Book Edges (Team Moneyline)

REAL model-vs-book comparison using current Bovada ML lines from matchups.json. Edge % = model_prob × book_decimal − 1, after de-vigging the book's pair to remove juice. Filter: model probability in [30%, 78%], 3%+ edge.

Today's Team ML Book Edges

Matchup · Side

Type

Model %

Calib %

Book %

Book ML

Raw Edge

Calib Edge

Kelly

Today's Total (O/U) Book Edges

Matchup · Side

Line

Model %

Calib %

Book %

Price

Raw Edge

Calib Edge

Kelly

How to read this: Edge is calculated against ACTUAL book prices (Bovada from matchups.json), not against model fair-odds. A +12% edge means the model thinks $1 on this side returns $1.12 long-run.

Caveat: Very large edges (20%+) often mean the model is overconfident, not that there's free money. The learning loop (learning_integrity + self_learn_weights) shrinks these over time as actual outcomes accumulate.

—