A calibrated sports predictor that publishes its own scorecard against a sharp market -- losses, nulls, and retractions included.
Every number below is copied verbatim from a committed artifact and wears its receipt: source path, as-of date, and verification status. No edge is claimed anywhere.
POSITIONING calibration, not profit · edge_claimed: false on every artifactModel total-runs forecast is sharper than the market's at end of inning 7 (n=55, 95% CI [+0.4131, +1.0810]). MODEL_SHARPER_PROVISIONAL -- still provisional on a small sample. A sharpness result measured leak-free, not an edge.
Across in-game win-prob checkpoints the market is as sharp or sharper. Model Brier 0.2377 vs market 0.2067 on MLB (n=78,986); NBA end-Q1 favors the market too. Both readings are true. We publish both.
Cross-sport calibration scoreboard
model vs market -- paired Brier / CRPS delta, 95% CI, verdict
| sport | market | checkpoint | n | model − market | 95% CI | verdict |
|---|---|---|---|---|---|---|
| nba | winprob (Brier) | end_q1 | 1592 | -0.0084 | [-0.0161, -0.0008] | MARKET SHARPER |
| nba | winprob (Brier) | halftime | 1593 | -0.0040 | [-0.0098, +0.0015] | UNDERPOWERED |
| nba | winprob (Brier) | end_q3 | 1593 | +0.0011 | [-0.0028, +0.0052] | UNDERPOWERED |
| nba | winprob (Brier) | q4_under5 | 1593 | +0.0019 | [-0.0010, +0.0048] | UNDERPOWERED |
| mlb_ingame | CRPS | home_margin|end_inning_3 | 20 | +0.4499 | [-0.0206, +1.0687] | UNDERPOWERED |
| mlb_ingame | CRPS | home_margin|end_inning_5 | 26 | +0.8555 | [+0.3259, +1.4888] | UNDERPOWERED |
| mlb_ingame | CRPS | home_margin|end_inning_6 | 22 | +0.9583 | [+0.2409, +1.8046] | UNDERPOWERED |
| mlb_ingame | CRPS | home_margin|end_inning_7 | 21 | +1.0660 | [+0.2656, +1.9602] | UNDERPOWERED |
| mlb_ingame | CRPS | home_margin|end_inning_8 | 21 | +1.7940 | [+0.8721, +3.0590] | UNDERPOWERED |
| mlb_ingame | CRPS | total_runs|end_inning_3 | 49 | +0.0891 | [-0.3032, +0.5064] | UNDERPOWERED |
| mlb_ingame | CRPS | total_runs|end_inning_5 | 54 | +0.2604 | [-0.0526, +0.5751] | UNDERPOWERED |
| mlb_ingame | CRPS | total_runs|end_inning_6 | 54 | +0.6392 | [+0.3174, +0.9632] | MODEL SHARPER |
| mlb_ingame | CRPS | total_runs|end_inning_7 | 55 | +0.7327 | [+0.4131, +1.0810] | MODEL SHARPER |
| mlb_ingame | CRPS | total_runs|end_inning_8 | 49 | +1.4201 | [+0.9659, +1.8977] | MODEL SHARPER |
| mlb_pregame | CRPS (total_runs) | pregame | 300 | -0.0523 | [-0.1657, +0.0482] | UNDERPOWERED |
| soccer_intl | winprob (Brier) | home_win_prob|minute_60 | 22 | -0.0931 | [-0.1570, -0.0336] | UNDERPOWERED |
| soccer_intl | winprob (Brier) | home_win_prob|minute_75 | 17 | -0.1075 | [-0.1885, -0.0316] | UNDERPOWERED |
nulls (351) outnumber confirms (168) 2.1x -- we publish our nulls
Two ways in
Entity Atlas
Descriptive cards across NBA, MLB, tennis, and soccer. Every card carries its sample floor and as-of date.
Evidence Explorer
Claim pages, each with its strongest receipt and a reproduce-on-a-fresh-clone block. Includes the retraction story: headline numbers this system refuted about itself.
Research labs
Counterfactual Explorer
Pick a game-state cell -- score margin x time -- and see the model forecast there next to the market and the realized outcome. Reads committed cells; runs no new simulation.
Market-Microstructure Lab
Outcome-free forecast-property instruments: information-arrival curves, model-market convergence, residual autocorrelation. Labelled STANDARD_INSTRUMENT.