Calculator
Frequently Asked Questions
What is the Kelly Criterion?
The Kelly Criterion is a mathematical formula for determining the optimal fraction of your bankroll to bet on any wager where you have a positive edge. Developed by John L. Kelly Jr. in 1956, it maximises the long-term growth rate of your bankroll. The formula is: f* = (bp − q) ÷ b, where b = decimal odds minus 1, p = probability of winning, and q = probability of losing (1 − p).
Why use fractional Kelly instead of full Kelly?
Full Kelly bets can cause very large drawdowns during normal variance, even when your edge estimate is correct. A half Kelly (50% of full Kelly) reduces drawdowns by approximately 75% while sacrificing only 25% of expected growth rate. Quarter Kelly (25%) is even more conservative and is recommended when your edge estimate is uncertain. Most professional bettors use 25–50% Kelly.
What happens if my edge estimate is wrong?
If you overestimate your win probability, Kelly will recommend bets that are too large, which can cause catastrophic losses. This is called estimation risk. Using fractional Kelly (25–50%) provides a safety buffer against edge overestimation. The Kelly Criterion assumes your probability estimate is accurate — always be conservative with your edge estimates.
What does a negative Kelly fraction mean?
A negative Kelly fraction (f* < 0) means you have no positive edge on the bet — the bet has negative expected value. The Kelly Criterion mathematically recommends not betting at all in this situation. This occurs whenever the bookmaker's implied probability exceeds your estimated true win probability.
Can Kelly Criterion be used for matched betting or exchange trading?
The Kelly Criterion is most useful for pure value betting where you have a genuine probability edge over the market. It is less relevant for matched betting where returns are near-certain regardless of outcome. For exchange trading and value betting with a real edge, Kelly sizing is an excellent tool for managing bankroll growth while controlling drawdown risk.