Enter decimal odds and your own estimated win probability to see your edge, the full Kelly stake, and a fractional (half or quarter) Kelly stake — the same 25% fractional approach we use ourselves.
The Kelly criterion is a bet-sizing formula: kelly = (p × b − q) / b, where p is your estimated true win probability, b is the decimal odds minus 1 (the net odds), and q is 1 − p (the probability of losing). The output is the fraction of your bankroll that maximizes long-term (compound) growth, given that your probability estimate is correct.
Kelly's core insight is that stake size should scale with your edge — how much your estimated win probability exceeds what the odds imply — not be a flat unit regardless of confidence. A bigger edge means a bigger stake; a small edge means a small stake; no edge (or a negative one) means no bet at all.
Full Kelly is mathematically optimal but brutally volatile — it maximizes growth but also maximizes drawdown risk, because it's extremely sensitive to any error in your probability estimate. That's why almost every serious, disciplined bettor uses fractional Kelly: staking only a fraction (commonly a half or a quarter) of what full Kelly recommends. This calculator's Kelly fraction selector defaults to a quarter — the same 25% fractional Kelly we use as our own staking rule (see our glossary definition).
The Kelly criterion is a formula for sizing bets to maximize long-term bankroll growth: kelly = (p × b − q) / b, where p is your estimated true win probability, b is the decimal odds minus 1, and q is 1 − p. It only recommends a positive stake when your estimated probability gives you an edge over the odds on offer.
Full Kelly maximizes expected long-term growth but comes with extreme bankroll swings — even a correctly-estimated full-Kelly strategy can see 50%+ drawdowns along the way. Fractional Kelly (staking half, a quarter, or less of the full recommendation) trades a modest amount of growth for a large reduction in variance and drawdown risk, which is why nearly every disciplined bettor uses it instead of full Kelly.
Kelly's stake size is directly proportional to your edge, so an inflated probability estimate produces an inflated — and unjustified — stake. This is the single biggest risk of using Kelly: the formula trusts your input completely and has no built-in way to detect overconfidence. Consistently overestimating your true win probability, even by a small margin, compounds into overbetting and accelerates bankroll ruin. Using a betting fraction below full Kelly is partly a hedge against exactly this risk.
We stake our own recommendations at 25% of full Kelly, capped at 5% of bankroll per bet. It's a deliberate trade-off: quarter Kelly gives up some theoretical long-term growth compared to full Kelly, but cuts bankroll variance and drawdown depth dramatically — and because probability estimates (ours or anyone else's) are never perfectly accurate, a smaller fraction is a buffer against that estimation error, not just against ordinary bad luck.