Kelly criterion calculator
The Kelly criterion converts an edge into a stake size. It needs two honest inputs — your probability and the price — and it will just as honestly tell you when the right stake is zero.
your estimate that the bet wins
the price actually available
money set aside for betting only
| Staking plan | % of bankroll | Stake |
|---|---|---|
| Full Kelly | 10.0% | 100.00 |
| Half Kelly | 5.0% | 50.00 |
| Quarter Kelly | 2.5% | 25.00 |
At 55% probability and odds of 2.00 your edge is 10.0% — full Kelly stakes 10.0% of the bankroll.
The formula
With decimal odds d and win probability p, the Kelly fraction of bankroll is:
f* = (d × p − 1) ÷ (d − 1)
The numerator, d × p − 1, is your edge — the expected profit per unit staked. At p = 55% and d = 2.00 the edge is 0.10, so f* = 0.10 ÷ 1.00 = 10% of bankroll. When the edge is zero or negative, f* is zero or negative — which is the formula saying don't bet, and this calculator says it out loud rather than showing a small number anyway.
Why serious bettors bet less than full Kelly
Full Kelly is optimal only if your probability input is exactly right — and in football it never is. Overestimate your edge and full Kelly systematically overbets, which costs more than underbetting saves; the drawdowns it tolerates are also brutal, routinely halving a bankroll on the way to its long-run growth. Half and quarter Kelly give up a little growth for much smoother variance and protection against the input being wrong — which is why the table above shows all three, and why the fraction, not the formula, is the real decision. The full treatment — the geometry of losses, worked examples, the ways the formula magnifies self-deception — is in our Kelly criterion guide.
Garbage in, garbage out
The calculator cannot rescue a bad probability. Every number it outputs is conditional on your estimate being better than the market's — a high bar that a realistic look at prediction accuracy should temper, and that only a few hundred recorded bets can begin to demonstrate. Our own match probabilities are published with a graded public record for exactly that reason: a probability with no record behind it is an opinion with decimals.
Stake only what you can afford to lose. The Kelly criterion manages the growth of a betting bankroll; it says nothing about whether betting is a good use of the money — most of the time, for most people, it isn't.