Find your edge.

Expected value is the average amount you'd win or lose per bet if you could make the same bet thousands of times. It's the formula sharp bettors use to decide whether a wager is worth making at all.

EV = (P_win × profit) − (P_loss × stake)

If EV is positive, the bet pays off in the long run. If negative, you're paying the sportsbook to take the action. A single positive-EV bet can still lose — but a steady stream of them is how sharps make money.

The Kelly criterion gives the mathematically optimal bet size to maximize bankroll growth — but it assumes you know your true win probability exactly. In reality, you don't. You're estimating.

Full Kelly maximizes growth when your estimate is perfect. When your estimate is off by even a little, full Kelly becomes punishingly volatile — you can hit long losing streaks that crater your bankroll.

Half Kelly captures about 75% of the growth with roughly half the variance. Quarter Kelly is even more conservative. Almost every disciplined sports bettor uses one of these instead of full Kelly.

That's the hard part of betting — and the part this calculator can't do for you. Common approaches:

• Sharp-book consensus. Pinnacle, Circa, Bookmaker — their lines are tight and assumed close to fair. Take their implied probability as your estimate.
• Your own model. If you build a predictive model (basketball regression, weather-adjusted MLB, etc.), the model spits out probabilities directly.
• Market average. Average implied probabilities across several books, then strip out the vig.

If your estimate matches the book's implied probability exactly, EV is zero. If your estimate is higher, the bet is +EV. Be honest — overconfident probability estimates are the most common way bettors lose money even when they think they're sharp.