We don't predict winners.
We predict value.
Most betting sites tell you who to pick. We publish where our model disagrees with the book's price. Then we show how every one of those disagreements actually settled.
Each sport gets its own model, built from data specific to that sport: team strength, matchups, injuries, rest, weather, and more. Built entirely separate from what the sportsbook is posting.
When our line and the market's line disagree significantly, we quantify the gap. That gap is where the model thinks the market has mispriced the game. Every one of those calls goes on the public record to be graded.
Every qualifying game is tiered by conviction and published before it's locked. We also show you the games we skip, because knowing when there's nothing to bet is part of the model, not an afterthought.
Our flagship model this season
Mike-Baseline is a 16-variable spread model (offensive and defensive rating, pace, the four factors, strength of schedule, home court, rest and travel) combined into a single rating and converted to a spread for every game. One rule decides whether it publishes a pick.
Our spread disagrees with the market's by 5 points or more, either direction. Our only published WNBA tier.
Everything below the 5-point bar. Tracked for transparency, never recommended.
A rule tested on five seasons of openers
Our CFB model projects a spread for every FBS-vs-FBS game and compares it to the game's opening line, frozen the moment the pick is made. A 2021-2025 walk-forward backtest against real opening lines is what set the two thresholds below, not a guess.
Both conditions matter. The spread cap exists because the model overstates its edge on early-season blowouts, and the backtest only holds up once that zone is excluded.
Below the edge threshold, or the game's opening spread is too lopsided for the rule to apply.
Three signals. Each one earned its slot.
NFL picks come from three independent, separately-validated rules rather than one model. Each publishes only when its own bar is cleared.
At outdoor stadiums, when the National Weather Service shows sustained wind at 12 mph or more in the forecast, captured within 48 hours of kickoff, we bet the Under on the game total. Wind measurably suppresses passing efficiency and scoring.
A separate points-projection model compares its own total to the opening total, all season, both directions (Over or Under), independent of the wind signal.
Weeks 1-4 only, vs. the opening spread. The market hasn't fully sharpened its numbers yet this early in the season. The rule is not validated against the market's closing line, so it never fires after week 4.
Paused since August 26, 2026
We paused MLB moneyline picks on August 26, 2026 while the model goes back to research. Every pick made before the pause stays public on the scoreboard, frozen at that date. Nothing was hidden or reset. Here's the rule as published, for reference:
The highest-conviction MLB tier: at least a 10% moneyline edge on the home or away team.
A home team priced above the book by a smaller margin than a Best Bet.
No edge worth betting, by this rule.
Picks resume next season
NBA picks are categorized by spread edge. Away underdogs are excluded entirely because our data showed they consistently lose money regardless of edge size. This is last season's rule, kept here for reference until picks resume.
A significant gap between our true line and the market spread.
A moderate edge where the model favors the other team than the market does.
Below threshold, or an away underdog (filtered out regardless of edge).
Capture first. Publish once it's proven.
We're starting NHL by capturing opening lines from opening night and backtesting candidate signals in parallel. That's the same discipline every other sport here went through before it ever published a pick. Nothing publishes until it clears the same bar.
One tier. Every sport.
$24.99/month or $249/year. Every pick, every sport, the full public scoreboard. Try it free for 7 days — no card required.
- Every recommended pick, across every sport we cover
- Our model line vs the book's line on every game
- Win probability breakdown — ours vs theirs
- Adjust odds to match your book, tiers update live
- Full scoreboard and results history — every pick graded publicly
- Pick alerts when official picks lock