Kelly Criterion Calculator

Find the optimal fraction of your bankroll

How it is calculated

You enter decimal odds (total payout per 1 staked, stake included), your estimated win probability in percent, and your bankroll. The tool sets b = odds − 1, p = probability ÷ 100, q = 1 − p, and computes the Kelly fraction f = (b × p − q) ÷ b. The recommended stake is bankroll × f.

Odds must be greater than 1, and probability must be above 0 and at most 100. When your edge is negative (b × p is less than q), the tool shows 0.00%, meaning the formula says not to bet at all. The result is only as good as your probability estimate; the tool cannot verify it. Many people stake only a fraction of the full Kelly amount, since small errors in p make the full amount risky.

Worked example

Decimal odds are 2.2, you estimate a 55% chance of winning, and the bankroll is $1,000. Then b = 1.2, p = 0.55 and q = 0.45. The fraction is (1.2 × 0.55 − 0.45) ÷ 1.2 = 0.21 ÷ 1.2 = 17.50%, so the stake is $175.00. A half-Kelly approach would use about 8.75%, or $87.50.

Try this example →

Common mistakes

  • Entering fractional or American odds. The field expects decimal odds such as 2.2, where 2.2 includes the returned stake.
  • Using odds minus one as the input, which understates b; the tool subtracts one itself.
  • Typing probability as 0.55 instead of 55. A value below 1 is read as a very small percentage.
  • Overestimating your win probability. The formula is sensitive to it, and an inflated p produces an oversized stake.
  • Applying it to a single bet and expecting a guaranteed result. It optimises long-run growth over many repeated bets and does not remove the risk of losing.

About this calculator

The Kelly criterion computes the bankroll fraction that maximizes long-run geometric growth: f* = p − q/b, where p is your win probability, q = 1−p, and b is the net decimal odds (payout multiple minus 1). It's derived by maximizing the expected logarithm of wealth rather than expected value directly, which is what makes it avoid the ruin risk of always betting the full edge you perceive.

Professional sports bettors and poker players use fractional Kelly (often 25–50% of the full Kelly stake) to size bets aggressively enough to compound an edge quickly while limiting the swings a full-Kelly bankroll would suffer from probability estimation errors. The same formula, reframed with expected return and variance, underlies position-sizing models used by quantitative traders and portfolio managers deciding how much capital to allocate to a single wager or investment with a known edge.

Was this helpful?

You can pick up to 3 reactions.

Comments (1)

  • Editorial team

    The answer is only as good as the win probability you feed in, and small errors there change the stake a lot. With decimal odds of 2.5 and a 50 percent chance the tool gives 16.67 percent of the bankroll, and when there is no edge it returns zero rather than a negative stake. How do you arrive at your probability estimates?

Popular calculators

All Calculators