Kelly Criterion Calculator
Enter your win rate and your average win ÷ average loss to get the growth-optimal fraction of the account to stake — full Kelly as the ceiling, half-Kelly as the recommendation, and both translated into dollars.
Your edge, measured
Percent of trades that close as winners, from your trade log.
1.5 means winners average 1.5× losers.
Risk capital used to turn the Kelly percent into dollar figures.
Full Kelly says 25.00% of the account ($2,500.00); practitioners bet half-Kelly — 12.50% ($1,250.00) — because estimation error at full Kelly is punishing.
- Full Kelly
- 25.00%
- Full Kelly in dollars
- $2,500.00
- Half-Kelly
- 12.50%
- Half-Kelly in dollars
- $1,250.00
How the Kelly fraction is calculated
The Kelly criterion comes from John L. Kelly Jr., a Bell Labs physicist who showed in 1956 that a bettor with an information edge maximizes the long-run growth rate of capital by staking a fixed fraction of it on every bet. Ed Thorp took the result out of information theory and into practice — first at the blackjack tables, then in the markets — and it has been the reference point for bet sizing with a measured edge ever since.
f* = W − (1 − W) ÷ R- f*
- Fraction of capital to stake per bet
- W
- Win probability (your win rate)
- R
- Win/loss ratio — average win ÷ average loss
Full Kelly is a ceiling, not a target. The formula assumes W and R are known exactly, but yours are estimates from a finite trade log — and the penalty for overshooting is severe, because betting above the true Kelly fraction lowers growth while raising volatility. Full Kelly also rides through brutal drawdowns even when the inputs are right. Betting half-Kelly captures roughly 75% of the maximum growth rate at about half the volatility, which is why practitioners treat it as the working number.
A 55% win rate with 1.5× payoffs
Suppose your trade log shows a 55% win rate and winners that average 1.5× your losers. Kelly says f* = 0.55 − 0.45 ÷ 1.5 = 0.25: full Kelly is 25% of the account, and half-Kelly is 12.5%. On a $10,000 account that is a $2,500 full-Kelly ceiling and a $1,250 half-Kelly stake.
Now flip it: a 40% win rate with a 1.0 win/loss ratio gives f* = 0.40 − 0.60 ÷ 1.0 = −0.20 — full Kelly is −20%. A negative Kelly fraction means the system has no positive expectancy, and the growth-optimal bet is nothing at all.
This is an educational calculation based only on the values you provide. It is not trading advice.
What this calculator assumes
- Bets are independent with fixed odds — the casino setting Kelly was derived for. Markets are not exactly that: payoffs vary, trades correlate, and regimes change.
- Your win rate and win/loss ratio are honest estimates from your own trade log. Garbage in, garbage out — optimistic inputs produce dangerously oversized stakes.
- Kelly sizes risk capital — money you can afford to trade through deep drawdowns — not your life savings.
- This is an educational sizing aid, not trading or investment advice.
Kelly criterion FAQ
Why bet half-Kelly?
Because the cost of overbetting is asymmetric. Your win rate and payoff ratio are estimates, and if they are even slightly optimistic, full Kelly quietly becomes over-Kelly — lower growth and higher volatility at the same time. Half-Kelly keeps roughly 75% of the maximum growth rate at about half the volatility, and it leaves a margin of safety for the estimation error that is always there.
What if Kelly is negative?
A negative Kelly fraction means the system has no positive expectancy: at that win rate and payoff ratio, every dollar staked loses money on average, so the optimal bet is zero. No position-sizing rule can rescue a losing system — fix the system, not the size.
Kelly vs the 1% rule?
They answer different questions. The 1–2% fixed-fractional rule in our position size calculator is a floor for unproven systems — it caps damage while you are still finding out whether you have an edge. Kelly needs a measured edge to work with: a win rate and payoff ratio estimated from a real sample, ideally 50 or more logged trades. Until you have that, size with the 1% rule.
Sources and review notes
- J. L. Kelly Jr., “A New Interpretation of Information Rate,” Bell System Technical Journal (1956)
- Edward O. Thorp, Fortune’s Formula chapters (W. Poundstone)
- Investor.gov — What is risk?
Methodology last checked Jul 16, 2026. Formula implementation is covered by deterministic unit tests. No financial professional review is claimed yet.