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=3,254 NCAAF N=960 TENNIS_WTA N=668 TENNIS_ATP N=567 MMA_MIXED_MARTIAL_ARTS N=522 LALIGA N=219 AMERICANFOOTBALL_NFL N=187 SERIEA N=171 EPL N=169 NHL N=162 LIGUE1 N=147 NFL_PRESEASON N=145 BUNDESLIGA N=128 UCL N=71 CRICKET_IPL N=27 NCAAB N=11 NBA N=9

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=bundesliga

Overall: N = 128 · mean predicted 47.5% · actual win rate 39.1% · gap -8.4pp · Brier 0.253 · log-loss 0.696

Calibration by predicted-probability bucket.
Predicted-prob bucket	N	Mean predicted	Actual win rate	Gap (actual − predicted)	Brier
<50%	67	37.6%	40.3%	+2.7pp	0.220
50-55%	15	52.4%	53.3%	+1.0pp	0.246
55-60%	27	57.4%	37.0%	-20.4pp	0.274
60-65%	12	62.2%	41.7%	-20.6pp	0.284
65-70%	7	67.6%	0.0%	-67.6pp	0.457

JSON feed · Model CLV · Methodology

Calibration is computed from the prediction_history table. Bucket edges are 5pp wide (50-55%, 55-60%, etc.) with widened tails (<50%, 80-90%, 90%+) where data is sparser. Negative gap is the dangerous direction — actual win rate below stated confidence.