Model calibration audit

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.

Calibration cohort by sport (click to filter):
All sports MLB N=3,353TENNIS_WTA N=753TENNIS_ATP N=748MMA_MIXED_MARTIAL_ARTS N=407LIGUE1 N=89NFL_PRESEASON N=65NBA N=42NHL N=33LALIGA N=26SERIEA N=19EPL N=15UCL N=13NCAAB N=11CRICKET_IPL N=6BUNDESLIGA N=3

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=mlb

Overall: N = 3,353 · mean predicted 54.8% · actual win rate 49.9% · gap -4.9pp · Brier 0.253 · log-loss 0.699
Calibration by predicted-probability bucket.
Predicted-prob bucket N Mean predicted Actual win rate Gap (actual − predicted) Brier
<50% 545 45.3% 41.8% -3.4pp 0.246
50-55% 1215 52.2% 50.8% -1.5pp 0.250
55-60% 748 57.2% 54.5% -2.6pp 0.249
60-65% 687 61.4% 48.2% -13.3pp 0.266
65-70% 147 66.9% 53.7% -13.1pp 0.266
70-75% 1 72.2% 100.0% +27.8pp 0.078
75-80% 1 76.5% 100.0% +23.5pp 0.055
80-90% 9 84.0% 88.9% +4.8pp 0.100