Free, no signup. Last updated 15 August 2026.
The Kelly criterion sizes a bet in proportion to your edge. Enter your estimated chance, the odds and your bankroll, and choose a fraction. Most disciplined bettors use half or quarter Kelly, and the reason is below.
f = (b × p − q) ÷ b, where b = odds − 1, p = your probability, q = 1 − p
The result is the share of your bankroll to stake. When the formula returns zero or less, it is telling you the bet has no edge at this price and the correct stake is nothing at all.
Kelly maximises long-term growth only if your probability estimate is exactly right. It never is. Overestimate your edge by a little and full Kelly overstakes by a lot, which turns a normal losing run into a serious drawdown. Half Kelly gives up roughly a quarter of the theoretical growth and cuts the swings dramatically, which is why it is the common choice.
There is a second reason to be careful: Kelly assumes you can bet a precise fraction of a bankroll that only this bet touches. In practice you have several bets running at once, so the effective exposure is larger than any single calculation suggests.
Every pick we publish carries a stake in units, where one unit is one percent of bankroll. That is deliberately simpler than Kelly: it is flat, it is easy to follow, and it does not depend on you trusting our probability estimates to the decimal point.
The Kelly criterion is a formula that sizes each bet in proportion to your edge and the odds, in theory maximising the long-term growth rate of a bankroll. It was developed at Bell Labs in 1956 and later adopted by professional gamblers and investors.
Half Kelly is the usual choice. Full Kelly only maximises growth when your probability estimates are exactly right, and any overestimate leads to overstaking and severe drawdowns. Half Kelly sacrifices about a quarter of the theoretical growth while roughly halving the volatility.
It means the price on offer does not compensate for the risk given your estimated chance, so there is no edge and the correct stake is nothing. A negative Kelly figure means the other side of the bet is the value one.
For most bettors, yes. Flat staking removes the dependency on precise probability estimates, is far simpler to follow, and makes a published record comparable. Kelly rewards accurate estimation and punishes overconfidence.
Three value picks a day in the free Telegram, each published with its price and stake, and every one of them ends up on the public record.
Join the free Telegram →Some keep the site working, others help us see what's working (analytics) and measure ads (marketing). You choose.