Apply Kelly sizing, expected value, and parlay math to NFL spreads, moneylines, totals, and same-game parlays. Realistic examples using actual NFL pricing conventions.
Bills −3 (−105) where you model a 56% cover probability
Or use the full Kelly Calculator with detailed breakdown.
A 3-leg parlay using common NFL markets:
Run your own combinations through the parlay calculator — it shows true vs offered probability and the bookmaker's hold across legs.
NFL spreads default to −110 (1.91) — that's a 4.5% vig per side. Your model edge needs to clear that before any bet has positive EV.
QB passing yards over + WR receiving yards over on the same game is correlated. Bookmakers price SGPs assuming partial correlation, but their model is usually conservative — sometimes you find edges, sometimes you pay heavy juice.
NFL probabilities are noisy. Full Kelly maximizes log growth only with perfect probability estimates. Half-Kelly reduces variance ~75% with a small EV cost.
Use the no-vig calculator. If both sides are −110 the implied probabilities are 52.4% each, summing to 104.8%. Strip the 4.8% overround proportionally to get fair probabilities of 50/50.
Compare your model's probability against the implied probability of the price you're offered. EV = (your probability × decimal odds) − 1. Anything above zero is positive expected value.
Standalone parlays compound the bookmaker's vig — three −110 legs give the book ~14% hold. Only parlay if you're chasing leverage on confirmed edges or exploiting correlation that the SGP pricer mismodels.