Free Trading Tool
Kelly Criterion Calculator
Enter your strategy's win rate and win/loss ratio to instantly calculate the Kelly Criterion position size — plus the half-Kelly and quarter-Kelly fractions traders actually use in practice. No signup required.
Percentage of trades your strategy wins (0-100).
Average win ÷ average loss. Enter 1.5 for a 1:1.5 payoff.
Enter as a positive number.
Your Result
Full Kelly (f*)
—
Fraction of capital the formula says maximises long-run growth.
Half-Kelly (practical)
—
What most practitioners size with instead of full Kelly.
Quarter-Kelly
—
Ratio Used (R)
—
Always paper trade first. This calculator is an analytical tool to help you evaluate a position-sizing model before you risk capital — it is not investment advice and does not guarantee results. Sizing errors compound faster than entry and exit errors, so test any sizing rule on a paper account before trading it live.
Want the full definition, the derivation, and the common mistakes? Read the full Kelly Criterion definition →
How to Calculate the Kelly Criterion
The Kelly Criterion answers one question: given a measured edge, what fraction of capital grows an account fastest over many repeated trades without risking ruin? It needs only two numbers, both of which come out of a backtest or a trade log:
- 1. Win rate (W) — the share of trades the strategy wins, expressed as a decimal (55% becomes 0.55).
- 2. Win/loss ratio (R) — average win size divided by average loss size. A strategy whose winners average 3,000 and losers average 2,000 has R = 1.5.
f* = W − (1 − W) / R
W = win rate as a decimal
R = average win ÷ average loss
f* = fraction of capital to riskWorked example: (illustrative, not real market data) — a strategy wins 55% of its trades, and its average winner is 1.5 times the size of its average loser.
- W = 0.55, so 1 − W = 0.45
- R = 1.5
- f* = 0.55 − (0.45 / 1.5) = 0.55 − 0.30 = 0.25
Full Kelly says risk 25% of capital on each trade. Almost nobody does that — see the half-Kelly section below. Half-Kelly on this setup is 12.5%, quarter-Kelly is 6.25%.
You may also see the formula written as f* = (bp − q) / b, where b is the payoff ratio, p the win probability, and q = 1 − p. It is the same equation rearranged — both forms return an identical f* for identical inputs.
Kelly Fractions at Common Win Rates
These are the numbers the calculator above produces, shown for a few typical win rate and payoff combinations. Note the last row: below a certain win rate the formula returns a negative f*, which means the setup has no edge at all and the optimal stake is zero.
| Win Rate | Win/Loss Ratio | Full Kelly | Half-Kelly |
|---|---|---|---|
| 50% | 1:1 | 10.0% | 5.0% |
| 55% | 1:1 | 10.0% | 5.0% |
| 55% | 1:2 | 32.5% | 16.3% |
| 60% | 1:1 | 20.0% | 10.0% |
| 40% | 1:3 | 20.0% | 10.0% |
| 45% | 1:1 | No edge | No edge |
The pattern worth internalising: raising the payoff ratio moves f* far more than nudging the win rate does. A 55% win rate at 1:1 sizes at 10%, but the same win rate at 1:2 sizes at 32.5% — because Kelly is sizing the expectancy, not the hit rate.
Why Traders Use Half-Kelly
Full Kelly is optimal under one assumption that never holds in trading: that the win rate and payoff ratio you entered are exactly correct. In reality they are estimates from a finite sample, and Kelly is very sensitive to them being overstated. Feed it a win rate that is five points too generous and it hands back a position size well past the genuinely optimal one — the zone where extra size actively reduces long-run growth.
The drawdown profile is the other half of the problem. Even with perfectly accurate inputs, full Kelly is a high-volatility growth path — deep, prolonged equity drawdowns are the normal behaviour of the strategy, not a sign it has broken. Most traders abandon a correctly-sized full-Kelly system mid-drawdown, which converts a paper-optimal sizing rule into a realised loss.
Half-Kelly is the practical compromise. Betting half the calculated fraction cuts volatility and drawdown substantially while sacrificing only about 25% of the theoretical long-run growth rate. Quarter-Kelly goes further still, and is the common choice when the win-rate and payoff estimates come from a limited backtest sample.
One rule has no exception: betting more than full Kelly is mathematically proven to reduce long-run growth while increasing ruin risk. There is no confidence level in an edge that makes over-betting correct.
How Quantzee Helps
Kelly is only as honest as the two numbers you feed it, and those numbers come from a backtest. A backtest built on repainting indicators overstates both the win rate and the average win — which inflates f* and produces dangerous over-sizing in live trading. Because Quantzee's indicators (see the AI Adaptive Quant Toolkit and SuperTrend Pro+) are non-repainting, the win rate and payoff ratio in the Strategy Tester are the ones that were genuinely available in real time — which is exactly what the Kelly formula assumes you are giving it.
FAQ