Kelly Criterion Calculator

Most bettors either wing their bet sizing or just bet the same amount every time. Kelly Criterion takes a different approach—it tells you exactly what percentage of your bankroll to bet based on your edge. The goal is to maximize long-term growth while keeping your risk of ruin basically at zero.

The formula was originally developed by John Kelly at Bell Labs back in 1956 for information theory. But it turned out to work perfectly for any situation where you're making repeated bets with an edge. Blackjack card counters, stock traders, poker players—they all use some version of Kelly.

In sports betting, Kelly answers the question: "I know this is a +EV bet, but how much should I actually risk?" Bet too little and you're not maximizing your edge. Bet too much and you risk going broke even with positive EV. Kelly finds the sweet spot mathematically.

The Kelly formula is: (bp - q) / b

Breaking it down: b is your decimal odds minus 1 (so 2.50 odds means b = 1.50). p is your estimated probability of winning. q is your probability of losing, which is just 1 - p.

Let's say you've got +200 odds (3.00 decimal) on something you think has a 40% chance of winning. That's b = 2.00, p = 0.40, q = 0.60. Plug it in: (2.00 × 0.40 - 0.60) / 2.00 = (0.80 - 0.60) / 2.00 = 0.10. Kelly says bet 10% of your bankroll.

If the result is negative or zero, that means there's no edge and you shouldn't bet at all. Kelly will never tell you to bet on something with negative EV—the math just doesn't work out to a positive number.

Full Kelly (betting the entire percentage it suggests) is mathematically optimal, but it leads to some wild swings in your bankroll. We're talking 20-30% drawdowns even when you're making all the right bets. Most people can't stomach that kind of volatility.

Fractional Kelly is where you bet a fraction of what Kelly recommends—typically 25% to 50%. This dramatically reduces variance. Half Kelly cuts your volatility in half but only reduces your long-term growth rate by 25%. For most people, that's a great trade-off.

There's another big reason to use fractional Kelly: you're probably not as accurate at estimating probabilities as you think. If you overestimate your edge by even a bit, full Kelly will have you overbetting and risking ruin. Fractional Kelly gives you a margin for error.

Using it on negative EV bets: Kelly only works if you actually have an edge. If you're just betting for fun or on a hunch, Kelly will tell you exactly how fast you'll go broke. The formula assumes you know what you're doing.

Not adjusting your bankroll: Your bankroll changes with every bet. Win and it goes up, lose and it goes down. You need to recalculate your bet size based on your current bankroll, not your starting bankroll. This is tedious but important.

Treating Kelly as gospel: The formula assumes your probability estimates are perfect, which they're not. It also assumes you can make infinite bets, which you can't. Real-world betting involves all kinds of factors Kelly doesn't account for, like bet limits or odds changes.

Ignoring correlation: If you're making multiple Kelly bets at the same time on correlated outcomes (like multiple player props in the same game), you're effectively over-leveraging. Kelly assumes independent trials.