How we pick

Each game is scored by date-based models built on research from Medici Codex (opens in a new tab). The models look at where the game sits in its season and schedule — rest, congestion, travel, tournament stage and the calendar itself — and produce one number: the probability of the outcome.

Most games don't produce a pick. If the probability isn't strong enough, we pass on the game rather than force a call. Passing is a decision, not a gap in coverage.

Every pick is posted before the game starts, and results are kept in each brand's free record.

Picks carry a confidence label so you can see how strong the probability behind them is. A label describes the strength of the read; it is not a promise about the result.