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=3,582 TENNIS_WTA N=909 TENNIS_ATP N=873 MMA_MIXED_MARTIAL_ARTS N=388 NFL_PRESEASON N=145 LALIGA N=116 NCAAF N=95 SERIEA N=87 EPL N=86 LIGUE1 N=84 UCL N=61 BUNDESLIGA N=53 NBA N=12 NCAAB N=11 NHL N=9 AMERICANFOOTBALL_NFL N=2

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=ncaaf

Overall: N = 95 · mean predicted 63.3% · actual win rate 64.2% · gap +0.9pp · Brier 0.191 · log-loss 0.548

Calibration by predicted-probability bucket.
Predicted-prob bucket	N	Mean predicted	Actual win rate	Gap (actual − predicted)	Brier
<50%	13	49.8%	61.5%	+11.8pp	0.251
50-55%	37	52.0%	45.9%	-6.0pp	0.252
55-60%	9	57.1%	44.4%	-12.6pp	0.268
60-65%	6	61.9%	66.7%	+4.8pp	0.237
65-70%	2	67.2%	50.0%	-17.2pp	0.258
70-75%	8	74.0%	87.5%	+13.5pp	0.128
80-90%	1	80.6%	100.0%	+19.4pp	0.038
90%+	19	92.2%	100.0%	+7.8pp	0.006

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.