When the model disagrees most with the market, the market is right
mlb: at largest disagreement (>=.10, n=8530), market usually right (model_closer_rate=0.199, model_brier=0.1778 vs market_brier=0.1298). soccer_intl: at largest disagreement (>=.10, n=2291), market usually right (model_closer_rate=0.105,...
Data integrity
These MLB/soccer numbers are revision 2, computed on the segment-clean corpus (2026-09-16). Revision 1 values are withdrawn and kept in the regeneration receipt. Segmentation excluded 51,635 of the 78,986 MLB ticks (65.4 percent); the specifically identified label mismatches were a smaller 27,076 (34.3 percent).
- market_disagreement_profile
How to read this figure
Each bucket is a disagreement size; a model_closer_rate below 0.5 means the market beats the model in that bucket more often than not.
Continue with the data
Open the interactive analysis: Model and market Brier by disagreement bucketWhat it means
Bucketing predictions by how far the model sits from the market is a direct test of who is right when they disagree. At small disagreements the two are effectively tied on Brier (0.129506 vs 0.129087). But every step up in disagreement, the model gets worse relative to the market -- at the widest gap it lands closer to the truth only about 20% of the time. Large model-market disagreements show the model is wrong.
Caveats and confounds
The same pattern holds in soccer (model_closer_rate 0.1048 at the widest gap), so it is not an MLB quirk -- though both rest on in-game corpora only.