A hit rate is the share of graded picks that won. It's the first number people look at and the easiest to misread.
Break-even is 52.4%, not 50%
At standard -110 juice you need to win 52.38% of bets just to break even, because a loss costs 110 to a win's 100. So:
- Under ~52% — losing after vig.
- 53–55% — a real, sustainable edge. This is what good looks like over hundreds of bets.
- 58%+ over a large sample — rare, and worth being skeptical about (small sample, cherry-picked lines, or soft books).
Props sold at plus money change the math — a 45% hit rate at +150 is profitable.
How big a sample you need
A 55% true edge still shows a losing week routinely. Rough guide:
200 picks: a trend, wide error bars
500+ picks: the hit rate is starting to mean something
1,000+ picks: you can trust it to a point or two
Judge a record by whether it's graded honestly and shown in full, over a real sample — not by a hot streak on a screenshot.
Calibration beats hit rate
A better question than "what's the hit rate" is "when the model says 60%, does it hit 60%?" That's calibration. A model can run at 51% and still be excellent if it's honest about which bets are close and which are strong. Props Edge publishes claimed win % next to actual on the track record page for exactly this reason.