College football, NFL, NHL, NBA and college basketball models graded in the open against the betting market: every game, the model's picks, and the record behind them, wins and losses both.
A model for each of college football, the NFL, the NHL, the NBA and men's college basketball, set against the betting market, with every pick graded in the open: wins, losses and the tests that failed.
Each sport has its own model. It prices every game, sets its number beside the betting line, and where the two disagree its side is the pick, locked before the start and graded in the open at that line and price. Each model's inputs, the tests it passed and failed, its limits and its version history are on its own page:
Every graded game is filed under two dozen situations - rest, travel, weather, total size, how far the model sits from the line, a quarterback change and more - and the model's picks are graded within each one over this season. Where the model has hit 60% or better in a situation (with at least 20 graded games), the trend supports the model ; where it has hit 40% or worse, the model struggles there. Either way the pick is still the model's: a card says "Trends disagree with the model. In similar situations, the model is 11–25 this season." and never tells you to take the other side. Tested over past seasons, picking sides from trends did not beat the break-even rate.
A game's pick on each market is the model's side, locked before kickoff. Until the site marked some picks 🔥 featured : . That label is retired. Its record, frozen at that date and replayed since 2024 from what was known before kickoff, is kept on the college football Methodology & archive page.
Every trend record on the site is the model's own record in that situation. A situation is listed because of its record, so it describes how the model has done there - it is not a guarantee of how it will do, and a game in two situations counts in both.
A model that is told the market's answer cannot tell you anything about the market. This one is deliberately kept independent, so the situations where it runs hot or cold against the line are information about the model rather than an echo of the market.
Weak, Moderate and Strong are the size of the gap between the model and the market, sized per sport (the glossary below lists each). A bigger gap is a bigger disagreement, not a surer thing. Which markets have shown a measured link between the gap and the result is on each sport's model page, under Validation.
Every sport's Journal grades every pick this season, by market, strength and situation. Retired records and earlier model details are on each sport's Methodology & archive page, linked from the footer.
Every term on the site, in the same words wherever it appears. Strength is sized per sport:
This site is published for informational and entertainment purposes only. Nothing on it is betting advice, a recommendation, or a solicitation to wager. Sports betting carries financial risk and is restricted or illegal in many jurisdictions; you are responsible for knowing the law where you are. If gambling is causing you harm, help is available at 1-800-GAMBLER in the United States.