Backtest
Projected vs actual
Every recorded next-season prediction, scored against the realized outcome from the frozen source. Retrospective plumbing, not prospective validation: treat calibration as pending until the 2026→2027 capture completes.
Score summary
| Role | N | Model Brier (lower is better) | Baseline Brier (0.5) | Edge |
|---|---|---|---|---|
| batter | 22 | 0.305 | 0.250 | -0.055 no edge claimed |
| pitcher | 21 | 0.280 | 0.250 | -0.030 no edge claimed |
| all | 43 | 0.293 | 0.250 | -0.043 no edge claimed |
Coverage: 43 of 50 cards scored, 7 excluded (no recording, no oracle, or no recorded Noul). Lower Brier is better; edge is baseline minus model.
Reliability buckets
| Role | P bucket | N | Mean P | Actual rate |
|---|---|---|---|---|
| batter | 0.2-0.4 | 15 | 0.349 | 1.000 |
| batter | 0.4-0.6 | 7 | 0.443 | 1.000 |
| pitcher | 0.0-0.2 | 5 | 0.178 | 1.000 |
| pitcher | 0.2-0.4 | 16 | 0.239 | 1.000 |
Every scored prediction
| Card | Projected | Actual | Loss |
|---|---|---|---|
| batter:acunaro01:2023 | 0.410 | happened | 0.348 |
| batter:altuvjo01:2022 | 0.360 | didn't | 0.130 |
| batter:alvaryo01:2022 | 0.450 | happened | 0.303 |
| batter:arozara01:2024 | 0.290 | didn't | 0.084 |
| batter:bellico01:2023 | 0.330 | happened | 0.449 |
| batter:bettsmo01:2023 | 0.370 | happened | 0.397 |
| batter:bogaexa01:2022 | 0.370 | didn't | 0.137 |
| batter:carroco02:2023 | 0.330 | happened | 0.449 |
| batter:freemfr01:2023 | 0.390 | happened | 0.372 |
| batter:goldspa01:2022 | 0.420 | happened | 0.336 |
| batter:guerrvl02:2021 | 0.390 | happened | 0.372 |
| batter:hendegu01:2024 | 0.370 | happened | 0.397 |
| batter:judgeaa01:2022 | 0.460 | happened | 0.292 |
| batter:judgeaa01:2024 | 0.520 | didn't | 0.270 |
| batter:ohtansh01:2024 | 0.410 | didn't | 0.168 |
| batter:ramirjo01:2022 | 0.350 | happened | 0.423 |
| batter:rodriju01:2023 | 0.310 | happened | 0.476 |
| batter:schwaky01:2022 | 0.340 | didn't | 0.116 |
| batter:semiema01:2021 | 0.360 | happened | 0.410 |
| batter:sotoju01:2024 | 0.430 | happened | 0.325 |
| batter:turnetr01:2023 | 0.290 | didn't | 0.084 |
| batter:wittbo02:2024 | 0.390 | happened | 0.372 |
| pitcher:alcansa01:2022 | 0.240 | happened | 0.578 |
| pitcher:biebesh01:2020 | 0.220 | happened | 0.608 |
| pitcher:burneco01:2021 | 0.240 | happened | 0.578 |
| pitcher:colege01:2023 | 0.240 | didn't | 0.058 |
| pitcher:degroja01:2020 | 0.210 | didn't | 0.044 |
| pitcher:degroja01:2021 | 0.240 | happened | 0.578 |
| pitcher:galleza01:2023 | 0.180 | didn't | 0.032 |
| pitcher:imanash01:2024 | 0.280 | happened | 0.518 |
| pitcher:kershcl01:2022 | 0.230 | happened | 0.593 |
| pitcher:ohtansh01:2022 | 0.210 | happened | 0.624 |
| pitcher:rayro02:2021 | 0.280 | didn't | 0.078 |
| pitcher:salech01:2024 | 0.200 | happened | 0.640 |
| pitcher:skenepa01:2024 | 0.260 | didn't | 0.068 |
| pitcher:skubata01:2024 | 0.190 | didn't | 0.036 |
| pitcher:snellbl01:2023 | 0.270 | didn't | 0.073 |
| pitcher:sotogr01:2024 | 0.260 | didn't | 0.068 |
| pitcher:steelju01:2023 | 0.180 | didn't | 0.032 |
| pitcher:valdefr01:2022 | 0.170 | didn't | 0.029 |
| pitcher:verlaju01:2022 | 0.240 | happened | 0.578 |
| pitcher:webblo01:2023 | 0.170 | didn't | 0.029 |
| pitcher:wheelza01:2024 | 0.210 | didn't | 0.044 |