A formula for calculating the optimal bet size to maximize long-term bankroll growth.
The Kelly Criterion is a mathematical formula that determines the ideal percentage of your bankroll to wager on a bet with a known edge. It balances the tradeoff between betting too little (slow growth) and betting too much (risk of ruin).
Full Kelly can be aggressive, so many professionals use fractional Kelly (typically 25-50% of the recommended amount) to reduce variance while still capturing most of the growth benefits.
The formula requires knowing your true win probability and the odds being offered. In sports betting, estimating your true probability is the hard part — the Kelly Criterion only works if your probability estimates are accurate.
Your bankroll is $10,000. You identify a Celtics moneyline at +120 where your model gives Boston a 50% win probability. Full Kelly stake = (bp − q)/b where b=1.2, p=0.5, q=0.5, yielding (0.6 − 0.5)/1.2 = 8.33% of bankroll, or $833.
Most pros run quarter Kelly to absorb model error and variance — that cuts the stake to roughly $208. Full Kelly mathematically maximizes long-run log growth, but a 20% model overestimate can produce catastrophic drawdowns. Fractional Kelly gives up a sliver of growth in exchange for sleeping through a 15-bet losing streak without blowing up.
Kelly % = (bp - q) / b, where b = decimal odds - 1, p = win probability, q = 1 - p<p>Your bankroll is <strong>$10,000</strong>. You identify a Celtics moneyline at <strong>+120</strong> where your model gives Boston a <strong>50% win probability</strong>. Full Kelly stake = (bp − q)/b where b=1.2, p=0.5, q=0.5, yielding (0.6 − 0.5)/1.2 = <strong>8.33% of bankroll</strong>, or $833.</p><p>Most pros run <strong>quarter Kelly</strong> to absorb model error and variance — that cuts the stake to roughly $208. Full Kelly mathematically maximizes long-run log growth, but a 20% model overestimate can produce catastrophic drawdowns. Fractional Kelly gives up a sliver of growth in exchange for sleeping through a 15-bet losing streak without blowing up.</p>
The practice of managing your gambling funds to minimize the risk of going broke.
The average amount you can expect to win or lose per bet over time.
A standardized bet size used to track performance, typically 1-2% of your bankroll.
The probability of losing your entire bankroll, even with a positive edge.
A formula for calculating the optimal bet size to maximize long-term bankroll growth.
Kelly % = (bp - q) / b, where b = decimal odds - 1, p = win probability, q = 1 - p
<p>Your bankroll is <strong>$10,000</strong>. You identify a Celtics moneyline at <strong>+120</strong> where your model gives Boston a <strong>50% win probability</strong>. Full Kelly stake = (bp − q)/b where b=1.2, p=0.5, q=0.5, yielding (0.6 − 0.5)/1.2 = <strong>8.33% of bankroll</strong>, or $833.</p><p>Most pros run <strong>quarter Kelly</strong> to absorb model error and variance — that cuts the stake to roughly $208. Full Kelly mathematically maximizes long-run log growth, but a 20% model overestimate can produce catastrophic drawdowns. Fractional Kelly gives up a sliver of growth in exchange for sleeping through a 15-bet losing streak without blowing up.</p>
Top-rated platforms reviewed by our editorial team