Model Record

Closing line value is the headline, not profit. Over a few hundred bets, results are mostly variance — a good model can be losing and a bad one winning for a long stretch. CLV measures whether the model got a better price than the market's final word, which converges roughly five times sooner and is the earliest honest evidence of an edge.

Picks are reconstructed from stored odds snapshots rather than logged by hand, so nothing can be added or removed after the fact.

This is the market's record, not the model's. Every side of every line in every stored snapshot is graded against the final score, so the sample cannot be narrowed to the ones that worked. The column that matters is the gap between how often a line converted and how often its price said it would.