Optimal bet sizing.

Kelly Criterion calculates optimal bet size to maximise long-term geometric growth: f* = (bp - q) / b, where b = win/loss ratio, p = win probability, q = loss probability. Used by Warren Buffett, Ed Thorp, professional gamblers and sports bettors. Tells you how much of capital to risk per opportunity.

Example: 60% win probability, win 200 / lose 100 (2:1 ratio). Kelly = (2 × 0.60 - 0.40) / 2 = 0.40 = 40% of bankroll. Aggressive but mathematically optimal for long-term growth. Most professionals use Half Kelly (20% in this example) - sacrifices some growth for dramatically lower volatility and drawdown risk.

Kelly limitations: assumes you know exact probabilities (rarely true in markets). Over-betting (above Kelly) causes ruin even with positive edge. Half Kelly is standard practice - retains 75% of growth at 25% of volatility. Kelly is impossible without edge: if expected value is negative (no edge), Kelly returns negative or zero - don't bet. Kelly is for repeated bets, not one-time decisions.

A worked example

Try the defaults: win probability of 60%, win payoff of 200, loss amount of 100. The tool returns 40.00%. You can adjust any input and the result updates as you type — no submit button, no reload. That's the real power here: seeing how sensitive the output is to one or two assumptions.

What moves the number most

The result responds to Win Probability %, Win Payoff, and Loss Amount. Not every input has equal weight. Flip one at a time toward extreme values to feel which ones move the needle most for your situation.

The formula behind this

Kelly fraction = (b×p - q) / b. Half Kelly = full Kelly / 2 (recommended for safety). Everything the calculator does is shown in the formula box below, so you can check the math against your own spreadsheet if you want.

Using this well

Treat the output as one point on a wider map. Run it three times — a pessimistic case, a central case, and a stretch case — and plan against the pessimistic one. That habit alone separates people who stick with an investment plan from those who bail at the first wobble.

What this doesn't capture

Steady-rate math ignores real-world volatility. Actual returns are lumpy; sequence-of-returns risk matters most in drawdown; fees and taxes drag on compound growth; and behaviour changes in drawdowns can reduce outcomes below the projection. Treat the number as one scenario, not a forecast.