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 · sport=tennis_wta
| Predicted-prob bucket | N | Mean predicted | Actual win rate | Gap (actual − predicted) | Brier |
|---|---|---|---|---|---|
| 50-55% | 126 | 52.5% | 55.6% | +3.1pp | 0.245 |
| 55-60% | 154 | 56.9% | 56.5% | -0.4pp | 0.245 |
| 60-65% | 65 | 62.2% | 55.4% | -6.8pp | 0.250 |
| 65-70% | 74 | 67.5% | 66.2% | -1.3pp | 0.225 |
| 70-75% | 40 | 72.9% | 75.0% | +2.1pp | 0.185 |
| 75-80% | 178 | 75.9% | 72.5% | -3.4pp | 0.203 |
| 80-90% | 80 | 87.4% | 87.5% | +0.1pp | 0.107 |
| 90%+ | 29 | 93.1% | 96.6% | +3.4pp | 0.033 |