Kelly Criterion Calculator
Calculate the mathematically optimal bet size to maximize long-term bankroll growth based on your edge.
Calculate Optimal Bet Size
Full Kelly
Half Kelly (Recommended)
Quarter Kelly
What Is the Kelly Criterion?
The Kelly Criterion is a mathematical formula that determines the optimal bet size to maximize long-term growth while minimizing the risk of ruin. It was developed by John Kelly at Bell Labs in 1956.
📐 The Kelly Formula
Kelly % = (bp - q) / b
Where: b = decimal odds - 1, p = win probability, q = loss probability (1-p)
Why Kelly Works
Kelly betting maximizes the geometric growth rate of your bankroll. Over thousands of bets, a Kelly bettor will almost certainly end up with more money than someone using any other staking strategy—assuming accurate probability estimates.
⚠️ The Catch: You Need Accurate Probabilities
Kelly is only as good as your probability estimates. Overestimating your edge leads to overbetting and potential ruin. This is why most sharp bettors use fractional Kelly (½ or ¼).
Full Kelly vs. Fractional Kelly
Full Kelly (100%)
Maximizes long-term growth but comes with high variance. You'll experience significant bankroll swings, and a few bad beats can cut your bankroll in half quickly.
Half Kelly (50%) — Recommended
Reduces variance by ~75% while only sacrificing ~25% of growth rate. This is the sweet spot for most bettors—still aggressive but with manageable swings.
Quarter Kelly (25%)
Very conservative approach. Good for bettors who are uncertain about their edge or have limited bankrolls they can't afford to lose.
💡 Pro Tip
Start with Quarter Kelly when you're new to a betting market. As you gain confidence in your probability estimates, you can move toward Half Kelly. Few professionals use Full Kelly due to variance.
When Kelly Says "Don't Bet"
If the Kelly percentage is zero or negative, the formula is telling you there's no edge—or worse, the book has the edge. In these cases:
- Kelly = 0%: Break-even bet, no edge
- Kelly < 0%: Negative EV, don't bet
- Kelly > 25%: Recheck your probability estimate—this high an edge is rare