The Kelly criterion is a formula that calculates the bet size which maximizes long-run capital growth, given your edge and odds. In simple form it weighs your win probability against your payoff ratio to produce a percentage of capital to risk. It's the mathematical answer to "how much should I bet?"

How it works

Kelly says the optimal fraction to risk rises with your edge and your payoff, and falls as your edge shrinks. A big, reliable edge justifies larger bets; a thin edge justifies small ones. With no edge, Kelly says bet nothing - which is itself a useful discipline.

Why full Kelly is brutal

Here's the catch: "optimal for growth" is not "comfortable to trade." Full Kelly produces enormous drawdowns - swings that would shake almost anyone out or trigger a ruinous losing streak if your edge estimate is even slightly off. And edge estimates are always uncertain. Full Kelly assumes you know your odds precisely; you never do.

Full Kelly is mathematically optimal and practically unbearable. That's why nobody trades it.

Fractional Kelly

The practical answer is fractional Kelly - betting a half or a quarter of what the formula suggests. This sacrifices a little theoretical growth for a large reduction in volatility and blow-up risk, and it cushions the error in your edge estimate. It's the same logic behind conservative position sizing: survival first, growth second. Bet small enough that being wrong about your edge can't end you.