Model-generated · Publicly graded

Free ML-Powered Sports Predictions

Machine-learning picks for MLB, NBA, NHL, and NCAAB — spread, total, and moneyline with a confidence rating. Every pick is graded against the final score and published, wins and losses alike.

3,346 picks graded in public

41,019 games in our history database

4 of 5 recent models rejected by our own gates

100% of results published — no removals

View the Full Ledger How the Model Works

←

Saturday, July 25, 2026

→

Confidence All Low (<60%) Medium (60-75%) High (>75%)

No Picks Match Filter

11 predictions available — try a different confidence level.

From Our Data Desk

All guides →

data analysis

Home Advantage, Measured Across 41,000 Games

We measured home win rate and margin across four sports in our history database. College basketball's home edge is enormous; MLB's is nearly zero.

our model

We Rejected 4 of Our Last 5 Models. Here's Why That's the Point

In July 2026 we tested five model candidates against strict holdout gates. Four failed. Here's the process, the numbers, and why we're publishing the losses.

data analysis

MLB Park Factors, Measured: What 1,400 Games Say About Every Ballpark

We measured run scoring in all 30 MLB parks across the 2026 season. Coors isn't even the highest — here's what actually matters for totals bettors.