The model's projection, next to the price.
Our model's projected side and confidence for every game, next to the best available price on that side.
Read this first. Our model is measured worse than the closing market (the "closing market" is the final price books post before a game starts — the best benchmark for a model). It went 49.8% ATS (against the spread: did the pick cover the point spread?) on 2,608 decided walk-forward games — 2,671 graded, 63 pushes — where 52.38% is the break-even needed just to profit at standard -110 odds. This page is transparency, not a pick. The durable edge on Sooth is price: take the best number and you gain expected value regardless of who wins. See the full record →
This page shows the model's projection next to the market. The full Pick Engine — the whole slate sealed before kickoff, divergence-ranked, graded in public after — lives at /picks. It is free, and it opens to everyone the moment it is sealed.
| MODEL PICK | MATCHUP | KICKOFF | CONFIDENCE | BEST PRICE | PRICE EDGE |
|---|---|---|---|---|---|
Raw data: best_lines.json. Confidence is capped at 85% (our 90%+ band was overconfident in testing). Prices move; read the timestamp and confirm at the book. Analysis only, not a sportsbook: we do not accept wagers or hold funds, here or anywhere on Sooth. Must be 21 or older where lawful. Help: 1-800-522-4700.