Stratsemble

Kelly Criterion Calculator

The Kelly criterion is the bet size, as a fraction of your account, that maximises the long-run growth rate of capital for a given edge. It is a mathematical ceiling rather than a target: the formula assumes you know your win rate and reward-to-risk exactly, so when they are only estimates — as they always are — it over-bets and invites deep drawdowns.

Formula

Full Kelly f* = W − (1 − W) ÷ R, where W is the win rate as a decimal and R is the reward-to-risk ratio. A result of zero or below means there is no edge, and Kelly's answer is to stake nothing.

Worked example: A 50% win rate at 2:1 reward-to-risk gives f* = 0.50 − 0.50 ÷ 2 = 0.25 — a full Kelly of 25% of the account, so half Kelly is 12.5%.

Kelly assumes a fixed, known edge and independent trades, so it over-bets whenever those are only estimates — which they always are. Educational tool, not advice.

Full Kelly
25.0%
of account per trade
Half Kelly
12.5%
what most people actually use
Quarter Kelly
6.3%
conservative
Why bet a fraction of Kelly
0full Kelly 25%half 13%over-betting →fraction of account bet per trade

This curve is the long-run growth rate IF your edge is exactly these inputs — which it never is. Real edges are uncertain and usually overstated, so the growth-optimal size sits below the peak; that's why practitioners bet a fraction of Kelly. Past the peak, growth falls off a cliff and over-betting destroys capital.

How to read it

Read full Kelly as a ceiling, not a target: because it assumes the edge is known exactly, any error in the win rate or payoff makes it over-bet, which is why almost nobody stakes it in full — half or quarter Kelly is the norm, keeping most of the growth with far smaller swings.

How is the Kelly fraction calculated?

Full Kelly is the win rate minus one-minus-the-win-rate divided by the reward-to-risk ratio. A 50% win rate at 2:1 comes to 25% of the account; a zero or negative result means the inputs describe no edge, and the growth-optimal stake there is nothing.

Why is full Kelly rarely used in practice?

Full Kelly assumes the win rate and payoff are known exactly. Because they are really estimates, full Kelly tends to over-bet and produce severe drawdowns, so half or quarter Kelly is the common choice — it keeps most of the long-run growth with far smaller swings.

What does a negative Kelly result mean?

It means the inputs average out to a losing system, so the growth-optimal stake is zero. No position size turns a negative edge into a winning one; the fix is a better edge, not a bigger bet.

Put this calculator on your site

A free, interactive calculator for any blog, newsletter or forum — same numbers as here, and it links back.

<iframe src="https://stratsemble.com/embed/tools/kelly-criterion-calculator" width="460" height="400" style="border:0;border-radius:12px;max-width:100%" loading="lazy" title="Kelly criterion calculator — Stratsemble"></iframe>
See which rules actually hold up

A calculator shows what a rule should do on paper. A backtest is how you find the rare few that actually beat simply buying and holding — costs on, losses shown, no look-ahead — so when you build one that survives, you'll know it isn't a fake edge, and can prove it forward before you risk real money. It won't tell you you'll win — nothing honest can.