to Nil - Home / No @ 41.00 - Stored outcome differed from strict Win to Nil settlement | bet365 | 03/04/2026, 14:21:34 | 1-0 | won | lost | -1.00u |
Grenoble Foot 38 vs Clermont Ligue 2 | Win to Nil - Home / No @ 29.00 - Stored outcome differed from strict Win to Nil settlement | bet365 | 03/04/2026, 14:21:24 | 2-2 | lost | won | +28.00u |
Grenoble Foot 38 vs Clermont Ligue 2 | Win to Nil - Home / No @ 29.00 - Stored outcome differed from strict Win to Nil settlement | bet365 | 03/04/2026, 14:18:32 | 2-2 | lost | won | +28.00u |
Nottingham Forest vs Burnley Premier League | Win to Nil - Home / No @ 41.00 - Highest contaminated stored odds | bet365 | 11/04/2026, 14:55:36 | 4-1 | won | won | +40.00u |
Rayo Vallecano vs Espanyol La Liga | Win to Nil - Home / No @ 41.00 - Highest contaminated stored odds | bet365 | 11/04/2026, 14:05:01 | 1-0 | lost | lost | -1.00u |
Cambridge United vs Notts County League Two | Win to Nil - Home / No @ 41.00 - Highest contaminated stored odds | bet365 | 11/04/2026, 13:46:51 | 4-0 | lost | lost | -1.00u |
Proof Chain Windows
Exact surfaced settled windows comparing older periods against recent behavior.
| Window | Sample | Brier | Market Brier | Alpha | ROI | Avg CLV | CLV Hit |
|---|---|---|---|---|---|---|---|
| Older 7d | 1,070 | 0.1133 | 0.3815 | +0.2681 | 41.72% | 10.85 | 2.2% |
| Recent 7d | 17,354 | 0.2010 | 0.1965 | -0.0045 | -3.78% | -1.01 | 1.7% |
| Older 30d | 15,596 | 0.2095 | 0.1927 | -0.0168 | 4.43% | 8.61 | 10.0% |
| Recent 30d | 41,496 | 0.2013 | 0.1994 | -0.0019 | -2.94% | 16.45 | 9.9% |
7-Day Improvement
Older 7d vs recent 7d, with 95% confidence intervals on the change.
Brier Improvement
-0.0877
95% CI -0.1021 to -0.0733
Statistically clear · worsened
Alpha Improvement
-0.2700
95% CI -0.3000 to -0.2500
Statistically clear · worsened
ROI Improvement
-45.51pp
95% CI -57.74 to -33.27pp
Statistically clear · worsened
Avg CLV Improvement
-11.86%
95% CI -25.20 to 1.48%
Directional only · worsened
CLV Hit Improvement
-0.57pp
95% CI -4.59 to 3.46pp
Directional only · worsened
30-Day Improvement
Older 30d vs recent 30d for a broader view of whether the learning story survives over longer horizons.
Brier Improvement
+0.0082
95% CI 0.0041 to 0.0122
Statistically clear · improved
Alpha Improvement
+0.0100
95% CI 0.0100 to 0.0200
Statistically clear · improved
ROI Improvement
-7.37pp
95% CI -10.23 to -4.51pp
Statistically clear · worsened
Avg CLV Improvement
+7.83%
95% CI 5.98 to 9.69%
Statistically clear · improved
CLV Hit Improvement
-0.17pp
95% CI -1.14 to 0.81pp
Directional only · worsened