For every bucket of predicted probability, what did the model actually hit? A well-calibrated model has actual win rate ≈ predicted prob. A negative gap means the model is overconfident in that bucket (dangerous); positive gap means it's underpredicting (safe). Rows turn red when |gap| > 5pp AND N ≥ 10.
Combined "all sports" is rarely meaningful — sports differ in market efficiency, signal availability, and base rates. Use the chips to drill into a single sport. N<50 (red) means the calibration is brittle; N≥200 (green) is trustworthy.
Filter: window=90d
| Predicted-prob bucket | N | Mean predicted | Actual win rate | Gap (actual − predicted) | Brier |
|---|---|---|---|---|---|
| <50% | 1594 | 42.7% | 39.3% | -3.4pp | 0.233 |
| 50-55% | 2229 | 52.1% | 51.2% | -0.9pp | 0.250 |
| 55-60% | 1568 | 57.4% | 55.6% | -1.8pp | 0.247 |
| 60-65% | 657 | 62.0% | 58.8% | -3.3pp | 0.243 |
| 65-70% | 532 | 67.2% | 61.8% | -5.4pp | 0.238 |
| 70-75% | 160 | 72.1% | 74.4% | +2.3pp | 0.189 |
| 75-80% | 333 | 76.6% | 75.7% | -0.9pp | 0.185 |
| 80-90% | 184 | 85.5% | 88.6% | +3.1pp | 0.101 |
| 90%+ | 158 | 93.8% | 96.8% | +3.0pp | 0.031 |