The Kelly Criterion: How Much to Stake (and Why Fractional)
Once you've found an edge, the Kelly criterion tells you how much of your bankroll to stake to grow it fastest over the long run. It's elegant and optimal in theory. Used at full strength on a mis-estimated edge, it will wreck you.
The Kelly formula
For a bet with win probability p, loss probability q = 1 − p, and decimal profit b per dollar, the Kelly fraction is f* = (b · p − q) / b. That fraction of your bankroll is the stake that maximizes the expected growth rate of your money. Bet more and growth falls off and volatility explodes; bet less and you leave growth on the table.
Why full Kelly is dangerous
Kelly is optimal only if your p is exactly right. It never is, because edge estimates are noisy. And Kelly is asymmetric: overestimating your edge and overbetting is far more punishing than underbetting by the same amount. A full-Kelly stake on an edge you've overestimated leads to deep drawdowns and, often, ruin.
Fractional Kelly
The standard fix is to bet a fixed fraction of the Kelly stake, a quarter or a half. Fractional Kelly keeps most of the long-run growth while sharply cutting volatility and drawdown, and it builds in a buffer against your own estimation error. Quarter-Kelly is a common default for exactly this reason.
For several bets at once, naive per-bet Kelly overallocates, especially for correlated outcomes in the same game. Size them jointly instead.
Worked example: a Kelly stake
Frequently asked questions
What is the Kelly criterion?
The Kelly criterion is a staking formula that sizes a bet to maximize the long-run growth rate of a bankroll, given a win probability and payout.
What is the Kelly criterion formula?
For decimal profit b per dollar and win probability p, the Kelly fraction is f* = (b · p − q) / b, where q = 1 − p.
Why use fractional Kelly instead of full Kelly?
Edge estimates are noisy, and overbetting is far more damaging than underbetting. Betting a quarter or half of the Kelly stake keeps most of the growth while cutting volatility and protecting against estimation error.
How do you size multiple simultaneous bets with Kelly?
Naive per-bet Kelly overallocates when bets run at the same time, especially correlated ones. Optimize the stakes jointly to maximize expected log wealth instead.