Kelly Criterion Calculator
Enter your estimated win probability and the decimal odds offered to calculate the optimal Kelly fraction — the percentage of your bankroll this formula suggests wagering to maximize long-term growth.
Kelly Criterion Explained: Optimal Bet Sizing for Long-Term Bankroll Growth
The Kelly Criterion is a mathematical formula developed by John L. Kelly Jr. at Bell Labs in 1956. Originally designed for signal transmission problems, it was quickly recognized as a powerful tool for optimal bet sizing in gambling and investing. The formula determines the fraction of a bankroll that, when wagered repeatedly on positive-expected-value opportunities, maximizes the long-term growth rate of capital. It has since become a foundational concept in quantitative finance, sports betting, and any field where sequential decisions involve uncertain outcomes.
The Core Formula
The Kelly formula is: f* = (bp − q) / b, where f* is the fraction of the bankroll to bet, b is the net odds received on the wager (decimal odds minus 1), p is the estimated probability of winning, and q = 1 − p is the probability of losing.
For example, suppose a bet pays 2.0 in decimal odds (net odds b = 1.0) and the bettor estimates a 60% win probability. The Kelly fraction is f* = (1.0 × 0.60 − 0.40) / 1.0 = 0.20, meaning the formula suggests wagering 20% of the bankroll on this bet.
When the Kelly fraction is negative, the formula indicates the bet has a negative expected value from the bettor's perspective and should not be placed. The formula does not recommend negative-sized bets; a negative result simply signals no bet.
Expected Value and the Prerequisite for Kelly Betting
The Kelly Criterion only produces a positive bet recommendation when the expected value per unit staked is positive: EV = bp − q > 0. This is equivalent to saying the bettor's estimated probability exceeds the breakeven probability implied by the odds.
Expected value represents the average return per unit wagered across many repetitions of the same bet. An EV of +0.10 means the bettor expects to gain 10 cents on average for every dollar staked. A negative EV means the house or bookmaker has an edge, and the Kelly formula will correctly return a negative fraction indicating no bet.
The critical dependency of the Kelly formula on accurate probability estimates cannot be overstated. If a bettor overestimates their win probability, the Kelly fraction will oversize the bet. This is why many practitioners deliberately use fractional Kelly strategies.
Half-Kelly and Quarter-Kelly: Conservative Variants
The full Kelly fraction maximizes the expected logarithm of wealth — and therefore the geometric growth rate — in theory. However, it assumes the bettor's probability estimates are exactly correct, which is rarely the case in practice. Errors in probability estimation translate directly into oversized bets and higher variance.
Half-Kelly (f*/2) and Quarter-Kelly (f*/4) are commonly used conservative variants that trade some expected growth rate for substantially reduced variance. A Half-Kelly bettor still captures most of the long-term growth advantage while halving the risk of large drawdowns. Many professional sports bettors and quantitative traders favor fractional Kelly for precisely this reason.
Research in mathematical finance has shown that betting more than the Kelly fraction — overbetting — can be particularly destructive. A bettor consistently wagering double the Kelly fraction will have the same expected growth rate as someone not betting at all, despite taking on far more risk. The Kelly fraction represents the maximum threshold, not merely a recommendation to wager exactly that amount.
Decimal Odds and Net Odds
Decimal odds (also called European odds) represent the total return per unit staked, including the original stake. Odds of 2.0 mean you receive 2.0 units back for every 1 unit bet — a profit of 1.0 unit on a winner. Odds of 1.5 mean a profit of 0.50 units per unit staked.
The Kelly formula uses b, the net odds (profit per unit staked), which equals decimal odds minus 1. This conversion is straightforward: for decimal odds of 2.5, the net odds b = 1.5, meaning a winning bet returns 1.5 units of profit for every 1 unit wagered.
The breakeven probability — the minimum win rate at which a bet has zero expected value — equals 1 / decimal odds. For 2.0 odds, the breakeven probability is 50%. For 3.0 odds, it is 33.3%. The Kelly Criterion only recommends a bet when the bettor's estimated probability exceeds this breakeven threshold.
Practical Considerations and Limitations
The Kelly Criterion assumes that the bettor accurately knows the true probability of winning. In real-world applications — sports betting, poker, financial markets — this probability must be estimated, and estimation errors are inevitable. The formula is only as reliable as the probability input.
Kelly also assumes that bets are divisible and that the bankroll can be fractioned arbitrarily. In practice, minimum bet sizes, transaction costs, and market liquidity constraints all limit how closely one can implement the full Kelly strategy.
The formula further assumes that bets are independent and that the bankroll compounds continuously without constraints. In contexts with correlated outcomes — such as a portfolio of sports bets on games played simultaneously — the single-bet Kelly formula underestimates the risk of ruin, and a simultaneous Kelly approach accounting for correlation should be used instead.
This calculator is provided for educational and entertainment purposes. It demonstrates the mathematical properties of the Kelly Criterion using user-supplied inputs. It does not constitute financial, investment, or gambling advice.
Historical Context and Applications
John Kelly published the criterion in a 1956 paper titled 'A New Interpretation of Information Rate' in the Bell System Technical Journal. The paper addressed the problem of maximizing the long-run growth rate of capital given repeated bets with known probabilities — framed around a gambler receiving advance information via a noisy telegraph channel.
Edward Thorp, the mathematician and blackjack pioneer, recognized the formula's applicability to card counting and later to financial markets. Thorp's hedge fund, Princeton-Newport Partners, applied Kelly-based position sizing throughout the 1970s and 1980s, achieving an estimated 20% annualized return before fees over a 20-year period.
The Kelly Criterion is now embedded in quantitative finance under various names: the log-optimal portfolio, the growth-optimal strategy, and the capital growth criterion. It forms the theoretical foundation for dynamic portfolio allocation strategies used in algorithmic trading, where the goal is to maximize the geometric mean of returns over time rather than the arithmetic mean.
Frequently Asked Questions
What does a negative Kelly fraction mean?
A negative Kelly fraction means the bet has a negative expected value from your estimated probability. The formula indicates you should not place the bet. This occurs when your estimated win probability is lower than the breakeven probability implied by the odds — meaning the bookmaker or house has a mathematical edge over you.
Why use Half-Kelly instead of full Kelly?
The full Kelly fraction maximizes long-term growth only when probability estimates are exactly correct. In practice, estimates contain errors, and overestimating your win probability causes the full Kelly to oversize bets significantly. Half-Kelly reduces variance substantially while retaining most of the long-term growth advantage, making it a popular choice among professional bettors and quantitative traders.
How do I convert fractional or American odds to decimal odds?
Fractional odds (e.g., 3/1) convert to decimal by dividing numerator by denominator and adding 1: 3/1 = 4.0 decimal. American odds convert differently: positive odds (e.g., +300) become (odds/100) + 1 = 4.0; negative odds (e.g., −150) become (100/abs(odds)) + 1 ≈ 1.67. This calculator uses decimal odds as input.
What is the breakeven probability for a given odds value?
The breakeven probability — the minimum win rate at which expected value equals zero — is 1 divided by the decimal odds. For odds of 2.0, breakeven is 50%. For odds of 3.0, breakeven is 33.3%. Your estimated probability must exceed this threshold for the Kelly Criterion to recommend a bet.
Does the Kelly Criterion apply to investing as well as betting?
Yes. The Kelly Criterion applies to any sequential decision under uncertainty where outcomes have known or estimated probabilities and returns. In investing, it is used to determine optimal position sizes in a portfolio. However, financial returns are continuous rather than binary, so the formula is adapted using expected return and variance rather than win/loss probabilities. The core principle — size positions to maximize long-term geometric growth — remains the same.