Where ideas go to be judged. Fit strictly on the past, bet at opening prices, graded against the closing line. Most ideas lose — that's the point of checking.
latest 15 shown
| at | 2026-07-14T19:16:16.167089+00:00 |
| failed | 0 |
| healthy | yes |
| check | detail | status |
|---|---|---|
| feed:api-football | last fetch 0.0h ago (max 8h) | ok |
| feed:kalshi | last fetch 5.1h ago (max 13h) | ok |
| feed:the-odds-api | last fetch 0.6h ago (max 26h) | ok |
| feed:espn | last fetch 14.7h ago (max 30h) | ok |
| odds:upcoming-24h | 2/2 priced contests fresh (<14h); stale: none | ok |
| derived:metrics | last metric computed 10.1h ago (max 30h) | ok |
| derived:sims | last sim finished 10.0h ago (max 30h) | ok |
| results:finished-unscored | 0 finished contests missing scores: none | ok |
| odds:multi-book | all upcoming competitions have >=3 fresh sportsbooks | ok |
| quota:api-football | 2649 requests today (plan ~150k/day) | ok |
first 10 of 11 rows
| note | survivor = +EV in train AND test at opening prices; gold = also +EV at closing (beats the sharpest line). Expect a few false positives from multiple testing even so. |
| n survivors | 4 |
| segments scanned | 127 |
| sel | band | comp | market | test n | train n | test roi |
|---|---|---|---|---|---|---|
| away side win | priced 1.00–1.80 | PPL | Match result | 211 | 597 | 6.56 |
| away side win | priced 1.00–1.80 | SA | Match result | 180 | 583 | 11.54 |
| away side win | even | SA | Match result | 319 | 1,144 | 5.61 |
| away side win | even | FL1 | Match result | 291 | 886 | 3.1 |
| sel | band | comp | market | test n | train n | test roi |
|---|---|---|---|---|---|---|
| away side win | priced 1.00–1.80 | PPL | Match result | 211 | 597 | 6.56 |
| away side win | priced 1.00–1.80 | SA | Match result | 180 | 583 | 11.54 |
| away side win |
| Backtest | Model | Comp | Bets | ROI | CLV | Brier edge |
|---|---|---|---|---|---|---|
| #29 | ml_gbm_v1 | PPL | 3,581 | -3.8% | 0.68% | -0.00063 |
| #28 | ml_gbm_v1 | DED | 4,313 | -1.8% | -0.14% | -0.00207 |
| #27 | ml_gbm_v1 | FL1 | 5,766 | -5.1% | 0.10% | -0.00221 |
| #26 | ml_gbm_v1 | BL1 | 5,468 | -4.4% | 0.49% | -0.00237 |
| #25 | ml_gbm_v1 | SA | 6,574 | -7.2% | 0.52% | -0.00243 |
| #22 | ml_gbm_v1 | PD | 6,843 | -7.3% | 0.15% | -0.00318 |
| #10 | ml_gbm_v1 | EPL | 336 | -4.7% | 0.28% | -0.00357 |
| #21 | ml_gbm_v1 | EPL | 8,430 | -3.3% | 0.40% | -0.00391 |
| #18 | ml_gbm_v1 | EPL | 9,586 | -1.5% | 0.53% | -0.00465 |
| note | skill_captured_vs_market: 1.0 = as sharp as the closing line, 0 = no better than the base rate. Biases marked EDGE are stable disagreements that PAY; MODEL_DEFECT are stable disagreements that COST (fix the model there). Signals go weak_or_decaying when any era's correlation flips sign. |
| model run | 18 |
| n | signal | status | target | corr overall | corr by window |
|---|---|---|---|---|---|
| 33,804 | match.ref_cards_avg | alive | cards | 0.241 | 0.2554, 0.2884, 0.2158, 0.1996 |
| 44,630 | match.style_clash_corners | alive | corners | 0.0816 | 0.086, 0.0701, 0.0865, 0.0693 |
| 137,822 | match.mismatch | alive | goal_margin | 0.1932 | 0.1518, 0.2123, 0.2149, 0.2031 |
| 131,137 | team.eye_test_v1 | alive | team_goals | 0.1856 | 0.1861, 0.1892, 0.1873, 0.1789 |
| 275,644 | team.elo | alive | team_goals | 0.2001 | 0.1853, 0.2167, 0.2086, 0.1955 |
| n | market | verdict | mean gap | selection | price band | following gap pays |
|---|---|---|---|---|---|---|
| 578 | Over/Under 2.5 goals | MODEL_DEFECT | 0.0661 | Under |
| n train | 13,698 |
| artifact | /app/data/downloads/gap_filter.joblib |
| base rate model right | 0.4915 |
| note | flat $1 at CLOSING prices (Bet365+Pinnacle best-of, the sharpest line there is); train 2020-2023, test 2024-2025. SURVIVES = +EV in BOTH windows with test n≥150. The always-favourite row is the margin baseline, not a strategy. |
| matches replayed | 14,400 |
| rule | test n | train n | verdict | test roi pct | train roi pct |
|---|---|---|---|---|---|
| book war fav only (crazy) | 16 | 51 | THIN | 23.38 | -18.14 |
| grinder fade (crazy) | 373 | 999 | DEAD | -1.15 | -1.16 |
| specialist gap 60 (ratio) | 2,443 | 3,720 | DEAD | -1.9 | -2.84 |
| public always favourite (baseline) | 5,333 | 9,067 | DEAD | -2.11 | -2.66 |
| clutch dog (crazy) | 372 | 721 | DEAD | -2.37 | -6.11 |
| fatigue fade 8sets (ratio) | 549 | 1,056 | DEAD | -2.44 | -8.59 |
| elo edge fav only (normal) | 1,528 | 1,776 | DEAD | -4.58 | -5.02 |
| elo edge 5 (normal) | 3,786 | 7,029 |
| note | market-shape re-score at consensus opening price, train<2022 / test>=2022. Model-gated rules scored on price band+selection only (gate ignored) — a SHAPE-NEGATIVE verdict is NOT a kill (e.g. wc_ko_unders is 5/6 in forward paper). NOT-REPLAYABLE strategies must be judged on their forward paper record (paper_bet.status in won/lost), not this proxy. |
| n live | 30 |
| n scored | 5 |
| n skipped | 25 |
| sel | band | code | comps | market | test n | train n |
|---|---|---|---|---|---|---|
| A | 1, 2.6 | away_fav_continental_v1 | SA, FL1, PPL | Match result | 1,331 | 4,396 |
| — | 2.2, 3.5 | mlb_dog_variance_v1 | all | Match winner (2-way) | 0 | 0 |
| D | 1, 1000 | ref_cards_draw_v1 | all | Match result | 12,649 | 55,726 |
| under | 1, 1000 | wc_ko_unders_v1 | all | Over/Under 2.5 goals | 12,649 | 10,313 |
| — | 4, 10 | wc_upset_hunter_v1 | all | Match result | 8,893 | 32,825 |
| code | reason |
|---|---|
| elo_value_mlb_v1 | model-only edge rule: no market shape to test |
| drought_fade_v1 | model-only edge rule: no market shape to test |
| note | pick = model edge>=min_edge on open price; _roi=(n, roi%). Skill only if pick_test ROI beats band_test AND is +EV. |
| min edge | 0.08 |
| EL2 | {"open":{"band_test":[165,-13.72],"pick_test":[72,-13.56],"b |
| EPL | {"open":{"band_test":[397,6.68],"pick_test":[93,42.43],"band |
| note | live paper record; statuses are won/lost/void/pending. Small early samples — treat as directional, not proof. This is how model-gated + soft/prop strategies must be judged (revalidate's shape proxy can't). |
| pnl | won | bets | lost | hit pct | pending | roi pct |
|---|---|---|---|---|---|---|
| 280.21 | 4 | 27 | 19 | 17.4 | 4 | 132 |
| 222.13 | 13 | 24 | 11 | 54.2 | 0 | 69.4 |
| 41 | 5 | 7 | 1 | 83.3 | 1 | 68.3 |
| 22 | 1 | 9 | 4 | 20 | 4 | 88 |
| 21.02 | 3 | 5 | 1 | 75 | 1 | 52.5 |
| 4.51 | 2 | 3 | 1 | 66.7 | 0 | 30.1 |
| 4.4 | 2 | 18 | 10 | 14.3 | 4 | 3.1 |
| 0 | 0 | 3 | 0 | 0 | 0 | 0 |
| 0 | 0 | 1 | 0 | 0 | 1 | 0 |
| 0 | 0 | 1 |
| note | same-day legs may be mildly correlated; trebles thin. If the single is +EV and legs ~independent, ROI should rise with legs (edge^n). |
| leg def | away, SA/FL1/PPL, price<= 2.6, 2015+ |
| price agg | max |
| single roi | 4.63 |
| compounds up | yes |
| bets | legs | hit pct | roi pct | biggest payout |
|---|---|---|---|---|
| 3,200 | 1 | 57.19 | 4.63 | 26 |
| 3,936 | 2 | 30.87 | 4.99 | 66.82 |
| 3,621 | 3 | 15.08 | -3.81 | 159.7 |
| note | in-sample pockets; same-day legs may be correlated; thin samples on doubles |
| closing | {"doubles":{"pnl":-2477.69,"bets":3369,"hit_pct":28,"roi_pct |
| opening | {"doubles":{"pnl":-582.06,"bets":6253,"hit_pct":31.8,"roi_pc |
| note | 1x2, sim-engine predictions, opening price, train<2022-07-01/test>=. Survive = +EV in BOTH halves, >=100 train & >=50 test bets. Multiple testing: ~a few false positives expected among the survivors; the train+test double-filter is the guard. Re-run after new odds land. |
| n survivors | 6 |
| predictions | 219,577 |
| cells scanned | 167 |
| sel | band | league | test n | train n | edge min | test roi |
|---|---|---|---|---|---|---|
| home side win | middog | EPL | 62 | 191 | 0.05 | 41.1 |
| home side win | middog | EPL | 78 | 223 | 0 | 22.6 |
| away side win | mid | EPL | 54 | 282 | 0.08 | 10.6 |
| away side win | mid | PPL | 81 | 290 | 0 | 6.5 |
| home side win | priced 1.00–1.80 | PPL | 90 | 600 | 0 | 4.7 |
| away side win | priced 1.00–1.80 | PPL | 63 | 252 | 0 | 3.5 |
| n | hit | roi | label |
|---|---|---|---|
| 540 | 0.35 | 0.1796 | EPL v1 (all 21 seasons) |
| 232 | 0.392 | 0.346 | EPL v2 (filters, SAME data — inflated by construction) |
| 104 | 0.394 | 0.3903 | EPL v2 on 2005-15 (derivation era) |
| 315 | 0.343 | 0.1504 | EPL v1 on 2016-26 (holdout) |
| 128 | 0.391 | 0.3101 | EPL v2 on 2016-26 (HOLDOUT — the number that matters) |
| 1,734 | 0.289 | -0.0416 | Other 6 leagues v1 |
| 768 | 0.279 | -0.067 | Other 6 leagues v2 (do the patterns travel?) |
| note | Real opening prices, 7 leagues x 21 seasons, per-season $1000 restarts. Cards/corners legs excluded (no historical odds exist) — their synthetic 5yr grading is soft_retro #185. Parlay legs assumed independent (different matches). |
| hot ref threshold | 3.6 |
| ref draw | {"bets":2876,"hit_rate":0.26,"roi_flat":-0.029,"worst_season |
| unders edge | {"bets":3050,"hit_rate":0.399,"roi_flat":-0.0518,"worst_seas |
| mid dog home | {"bets":4030,"hit_rate":0.294,"roi_flat":-0.0435,"worst_seas |
| PD | {"per_class":{"ref_draw":{"bets":105,"hit_rate":0.248,"roi_f |
| SA | {"per_class":{"ref_draw":{"bets":466,"hit_rate":0.27,"roi_fl |
| BL1 | {"per_class":{"ref_draw":{"bets":196,"hit_rate":0.265,"roi_f |
| DED | {"per_class":{"ref_draw":{"bets":32,"hit_rate":0.219,"roi_fl |
| EPL | {"per_class":{"ref_draw":{"bets":2065,"hit_rate":0.259,"roi_ |
| FL1 | {"per_class":{"ref_draw":{"bets":12,"hit_rate":0.167,"roi_fl |
| PPL | {"per_class":{"ref_draw":{"bets":0,"hit_rate":null,"roi_flat |
| bets | 1,765 |
| hit rate | 0.095 |
| roi flat | -0.1136 |
| worst season | 720.88 |
| league seasons | 21 |
| winning seasons | 7 |
| median season end | 978.26 |
| bets |
| note | posterior_vs_table counts seasons where the Bayesian leader differed from the table leader at that stage, and who won the title. Live posteriors get priced when the Monday outrights archive has real futures odds to compare against. |
| league seasons | 146 |
| n | band | mean p | realized | checkpoint |
|---|---|---|---|---|
| 102 | 20%-50% | 0.322 | 0.333 | 10% |
| 73 | 50%-80% | 0.635 | 0.507 | 10% |
| 50 | 80%-100% | 0.888 | 0.74 | 10% |
| 99 | 20%-50% | 0.334 | 0.374 | 20% |
| 72 | 50%-80% | 0.638 | 0.542 | 20% |
| 50 | 80%-100% | 0.912 | 0.78 | 20% |
| 94 | 20%-50% | 0.321 | 0.309 | 30% |
| 72 | 50%-80% | 0.64 | 0.639 | 30% |
| 56 | 80%-100% | 0.914 | 0.821 | 30% |
| 92 | 20%-50% | 0.339 | 0.326 | 40% |
first 10 of 27 rows
| league |
|---|
| Engine | Rule | Staking | Median end | Worst season | Winning |
|---|---|---|---|---|---|
| GBM model #28 → | when the expected tempo is high (top third of games): back the underdog (3.0+) | Fibonacci ladder | $2151 | $5 | 5/9 |
| GBM model #16 → | back away wins at any price, any positive model edge | target-chase ×2 | $2000 | $2 | 12/21 |
| GBM model #16 → | back home wins priced 2.60–4.50, any positive model edge | target-chase ×2 | $2000 | $7 | 13/21 |
| GBM model #16 → | back home wins priced 2.60–4.50, model edge over 2% | target-chase ×2 | $2000 | $13 | 13/21 |
| GBM model #16 → | when the skill mismatch is low (bottom third of games): back under 2.5 goals | target-chase ×2 | $2000 | $3 | 4/7 |
| note | Real historical slates from the walk-forward predictions at real opening prices, all leagues. P(hit target) is the mode's objective; mean_end shows what that shot at glory costs in EV. Parlay legs assumed independent (different matches). |
| target | turn $100 into $1000 same day |
| days replayed | 2,070 |
| avg slate size | 14 |
| flat 10 | {"mean_end":93.11,"bust_rate":0.0923,"median_end":85.7,"p_hi |
| half kelly | {"mean_end":96.6,"bust_rate":0,"median_end":93.69,"p_hit_tar |
| half ladder | {"mean_end":88.66,"bust_rate":0.4092,"median_end":3.12,"p_hi |
| parlay top4 | {"mean_end":63.66,"bust_rate":0.9942,"median_end":0,"p_hit_t |
| allin ladder | {"mean_end":93.26,"bust_rate":0.9464,"median_end":0,"p_hit_t |
| roi | 1.3644 |
| bets | 793 |
| note | Futures prices are SYNTHETIC: naive previous-rank base rates + 35% futures-book margin. Prediction-quality numbers (top-pick hit rate, P on actual champion) are price-free and fully honest; the ROI says whether season-sims beat a last-season-rank bookmaker, not a real one. |
| league seasons | 144 |
| mean p on actual champion | 0.4658 |
| top pick champion hit rate | 0.542 |
| top4 | {"roi":1.9944,"bets":209,"hit_rate":0.311} |
| bottom3 | {"roi":0.4825,"bets":365,"hit_rate":0.332} |
| champion | {"roi":2.2328,"bets":219,"hit_rate":0.352} |
| note | Live market is SYNTHETIC (closing-line lambdas conditioned on the live state) with realistic book shading: base vig + tail penalty growing with price, quotes capped at 34. roi_fair grades at unshaded Poisson-fair prices (upper bound / diagnostic only — real books fatten comeback tails). Overreaction edges are invisible here by construction; grading vs REAL live odds accumulates in af_live_odds. |
| model run | 18 |
| in play vig | 0.06 |
| matches replayed | 3,800 |
| bets | policy | staked | seasons | hit rate | roi fair | roi at vig |
|---|---|---|---|---|---|---|
| 1,324 | fade_overreaction | 13,240 | 10 | 0.167 | 0.5105 | 0.1168 |
| 4,063 | double_down | 40,630 | 10 | 0.294 | 0.1444 | 0.0572 |
| 3,216 | prematch_only | 32,160 | 10 | 0.329 | 0.0112 | 0.0112 |
| 5,042 | cover_lock | 33,584 | 10 | 0.252 | 0.0161 | 0.008 |
| 3,789 | value_stream | 37,890 | 10 | 0.23 | 0.2646 | -0.022 |
| note | identical picks and prices; only the structure differs. 'covered' refunds the whole outlay on a draw, 'half_cover' refunds half. median_trough = median of each season's lowest bankroll (the stability number). |
| model run | 18 |
| picks considered | 5,068 |
| covered | {"bets":4650,"wins":1650,"seasons":21,"best_season":1856.65, |
| straight | {"bets":5057,"wins":1947,"seasons":21,"best_season":2247,"fu |
| half cover | {"bets":5006,"wins":1900,"seasons":21,"best_season":2047.73, |
| Better at | Gets more | P(better side) | …when gap is big | n | Stable |
|---|---|---|---|---|---|
| team.elo | goals | 0.681 | 0.8295 | 40,458 | ⏳ yes |
| team.state_index | goals | 0.6666 | 0.7991 | 40,374 | ⏳ yes |
| team.venue_ppg_l5 | goals | 0.6484 | 0.7561 | 34,922 | ⏳ yes |
| team.goal_diff_avg_l5 | goals | 0.631 | 0.7413 | 36,833 | ⏳ yes |
| team.season_ppg | goals | 0.6327 | 0.7256 | 34,791 | ⏳ yes |
| team.league_points | goals | 0.6353 | 0.7257 | 29,292 | ⏳ yes |
| team.form_points_l5 | goals | 0.6207 | 0.7175 | 35,775 | ⏳ yes |
| team.league_rank | goals | 0.3715 | 0.3137 | 30,742 | ⏳ yes |
| team.goals_against_avg_l5 |
| Gap size | Direction | n | Model advantage | Reality | Model said | Market said | Stable |
|---|---|---|---|---|---|---|---|
| large | model_lower | 1,586 | -0.04565 | 0.483 | 0.3504 | 0.4813 | ⏳ yes |
| large | model_higher | 1,777 | -0.04481 | 0.3748 | 0.5171 | 0.3831 | ⏳ yes |
| medium | model_lower | 3,515 | -0.02048 | 0.3991 | 0.3165 | 0.3878 | ⏳ yes |
| medium | model_higher | 3,109 | -0.01393 | 0.377 | 0.4519 | 0.3802 | ⏳ yes |
| small | model_higher | 2,874 | -0.00375 | 0.3619 | 0.3976 | 0.3632 | ⏳ yes |
| small | model_lower | 3,286 | -0.00215 | 0.3478 | 0.3198 | 0.3543 | ⏳ yes |
| tiny | model_higher | 2,339 |
| note | walk-forward chronological folds; players need 180+ prior minutes |
| brier | 0.0727 |
| auc mean | 0.7623 |
| base rate | 0.0879 |
| n player matches | 151,430 |
| n | bin | p pred | p actual |
|---|---|---|---|
| 46,694 | 0.00-0.07 | 0.0354 | 0.0364 |
| 15,577 | 0.07-0.15 | 0.106 | 0.1102 |
| 7,299 | 0.15-0.22 | 0.1815 | 0.1841 |
| 3,697 | 0.22-0.30 | 0.2555 | 0.2589 |
| 1,680 | 0.30-0.37 | 0.329 | 0.344 |
| 662 | 0.37-0.45 | 0.4007 | 0.4048 |
| 104 | 0.45-0.52 | 0.4654 | 0.4808 |
| 2 | 0.52-0.59 | 0.5643 | 0.5 |
| days | 30 |
| note | avg_edge_bps > 0 = the book prices better than the cross-book median (generous or slow); best_share = how often it posts the top price. Line-shop these first; they are also the books whose stale quotes the sniper strategies hit. |
| n quotes | 4,744 |
| book | quotes | best share | avg edge bps |
|---|---|---|---|
| betfair_ex_eu | 54 | 0.3519 | 122.5 |
| matchbook | 54 | 0.3148 | 122 |
| onexbet | 54 | 0.2778 | 89.5 |
| af_pinnacle | 130 | 0.2077 | 76.6 |
| af_1xbet | 182 | 0.2747 | 63.9 |
| unibet_nl | 54 | 0.0926 | 61 |
| af_betano | 182 | 0.1758 | 54.9 |
| unibet_se | 54 | 0.0556 | 52.1 |
| pinnacle | 54 | 0 | 35.6 |
| af_marathonbet | 182 | 0.0934 | 33.8 |
first 10 of 37 rows
| book | market | quotes | best price share | generosity p90 bps |
|---|
| n refs known | 400 |
| tercile threshold | 3.8 |
| match | kickoff | referee | verdict | cards line | contest id | ref avg cards |
|---|---|---|---|---|---|---|
| United States v Belgium | 2026-07-07 00:00:00+00:00 | A. Makhadmeh | ref_unknown_to_us | 3.5 | 9,933 | — |
| Argentina v Egypt | 2026-07-07 16:00:00+00:00 | F. Letexier | ref_unknown_to_us | 3.5 | 105,099 | — |
| Switzerland v Colombia | 2026-07-07 20:00:00+00:00 | I. Barton | ref_unknown_to_us | 3.5 | 105,100 | — |
| note | teams whose rolling-form numbers may describe a team that no longer exists — the market prices form, form lies after regime changes. Grade after a season of alerts. |
| n games | 42 |
| n alerts | 1 |
| side | team | flags | match | kickoff | severity | contest id |
|---|---|---|---|---|---|---|
| away | Egypt | 1 items | Argentina v Egypt | 2026-07-07 16:00:00+00:00 | 2 | 105,099 |
| note | share_after_75 vs uniform (20/95≈0.21) quantifies the late-game spike; price_window() turns a match's lambda into P(goal in window). |
| n games | 5,521 |
| n goals | 17,000 |
| share after 75 | 0.243 |
| share after 85 | 0.1375 |
| share first 15 | 0.1167 |
| n games with goals | 5,521 |
| uniform share after 75 | 0.2105 |
| goal in both halves given any | 0.6283 |
| 0 | 0.0275 |
| 5 | 0.0439 |
| 10 | 0.0453 |
| 15 | 0.0462 |
| 20 | 0.0461 |
| 25 | 0.0466 |
| 30 | 0.0493 |
| 35 | 0.0463 |
| 40 | 0.0502 |
| 45 | 0.0789 |
| 50 | 0.0572 |
| 55 | 0.0562 |
| 0 | 0.0831 |
| 5 | 0.1213 |
| 10 | 0.1087 |
| 15 | 0.0976 |
| 20 | 0.0819 |
| 25 | 0.077 |
| 30 | 0.0656 |
| 35 | 0.054 |
| 40 |
| note | real prices only: Kalshi fee-adjusted candlestick asks for 1x2 everywhere; our archived/harvested book prices for totals/cards/corners where they exist. Walk-forward sims per match day; no synthetic pricing anywhere. |
| games | 92 |
| games with 1x2 price | 77 |
| games with soft prices | 24 |
| wc ko unders v1 | {"won":3,"bets":4,"lost":1,"final":1025.7,"history":[{"bet": |
| wc sim value v1 | {"won":8,"bets":57,"lost":49,"final":786.92,"history":[{"bet |
| cards over ref v1 | {"won":0,"bets":0,"lost":0,"final":1000,"history":[]} |
| wc upset hunter v1 | {"won":4,"bets":28,"lost":24,"final":988.32,"history":[{"bet |
| corners style over v1 | {"won":0,"bets":0,"lost":0,"final":1000,"history":[]} |
| sim value kelly of ruin | {"won":8,"bets":57,"lost":49,"final":620.51,"history":[{"bet |
| from | 2021-07-01 |
| price model | prior base rate x (1+0.06/2) implied, pessimistic +0.04 shade variant |
| fair margin | {"combined":{"bets":0,"path":[],"hit_rate":null,"flat_peak": |
| overs shaded | {"combined":{"bets":0,"path":[],"hit_rate":null,"flat_peak": |
| Gap size | Direction | n | Model advantage | Reality | Model said | Market said | Stable |
|---|---|---|---|---|---|---|---|
| large | model_lower | 1,616 | -0.04316 | 0.4666 | 0.3368 | 0.4679 | ⏳ yes |
| large | model_higher | 1,833 | -0.04063 | 0.3944 | 0.5228 | 0.3884 | ⏳ yes |
| medium | model_lower | 3,404 | -0.01534 | 0.386 | 0.3134 | 0.3848 | ⏳ yes |
| medium | model_higher | 2,922 | -0.01351 | 0.3765 | 0.4506 | 0.3788 | ⏳ yes |
| small | model_lower | 3,405 | -0.00381 | 0.3471 | 0.3185 | 0.3528 | ⏳ yes |
| small | model_higher | 2,911 | -0.00195 | 0.3748 | 0.4001 | 0.3661 | ⏳ yes |
| tiny | model_higher | 2,399 |
| even |
| SA |
| Match result |
| 319 |
| 1,144 |
| 5.61 |
| away side win | even | FL1 | Match result | 291 | 886 | 3.1 |
| #1 | poisson_tw | EPL | 353 | -5.8% | -1.05% | -0.00526 |
| #19 | ml_gbm_v1 | EPL | 9,793 | -1.9% | 0.38% | -0.00563 |
| #16 | ml_gbm_v1 | EPL | 9,588 | -2.9% | 0.42% | -0.00575 |
| #12 | poisson_tw | EPL | 1,442 | -11.4% | -0.44% | -0.00595 |
| #17 | poisson_tw | EPL | 1,442 | -11.4% | -0.44% | -0.00595 |
| #2 | poisson_tw | EPL | 2,322 | -7.0% | -0.25% | -0.00617 |
| #20 | ml_gbm_v1 | EPL | 9,900 | -2.4% | 0.40% | -0.00637 |
| #14 | ml_gbm_v1 | EPL | 9,248 | -2.2% | 0.31% | -0.00702 |
| #15 | ml_gbm_v1 | EPL | 9,659 | -1.0% | 0.31% | -0.00718 |
| #13 | ml_gbm_v1 | EPL | 9,247 | -1.4% | 0.42% | -0.00764 |
| #7 | ml_gbm_v1 | EPL | 8,235 | -1.3% | 0.28% | -0.00875 |
| #11 | ml_gbm_v1 | EPL | 8,203 | -2.7% | 0.42% | -0.00966 |
| #6 | ml_gbm_v1 | EL2 | 9,818 | -5.5% | -0.04% | -0.01067 |
| #4 | ml_gbm_v1 | EPL | 10,506 | -2.1% | 0.07% | -0.02901 |
| #3 | ml_gbm_v1 | EPL | 10,616 | -3.0% | 0.06% | -0.03131 |
| Code | Status | Description |
|---|---|---|
| draw_family_v1 | dormant | The sweep survivor: back the DRAW when model edge > 10% in high-tempo matchups. Verdict window +12.6% on 324 unseen bets (upper bound). [BLOCKED: model 'gbm' needs a forward-serving job the executor doesn't have yet — review finding #12; was mislabeled 'armed'.] |
| draw_calm_books_v1 | dormant | Sibling survivor: back the DRAW when edge > 15% and bookmaker margin is low. Verdict +4.5% on 429 unseen bets. [BLOCKED: model 'gbm' needs a forward-serving job the executor doesn't have yet — review finding #12; was mislabeled 'armed'.] |
| mlb_dog_variance_v1 | live | Baseball is the highest-variance major sport (rating->win only +0.12) yet the public loves favorites. Back MLB underdogs priced 2.2-3.5 when our Elo edge is at worst -3%. |
| elo_value_mlb_v1 | live | Straight Elo value on MLB moneylines: bet either side when Elo win probability beats the best price by 4%. |
| drought_fade_v1 | live | Law-derived (winless-run replicates in 4 sports): bet AGAINST any side whose opponent... no — whose own winless run is 4+ games, when priced under 2.8. Fade the drought. |
| steam_confusion_v1 | live | Market-structure play: when bookmakers DISAGREE hard on a price (cross-book spread > 8%), take the best available price on the side with any positive Elo edge — someone is wrong, take the generous one. |
| fatigue_fade_nba_v1 | armed | Law-derived (NBA congestion is real, r=-0.067): fade NBA teams playing their 4th+ game in 7 days against fresher opponents. |
| hot_hand_fade_v1 | live | Mean-reversion creative: fade sides whose Elo momentum is scorching (opponent momentum in top band) — the market overprices streaks. Expected to be humbled; that is the point of testing it. |
| chaos_parlay_v1 | live | The fun one: a $2 three-leg parlay of the day's highest-edge independent legs across all live strategies' candidates. Margin math says it should bleed; it exists to make the math visible. |
| placebo_rain_unders_v1 | dormant | CONTROL: bet football unders when heavy rain is forecast. Our own discovery says rain does NOT reduce goals — so this placebo SHOULD lose at the margin rate. If it wins long-term, our discovery engine has a bug. Science. [BLOCKED: model 'gbm' needs a forward-serving job the executor doesn't have yet — review finding #12; was mislabeled 'armed'.] |
| cards_closeness_v1 | dormant | Discovery-derived (closeness->fouls, r=+0.19 emergent): cards/fouls overs in evenly-matched games. Dormant until a live cards-odds feed exists. UPDATE 2026-07-03: referee identity alone swings P(4+ cards) from 33.6% to 54.0% (quartile split, n=21,870) — add match.ref_cards_avg regime when the odds feed exists. |
| value_all_v1 | retired | Baseline engine: any market, 5% edge, fractional Kelly. Backtest verdict: -2.05% over 20 EPL seasons. Kept as the reference corpse. |
| wc_sim_value_v1 | live | World Cup: back any 1x2 selection where the Poisson fixtures-sim beats the best live price by 8%. Caveat honored: the sim runs hot on tournament data, hence the high bar. [2026-07-07: gap-filter gate mandatory — naked longshot gaps went 8W-49L in the cup retro and the gf twin vetoed them all.] |
| wc_upset_hunter_v1 | live | World Cup knockout chaos: longshots priced 4-10 where the sim still sees positive edge. Small stakes, big stories. |
| wc_ko_unders_v1 | live | Knockout football is cagey (legs tighten when losing means going home): under 2.5 goals when the sim agrees (any positive edge). Hypothesis-tagged. |
| steam_chaser_v1 | live | Line-movement study (78k selections, 20yr): big steam (4+ prob-pt moves) is the one bucket the close doesn't fully absorb (+2.3pts residual), concentrated in soft leagues (+5.9% EL2) and away sides (+6.4%). Hypothesis-grade (CI spans zero) — this forward trial is the judge. Bets any selection whose consensus implied prob rose 4+ pts across our own daily snapshots. |
| portfolio_v1 | live | Layer 3, the bet-maker: builds the best BASKET across everything priced (ballast/value/longshot buckets, one bet per contest, quarter-Kelly, model blended 50/50 with the market) and Monte Carlos each slate before betting. Historical replay verdict: with the current football model's edge it would have gone bust over 21 seasons ($1000 -> $0.76) — this forward trial tests whether the newer, humbler probabilities change that. |
| pitcher_edge_v1 | live | Who's on the mound matters most in baseball. Bet an MLB moneyline side when its probable starter's season ERA is at least 1.00 better than the opponent's (mlb.starter_era_gap, signed to the bet side) AND our Elo sees any positive edge at the best live price. First strategy to use contest-scoped context metrics. |
| prop_scorer_v1 | live | The first player-level strategy: a gradient-boosted scorer model (rolling goals/shots per 90, starting rate, team attack vs opponent defence, confirmed lineups when available) prices every Kalshi anytime-goalscorer market and bets where its probability beats the fee-adjusted ask by 10%+. Settles from Kalshi's own results. |
| ref_cards_draw_v1 | live | Promoted from the matrix tournament (won ALL 4 backtest seasons, worst season +$471): when the referee's card average is in the top tercile (>=3.52/game), the match tightens — back the DRAW blind. Model-free context bet; forward trial to survive multiple-comparisons doubt. |
| cards_over_ref_v1 | live | The referee signal finally bettable: card-happy ref (top tercile, >=3.52/game) -> back total cards OVER at the main line from the new odds feed. Settles from post-match team stats. |
| corners_style_over_v1 | live | Two corner-hungry teams meet (style clash top tercile, >=11.4 combined corner averages) -> back total corners OVER at the main line. Model-free; the relational miner showed corners follow style and mismatch. |
| wc_sim_value_kor_v1 | live | Staking A/B twin of wc_sim_value_v1: identical signal (Poisson sim beats best price by 8%) but sized by kelly_of_ruin — half-Kelly scaled by the posterior probability this bankroll's edge is real, from its own settled record. Measures the staking policy, not the signal. [2026-07-07: gap-filter gate mandatory — naked longshot gaps went 8W-49L in the cup retro and the gf twin vetoed them all.] |
| mls_sim_value_v1 | live | MLS forward trial: back any 1x2 selection where the fixtures sim beats the best price by 10%. Higher bar than WC — the sim is new to this league and MLS draws are notoriously frequent. |
| wc_sim_value_gf_v1 | retired | Gap-filtered twin of wc_sim_value_v1: identical sim signal, but each edge must ALSO be blessed by the gap-survival classifier (P(model side right) >= 0.55, trained on 21 seasons of model-vs-market disagreements). The meta-filter's live A/B. |
| dutch_cover_v1 | live | Half-cover dutch: back the sim's 1x2 pick (edge >= 5%) AND the draw, draw stake sized to refund HALF the outlay on a draw. The 21-season backtest (exp #203) found full cover overpays for insurance (median $898) while half-cover beats straight outright: median $1,060 vs $995, worst season $87 vs $17. |
| dd_live_v1 | live | In-play double-down: when a pending pre-match 1x2 single is TRAILING at a poll and the conditioned model posterior still shows >=8% edge at the real live price, add one $10 bet on the same side. The one in-play behavior that beat its control in the 3,454-match bayes replay (+8.1% vs +2.5%, exp #206). Placed by the inplay watcher, not the executor; graded against REAL live odds. |
| simple_parlay_v1 | live | Simple 2-leg parlay of the day's PROVEN-class bets only (ko unders, ref-gated cards over, style-gated corners over — the classes that are 11-for-13 live). Two +EV legs compound edge multiplicatively; small flat stake because variance compounds too. Legs come from bets the proven strategies actually placed today. |
| epl_mid_dog_v1 | live | The tuning sweep's best honest cell, armed for 2026-27: EPL home mid-dogs (price 2.6-4.5) when the engine's edge >= 8% -> +18% ROI over 540 real-priced bets, era-stable (+15%/+12%/+10% across 05-10/17-21/22-26). Half-Kelly per the matrix law. HONEST CAVEAT: thresholds chosen in-sample; this live trial is the out-of-sample test. |
| epl_mid_dog_v2 | live | v1 + the loss-pattern filters baked in: real favourite only (away rank <= 14), our dog not in the drop zone (home rank <= 17), home side alive (form >= 4 pts l5), and no Aug-Oct bets (early-season inputs are noise). Holdout 2016-26: +31.0% over 128 bets vs v1's +15.0% on the same seasons; filters do NOT travel to other leagues (-6.7%) — this is an EPL edge. A/B twin of v1, ~13 bets/season. |
| prop_scorer_v2 | live | Anytime-scorer v2.1: involvement-first gates (forwards gi>=0.40/90, attacking mids gi>=0.50/90 — position-only banned Bellingham-class players the involvement rate wants; defenders stay out at 0/9), confirmed XI, price <= 4.5, full scorer model, edge >= 10%. Retested on the real graded window: rejects all five of v1's losers, admits nothing that lost. A/B twin of v1. |
| prop_soa_v1 | live | Score-or-assist singles with the context gates baked in from day one (graded WC window: naked SoA -60%, elite lane -51%, price 5+ 0-for-13 -> tails banned): goal involvement >= 0.40/90 over last 12 apps, price 2.0-4.5 only, expected starter, elo not against (within 50), and the keeper gate — only bet INTO a cold opposing keeper (team.keeper_form_v1 below pool median). Kalshi SoA settles via settle-kalshi-props. |
| soa_duo_parlay_v1 | live | Tom's both-teams structure: when EACH side of one match has a qualifying prop_soa_v1 player, parlay the pair ($3). Mildly negatively correlated legs (one side dominating starves the other) — the trial measures whether the price makes up for it. |
| fade_goal_live_v1 | live | In-play fade-the-goal: when a goal goes in, back the CONCEDING side at the real live price if the conditioned posterior still shows >=10% edge. The bayes replay's best surviving policy (+11.7% ROI over 1,324 bets on the own-goal-corrected sample, 6/10 winning seasons vs +1.1% control). Placed by the watcher, max 2 per contest, savepoint-safe. The definitive grade is this live sample at real odds. |
| mad_game_live_v1 | live | Mad-game mode, live: the whole $1000 game budget committed to ONE match. Plan: spray_and_pray — equal-revenue dutch with a +10% return target, the one-game-format champion of the 12-plan mad-game replay (exp #225: median $1,100, 75.6% hit rate, 0.4% bust vs tranche_ladder's $758/10.7%). By 12' the watcher covers the largest outcome set whose dutch return clears the target at real live prices, stakes ~1/price (small on longshots, big on favourites), and holds the unstaked remainder as the designed loss buffer. tranche_ladder remains available via bet_rule.plan. One contest per day. |
| el2_mid_dog_v1 | retired | League Two home mid-dogs in the 3.5-4.5 price band when the sim engine's edge >= 8% -> +16% ROI over 161 real-priced bets (2017-25), positive in the last 5 straight seasons (+10/+20/+7/+5/+4%), one COVID-empty-stadium down year. The segmentation finder showed the edge concentrates in 3.5-4.5, not the wider 2.6-4.5 band (which was diluted to +3.5%). Half-Kelly per the matrix law (its honest-staking cells on League Two average ~+20%/season and never bust). HONEST CAVEAT: band + threshold chosen in-sample on the baseline poisson engine; this live trial is the out-of-sample test, and the tuned engine hasn't been applied yet. |
| goal_bets_v1 | live | Pre-match goal-market value: price total-goals (O/U 2.5) and BTTS off the Dixon-Coles score matrix built from each contest's stored sim lambdas (home/away-split, shrunk), bet where the book is soft by >=8%. Also surfaces the two likeliest correct scores. Placed by the goal-bets command, graded from the scoreline. Tom's goal-bets idea; the score matrix also feeds the half-time Bayesian re-bet (goal_bets_ht_v2). |
| goal_bets_ht_v2 | live | Bayesian half-time re-bet on total goals: from ~40' the live score is data, so the score matrix conditions on it (remaining goals ~ Poisson on the remaining lambda share) and the final O/U 2.5 posterior updates exactly. If it beats the live over/under price by >=8%, back it. Placed by the in-play watcher at real live odds, one re-bet per contest. v2 of goal_bets_v1; debuts on the WC quarter-finals. |
| goal_bets_late_over_v3 | live | Late-overs goal variant: from 50' back OVER 2.5 only when the score-matrix posterior beats the live over price by >=8%. Top of the goal-replay family (exp: +11.6% ROI, 8/10 winning seasons over 1225 corpus bets) BUT on synthesised live lines where late-over longshots may be flattered by the vig model — this WC trial is the real test. Backing late UNDERS was a -20.7% trap in the same backtest, so this variant is overs-only. Watcher-placed, one bet per contest. |
| away_fav_continental_v1 | live | edge_hunt finding (2026-07-09): back the AWAY side at short/even odds (<=2.6) in Serie A / Ligue 1 / Primeira. +3.0% ROI at consensus market price, +4.7% best-of-book over 3,366 bets (2015+). Survives a temporal train/test split, BEATS the closing line, positive in 13/14 recent seasons, holds at consensus (not a best-of-book artifact; ~3 books/game). Maps to the home-favouritism bias. CAVEATS: small edge (~3%); ~2012 regime change means exploitable-now not timeless; this live paper trial is the out-of-sample test. |
| wc_form_edge_v1 | live | World Cup: back any 1x2 pick where the fixtures-sim beats the best live price by 5%+ AND the bookmakers disagree with each other by 4%+ on that price — tournament markets get confused, and confusion is where a model edge is most believable. The lower edge bar than wc_sim_value_v1 is paid for by the confusion gate. |
| wc_keeper_wall_v1 | live | World Cup club-form carryover: a keeper doesn't forget how to save when he swaps his club shirt for his country's. Back a team (1x2) when THEIR keeper's rolling save% — built from ALL his matches, club included — is hot (>=75%) and the OPPONENT'S keeper is leaky (<=68%), requiring only a non-negative sim edge so the keeper signal is the driver. NOTE: fires only once nations carry team.keeper_form values (needs AF starter data for national-team matches — the daily catchup is filling this); silent until then by design, never guessing. |
| wc_parlay_v1 | live | World Cup accumulator: combines today's two best singles actually placed by the WC strategies (wc_sim_value_v1 / wc_upset_hunter_v1 / wc_form_edge_v1 — every leg already passed its own gates) into one 2-leg parlay. Small stake, tournament-sized price. |
| priced 2.60–4.50 |
| -0.0185 |
| 326 | Over/Under 2.5 goals | MODEL_DEFECT | -0.0495 | Under | priced 1.00–1.80 | -0.0221 |
| 1,190 | Over/Under 2.5 goals | MODEL_DEFECT | -0.044 | Over | priced 1.00–1.80 | -0.0095 |
| 1,428 | Over/Under 2.5 goals | MODEL_DEFECT | 0.0213 | Over | even | -0.0129 |
| 834 | Match result | MODEL_DEFECT | 0.0186 | home side win | priced 4.50+ | -0.0072 |
| 1,331 | Match result | MODEL_DEFECT | 0.0146 | home side win | priced 2.60–4.50 | -0.0103 |
| 945 | Match result | MODEL_DEFECT | -0.0139 | away side win | even | -0.0082 |
| 1,628 | Match result | noise | 0.0124 | home side win | even | -0.0124 |
| 1,491 | Match result | MODEL_DEFECT | 0.0083 | Draw | priced 4.50+ | -0.015 |
| 3,829 | Match result | noise | -0.0079 | Draw | priced 2.60–4.50 | -0.0094 |
first 10 of 15 rows
| n | question | by window | brier base | brier model | brier market | skill captured vs market |
|---|---|---|---|---|---|---|
| 15,960 | 1x2 | 0.19467, 0.1883, 0.19234, 0.1965 | 0.2222 | 0.1929 | 0.1885 | 0.871 |
| 5,320 | ou_goals_2_5 | 0.24364, 0.24554, 0.24695 | 0.25 | 0.2457 | 0.2402 | 0.437 |
| DEAD |
| -4.93 |
| -7.07 |
| elo edge 10 (normal) | 2,952 | 5,872 | DEAD | -5.64 | -7.2 |
| surface elo edge 5 (normal) | 4,013 | 7,477 | DEAD | -7.43 | -6.55 |
first 10 of 12 rows
| steam_confusion_v1 | model-only edge rule: no market shape to test |
| hot_hand_fade_v1 | model-only edge rule: no market shape to test |
| chaos_parlay_v1 | composite: depends on its leg strategies |
| wc_sim_value_v1 | model-only edge rule: no market shape to test |
| steam_chaser_v1 | steam: needs our snapshot-move history, not price levels |
| portfolio_v1 | composite: depends on its leg strategies |
| pitcher_edge_v1 | model-only edge rule: no market shape to test |
| prop_scorer_v1 | soft/prop market: ~no odds history (forward-capture only) |
first 10 of 25 rows
| 0 |
| 0 |
| 1 |
| 0 |
first 10 of 24 rows
| pnl | 429.84 |
| pending | 43 |
| roi pct | 30.4 |
| settled | 134 |
| strategies | 24 |
| 11,721 |
| hit rate | 0.283 |
| roi flat | -0.0526 |
| worst season | 441.76 |
| league seasons | 167 |
| winning seasons | 67 |
| median season end | 973.7 |
| bets | 40,977 |
| hit rate | 0.334 |
| roi flat | -0.0507 |
| worst season | 40.6 |
| league seasons | 146 |
| winning seasons | 47 |
| median season end | 854.4 |
| played |
|---|
| season |
|---|
| remaining |
|---|
| p champion top6 |
|---|
| MLS | 218 | 2026 | 292 | 6 items |
| checkpoint | table right | disagreements | posterior right |
|---|---|---|---|
| 10% | 16 | 91 | 49 |
| 20% | 19 | 64 | 33 |
| 30% | 14 | 54 | 27 |
| 40% | 14 | 41 | 18 |
| 50% | 9 | 35 | 19 |
| 60% | 9 | 25 | 12 |
| 70% | 5 | 18 | 10 |
| 80% | 5 | 19 | 11 |
| 90% | 5 | 14 | 8 |
| GBM model #16 → | when the Elo gap is low (bottom third of games): back over 2.5 goals | target-chase ×2 | $2000 | $9 | 4/7 |
| GBM model #16 → | when the referee's card average is high (top third of games): back under 2.5 goals | target-chase ×2 | $2000 | $2 | 5/7 |
| GBM model #16 → | when the match stakes is high (top third of games): back under 2.5 goals | target-chase ×2 | $2000 | $112 | 6/7 |
| GBM model #18 → | back over 2.5 goals priced 1.00–1.80, model edge over 2% | target-chase ×2 | $2000 | $14 | 5/7 |
| GBM model #18 → | back home wins priced 1.80–2.60, model edge over 5% | target-chase ×2 | $2000 | $25 | 12/21 |
| GBM model #18 → | back anything the model likes priced 1.80–2.60, model edge over 10% | target-chase ×2 | $2000 | $2 | 11/21 |
| GBM model #18 → | back match-result picks priced 1.80–2.60, model edge over 10% | target-chase ×2 | $2000 | $9 | 11/21 |
| GBM model #18 → | when the Elo gap is high (top third of games): back the favourite | target-chase ×2 | $2000 | $35 | 11/21 |
| GBM model #18 → | when the Elo gap is low (bottom third of games): back under 2.5 goals | target-chase ×2 | $2000 | $3 | 4/7 |
| GBM model #18 → | when the referee's card average is low (bottom third of games): back over 2.5 goals | target-chase ×2 | $2000 | $30 | 5/7 |
| GBM model #18 → | when the match stakes is high (top third of games): back under 2.5 goals | target-chase ×2 | $2000 | $2000 | 7/7 |
| GBM model #18 → | when the match stakes is low (bottom third of games): back over 2.5 goals | target-chase ×2 | $2000 | $55 | 4/7 |
| GBM model #21 → | back anything the model likes priced 1.80–2.60, any positive model edge | target-chase ×2 | $2000 | $2 | 13/21 |
| GBM model #21 → | back over 2.5 goals priced 1.80–2.60, any positive model edge | target-chase ×2 | $2000 | $2 | 5/7 |
| GBM model #21 → | back anything the model likes priced 1.00–1.80, model edge over 2% | target-chase ×2 | $2000 | $3 | 11/21 |
| GBM model #21 → | back over 2.5 goals priced 1.00–1.80, model edge over 2% | target-chase ×2 | $2000 | $4 | 6/7 |
| GBM model #21 → | back over 2.5 goals priced 1.80–2.60, model edge over 2% | target-chase ×2 | $2000 | $7 | 5/7 |
| GBM model #21 → | back home wins at any price, model edge over 5% | target-chase ×2 | $2000 | $2 | 14/21 |
| GBM model #21 → | back over 2.5 goals at any price, model edge over 5% | target-chase ×2 | $2000 | $2 | 4/7 |
| GBM model #21 → | back anything the model likes priced 1.80–2.60, model edge over 5% | target-chase ×2 | $2000 | $2 | 12/21 |
| GBM model #21 → | when the skill mismatch is high (top third of games): back under 2.5 goals | target-chase ×2 | $2000 | $2 | 4/7 |
| GBM model #21 → | when the Elo gap is low (bottom third of games): back over 2.5 goals | target-chase ×2 | $2000 | $3 | 5/7 |
| GBM model #21 → | when the expected tempo is high (top third of games): back the favourite | target-chase ×2 | $2000 | $23 | 14/21 |
| GBM model #21 → | when the referee's card average is high (top third of games): back under 2.5 goals | target-chase ×2 | $2000 | $514 | 6/7 |
| GBM model #21 → | when the match stakes is high (top third of games): back under 2.5 goals | target-chase ×2 | $2000 | $75 | 5/7 |
23,640 cells total · leaderboard shows the top by median season end.
| goals |
| 0.3872 |
| 0.2948 |
| 35,384 |
| ⏳ yes |
| team.goals_for_avg_l5 | goals | 0.6075 | 0.6891 | 35,563 | ⏳ yes |
| team.matches_since_win | goals | 0.4068 | — | 31,368 | ⏳ yes |
| team.h2h_win_rate_l5 | goals | 0.64 | — | 13,815 | ⏳ yes |
| team.shot_diff_avg_l5 | goals | 0.6111 | 0.7171 | 18,801 | ⏳ yes |
| team.dominance_l5 | goals | 0.6093 | 0.7144 | 19,055 | ⏳ yes |
| team.goals_volatility_l10 | goals | 0.5763 | 0.6308 | 37,783 | ⏳ yes |
| 0.3442 |
| 0.3573 |
| 0.3476 |
| ⏳ yes |
| tiny | model_lower | 2,532 | -0.00067 | 0.3424 | 0.3263 | 0.3362 | no |
Gap-survival classifier holdout AUC: 0.6778 · top features: rel_gap, price_open, prob_market, gap
Hurricane test over 4,560 shared predictions — Brier: engine A 0.19805, engine B 0.19611, engine mean 0.19499, market 0.19016. Averaging wrong models beats both members, but the market still beats the ensemble.
| avg edge vs consensus bps |
|---|
| betfair_ex_eu | Match result | 54 | 0.3519 | 195.5 | 122.5 |
| matchbook | Match result | 54 | 0.3148 | 220.9 | 122 |
| af_1xbet | Match result | 78 | 0.3077 | 205.4 | 108.5 |
| onexbet | Match result | 54 | 0.2778 | 170.1 | 89.5 |
| af_pinnacle | Over/Under 2.5 goals | 52 | 0.3077 | 176.3 | 80.1 |
| af_betano | Match result | 78 | 0.2051 | 239.9 | 79.4 |
| af_pinnacle | Match result | 78 | 0.141 | 152.2 | 74.3 |
| af_1xbet | Over/Under 2.5 goals | 52 | 0.1923 | 156 | 65.5 |
| af_marathonbet | Both teams to score | 52 | 0.25 | 191.6 | 64.3 |
| af_betano | Over/Under 2.5 goals | 52 | 0.2308 | 155.9 | 62 |
first 10 of 40 rows
| 0.0496 |
| 45 | 0.0638 |
| 50 | 0.0379 |
| 55 | 0.0357 |
| 0.3443 |
| 0.3633 |
| 0.3535 |
| ⏳ yes |
| tiny | model_lower | 2,460 | -0.00038 | 0.3394 | 0.3273 | 0.3372 | ⏳ yes |
Gap-survival classifier holdout AUC: 0.6787 · top features: rel_gap, price_open, prob_market, gap
Hurricane test over 4,560 shared predictions — Brier: engine A 0.19489, engine B 0.19611, engine mean 0.19337, market 0.19016. Averaging wrong models beats both members, but the market still beats the ensemble.