What is the Kelly Criterion in trading?

The Kelly Criterion is a formula that calculates the optimal position size as a percentage of your trading capital to maximise long-run growth. It takes into account your win rate and the ratio of average wins to average losses. Full Kelly is often considered too aggressive; most traders use Half Kelly (50% of the Kelly output).

What is the Kelly Criterion formula?

Kelly % = W - [(1 - W) / R], where W is win rate (as a decimal) and R is the win/loss ratio (average win ÷ average loss). For example, with a 55% win rate and 1.5:1 R:R, Kelly = 0.55 - (0.45/1.5) = 0.55 - 0.30 = 25%.

Is Full Kelly too risky for trading?

Yes. Full Kelly maximises geometric growth but causes very high drawdowns. Most professional traders use Half Kelly (50% of the Kelly %) or Quarter Kelly (25%) to reduce drawdown volatility while still benefiting from optimal sizing principles.

← All tools

Kelly Criterion Calculator

Find the mathematically optimal position size to maximise account growth based on your historical win rate and reward:risk ratio.

Calculate Kelly position size

Win rate (%)

Average win ($)

Average loss ($)

Account balance (optional, $)

Why most traders should use Half Kelly

The Kelly Criterion assumes perfectly accurate estimates of win rate and R:R. In practice, a small error in your statistics means over-betting. Half Kelly captures approximately 75% of Kelly's theoretical growth advantage while cutting the variance (drawdown) by about half.

Before using Kelly sizing, ensure you have at least 100 trades of reliable data. Tradapt's analytics section shows your rolling win rate and average R across any date range, so your Kelly inputs stay accurate.

Get your real win rate and R:R from your Tradapt analytics — the inputs Kelly needs.

Start free

Related tools

Tool

Win Rate Calculator

Calculate your true win rate and expectancy.

Tool

Risk-Reward Calculator

Calculate R:R before entering trades.

Glossary

Position Sizing

How to size positions to manage risk correctly.