Kelly Criterion: Optimal Bet Sizing for Maximum Growth
The Kelly Criterion formula is f* = (bp - q) / b. Learn what each variable means, how to use it for trading and betting, why half-Kelly beats full Kelly, and common mistakes.
The Kelly Criterion Formula
The Kelly Criterion formula is: f* = (bp - q) / b. Here f* is the fraction of your bankroll you should bet, b is the net odds you receive (e.g. 2-to-1 means b = 2), p is your probability of winning, and q is your probability of losing (which is just 1 - p). Example: if a bet pays 2-to-1 and you win 60% of the time, Kelly says bet f* = (2 × 0.6 - 0.4) / 2 = 40% of your bankroll. The formula was developed by John Kelly at Bell Labs in 1956 and later applied to investing by Ed Thorp. For investing with continuous outcomes, it adapts to: f* = expected excess return / variance of returns. The key insight is that Kelly maximizes long-term compound growth — not the expected value of any single bet. Maximizing expected value can lead to overbetting and eventual ruin; Kelly avoids this by balancing growth against risk.
Edge vs Odds: What Drives Kelly Sizing
The Kelly fraction is determined by the relationship between your edge (how often and by how much you win) and the odds (the payoff structure). A large edge with symmetric odds produces a large Kelly bet. A small edge, even with favorable odds, produces a tiny Kelly fraction. Critically, Kelly bets zero when there is no edge — it never gambles. This reveals an uncomfortable truth: most retail investors who cannot demonstrate a quantifiable edge should, under Kelly logic, bet conservatively or not at all. The formula also shows that edge quality matters more than edge size. A small but reliable edge with low variance justifies a larger Kelly fraction than a large but uncertain edge with high variance. The variance in the denominator penalizes uncertainty, naturally producing smaller bets when you are less sure of your advantage.
Free charts, alerts, and screeners for every asset discussed on this page. Used by 50M+ traders.
Fractional Kelly for Real-World Use
Full Kelly sizing produces maximum long-term growth but at the cost of enormous volatility and drawdowns along the way. The theoretical maximum drawdown of a full Kelly strategy approaches 100% given enough time. For any real investor with finite patience and finite capital, full Kelly is too aggressive. The standard practical recommendation is half-Kelly — betting exactly half the formula's output. Half-Kelly achieves approximately 75% of the growth rate of full Kelly while dramatically reducing variance and drawdown risk. Quarter-Kelly reduces drawdowns even further at a still-acceptable growth rate reduction. Most professional investors and fund managers who use Kelly-based sizing operate at one-quarter to one-half Kelly. The reduction in volatility and peak-to-trough drawdowns at fractional Kelly makes it psychologically sustainable in a way that full Kelly is not.
Why Most Investors Over-Bet
The most common portfolio management error, by a wide margin, is overbetting — allocating more capital to individual positions than the Kelly Criterion would recommend given the actual edge and variance. Overconfidence leads investors to overestimate their edge while underestimating variance, producing position sizes far larger than optimal. Betting more than the Kelly fraction does not just increase volatility — it actually reduces long-term compound growth. This is counterintuitive: betting more aggressively produces lower terminal wealth over long horizons. The region above full Kelly is called the overbetting zone, where expected geometric return declines despite larger expected arithmetic return. Many retail portfolios operate deep in this overbetting zone with concentrated positions representing 20 to 50 percent of their capital, sizes that Kelly would only justify for strategies with enormous, reliable edges that almost no individual investor possesses.
Kelly in a Portfolio Context
Extending Kelly to multi-asset portfolios requires accounting for correlations between positions. The portfolio Kelly solution uses a covariance matrix to determine optimal weights across all positions simultaneously, accounting for the diversification benefit of imperfectly correlated bets. In practice, the multi-asset Kelly often recommends more diversified portfolios than investors expect because diversification allows higher total exposure at the same risk level. The Kelly framework also provides a natural way to think about leverage: if the optimal Kelly portfolio sums to more than 100% allocation, it implies that leverage is mathematically justified given the edges and correlations, though fractional Kelly would scale this down. The framework's greatest practical value may not be the precise numbers it produces but the discipline it imposes: it forces you to explicitly quantify your edge, estimate variance, and size positions accordingly rather than relying on intuition.
Free charts, alerts, and screeners for every asset discussed on this page. Used by 50M+ traders.
Frequently Asked Questions
What is the Kelly Criterion formula?▼
The Kelly Criterion formula is f* = (bp - q) / b, where f* is the optimal fraction of your bankroll to bet, b is the odds received (net odds), p is the probability of winning, and q is the probability of losing (1 - p). It was developed by John Kelly at Bell Labs in 1956 and later applied to investing by Ed Thorp.
How do you use the Kelly Criterion for trading?▼
For trading and investing, the Kelly formula adapts to: f* = expected excess return / variance of returns. You estimate your edge (expected return above the risk-free rate) and divide by the variance of those returns. Most practitioners use half-Kelly or quarter-Kelly to reduce volatility while maintaining most of the growth rate.
What is fractional Kelly and why use it?▼
Fractional Kelly means betting a fraction of what the full Kelly formula recommends — typically half-Kelly (50%) or quarter-Kelly (25%). Full Kelly maximizes long-term growth but with extreme drawdowns approaching 100%. Half-Kelly achieves roughly 75% of the growth rate while dramatically cutting drawdown risk, making it psychologically sustainable for real investors.
What happens if you bet more than the Kelly Criterion suggests?▼
Betting above the Kelly fraction is called overbetting and it actually reduces your long-term compound growth — not just increases risk. This is counterintuitive: larger bets produce lower terminal wealth over long horizons. The region above full Kelly is where expected geometric return declines despite larger expected arithmetic return.
Is the Kelly Criterion good for sports betting?▼
Yes, the Kelly Criterion was originally designed for binary bets and works well for sports betting when you have a genuine edge. The formula f* = (bp - q) / b directly applies to fixed-odds betting. However, most sports bettors overestimate their edge — if you cannot demonstrate a quantifiable edge, Kelly recommends betting zero.
Related Topics
Track These Indicators Live
The Alphameter synthesizes six macro indicators into a single regime score — updated daily. See the current reading and full indicator breakdown on the dashboard.
Get notified when the market regime changes
Regime alerts + weekly macro brief. Unsubscribe anytime.