Calculator · Free · No signup
Kelly criterion & expected value calculator
Expected value tells you whether a bet is worth making; the Kelly criterion tells you how much of your bankroll to put on it. Enter a win chance and a payout to see both, and compare how full Kelly, half Kelly and flat betting grow a bankroll when you do have an edge.
EV & Kelly Calculator
Free toolThe win chance is your estimate. Kelly is only as good as that number: overestimate your edge and full Kelly overbets. Nothing here makes a negative-EV casino game profitable.
Play it for real
Lower the edge you play against with rakeback
Expected value of a bet
Expected value (EV) is the average result per unit staked if you repeated the bet many times. With win probability p and a payout multiplier m (decimal odds, stake included), EV per $1 = p × m − 1. A 55% chance at odds of 1.95 gives 0.55 × 1.95 − 1 = +0.0725, a 7.25% edge. A 49.5% dice roll paying 2× gives 0.99 − 1 = −0.01, the 1% house edge.
The implied probability of a payout is 1 ÷ m. If your honest estimate of the win chance is higher than the implied probability, the bet has positive EV.
The Kelly criterion formula
Kelly picks the stake that maximises the long-run growth rate of your bankroll. With net odds b = m − 1 and loss probability q = 1 − p, the Kelly fraction is f* = (b × p − q) ÷ b, which simplifies to (p × m − 1) ÷ (m − 1). For the 55% at 1.95 example that is 0.0725 ÷ 0.95 = 7.63% of your current bankroll.
When EV is zero or negative the formula returns zero or a negative number, which means do not bet. Every casino original, slot and table game has a built-in house edge, so Kelly's answer for all of them is 0%, no matter which multiplier you choose.
Full Kelly, half Kelly and flat betting
The chart plots median bankroll paths. For a fraction f of bankroll per bet, the expected log growth per bet is g = p × ln(1 + f × b) + q × ln(1 − f), so the median bankroll after n bets is B × e^(n × g). Flat betting a fixed dollar amount grows linearly at stake × EV per bet.
Full Kelly grows fastest in the long run but has deep drawdowns: there is a 50% chance of falling to half your bankroll at some point. Half Kelly gives up about a quarter of the growth rate in exchange for much smaller swings, and it protects you if your edge estimate is too optimistic. Betting more than twice the Kelly fraction gives negative growth even with a real edge. Press Simulate a run to see one random sequence of outcomes applied to all three strategies.
Where positive EV actually exists
Sports bettors find +EV when their probability model beats the bookmaker's line. In a casino, the only levers are external: rakeback, deposit bonuses with light wagering, and promotions. Those reduce your expected loss but rarely flip a game to positive EV on their own.
FAQ
EV & Kelly Calculator questions
How do you calculate expected value in betting?
Multiply the win probability by the decimal payout and subtract 1. A positive result is your average profit per $1 staked; a negative one is your average loss.
What is the Kelly criterion?
A staking formula that maximises long-run bankroll growth: bet (p × m − 1) ÷ (m − 1) of your current bankroll, where p is the win chance and m the decimal payout.
What does Kelly say for casino games?
Bet nothing. With a house edge the expected value is negative, so the Kelly fraction is zero or below, whatever multiplier you pick.
Why use half Kelly?
It keeps roughly 75% of full Kelly's growth rate with far smaller drawdowns, and limits the damage when your estimated edge is too high.
Is Kelly the same as Martingale?
No. Kelly bets a fixed fraction of the current bankroll and only when EV is positive. Martingale doubles after losses and has no effect on expected value.
More tools
Other free tools
Crash Simulator
Set an auto cash-out, watch the multiplier climb and bust, then run 1,000 rounds to see the real crash game odds.
◆SimulatorMines Simulator
Play a 5×5 mines board with any number of mines, see every multiplier in advance, and auto-play 1,000 rounds.
◎SimulatorPlinko Simulator
Drop balls on an 8–16 row board, compare where they land with the exact binomial odds, and check the RTP of every risk level.
⚂SimulatorDice Simulator
Set a target, see the true win chance and payout, then roll 1,000 times to watch the edge show up.