What Is the Kelly Criterion?
The Kelly Criterion is a mathematical formula that calculates the optimal percentage of your bankroll to wager on a given bet. Developed by physicist John L. Kelly Jr. in 1956 for signal processing, it was quickly adopted by gamblers and investors who recognized its value for managing capital growth.
The core idea is simple: bet too little and you grow your bankroll slowly. Bet too much and you risk ruin, even with a genuine edge. The Kelly Criterion finds the exact stake size that maximizes long-term bankroll growth based on your edge on a bet.
The Kelly Criterion Formula
The standard Kelly formula is:
f = (bp - q) / b
Where:
- f is the fraction of your bankroll to bet
- b is the net odds received on the bet (decimal odds minus 1)
- p is your estimated probability of winning
- q is your estimated probability of losing (1 - p)
It becomes intuitive once you work through an example.
Practical Example: Kelly on a Moneyline Bet
Say you find a team listed at +110 on the moneyline. A $100 bet wins $110 in profit. In decimal odds, that is 2.10, so b = 1.10.
You have done your handicapping and believe this team has a 55% chance of winning, so p = 0.55 and q = 0.45.
Plug it in:
f = (1.10 x 0.55 - 0.45) / 1.10
f = (0.605 - 0.45) / 1.10
f = 0.155 / 1.10
f = 0.141, or about 14.1% of your bankroll
On a $1,000 bankroll, Kelly says to bet roughly $141. That number feels high because it is the mathematically aggressive maximum. More on that below.
To quickly convert between American, decimal, and fractional odds before plugging into the formula, the Odds Converter handles that instantly.
Why Edge Estimation Matters
The Kelly Criterion is only as good as your probability estimate. If you think a team has a 55% chance of winning but their true probability is 50%, you have no edge. Kelly will still output a bet size, but you will be wagering into negative expected value, which destroys bankrolls over time.
Disciplined handicappers spend more time building accurate probability models than hunting for the best number. The formula rewards accuracy and punishes overconfidence.
Before sizing any bet with Kelly, confirm your expected value first. The EV Calculator lets you input your probability estimate and the available odds to verify you are working with a positive-EV spot before committing a stake.
Fractional Kelly: The Practical Version
Full Kelly sizing is mathematically optimal under one condition: your probability estimates are perfectly accurate. In practice, they never are. A small miscalculation leads to overbetting, sharp bankroll swings, and potential ruin.
Most serious bettors use fractional Kelly. The concept is straightforward: take the Kelly output and bet a fixed fraction of it.
Common choices are:
- Half Kelly (0.5x): Bet 50% of what Kelly recommends
- Quarter Kelly (0.25x): Bet 25% of what Kelly recommends
Using the earlier example, full Kelly recommended $141 on a $1,000 bankroll. Half Kelly cuts that to roughly $70. Quarter Kelly brings it down to about $35.
Fractional Kelly reduces variance significantly. Your bankroll will not swing as violently between wins and losses. The tradeoff is a lower long-term growth rate, but that is a sensible exchange for bettors who cannot verify their edge estimates with precision.
How Bettors Choose Their Fraction
There is no universal answer. Bettors with strong confidence in their models, such as those using sharp line movement data or detailed statistical systems, may use half Kelly. Recreational bettors or those still developing their handicapping process are better served by quarter Kelly or lower.
A practical rule: if the full Kelly output makes you uncomfortable with the bet size, that is useful signal. Reduce the fraction.
A Second Example: A Spread Bet
You want to bet a side against the spread. Standard juice is -110, meaning you risk $110 to win $100. In decimal terms, b = 0.909 (net profit per dollar wagered).
You estimate a 54% chance of covering, so p = 0.54 and q = 0.46.
f = (0.909 x 0.54 - 0.46) / 0.909
f = (0.491 - 0.46) / 0.909
f = 0.031 / 0.909
f = 0.034, or about 3.4% of your bankroll
On a $1,000 bankroll, that is $34 at full Kelly, or $17 at half Kelly.
This example shows how quickly the vig erodes edges. Even at 54% against -110, the edge is modest and Kelly sizes it accordingly. Shopping for the best available line is one of the most effective ways to widen that edge. Line Whale's live odds comparison makes it easy to compare prices across books in real time.
When Not to Use Kelly
Kelly is not designed for parlays. The compounding of multiple legs inflates both the potential payout and the probability of being wrong. If you want to understand parlay math, the Parlay Calculator breaks down implied probability versus actual payout on any parlay combination.
Kelly also does not account for correlated outcomes or bets with highly uncertain edges, such as futures markets. In those situations, a flat 1-2% of bankroll is a safer default.
Key Takeaways
- The Kelly Criterion calculates the optimal bet size to maximize bankroll growth based on your estimated edge.
- The formula is f = (bp - q) / b, where b is net odds, p is your win probability, and q is your loss probability.
- Full Kelly produces aggressive bet sizes. Most bettors use half or quarter Kelly to reduce variance and protect against estimation errors.
- Your probability estimate is the foundation. An inaccurate estimate means Kelly will size your bet incorrectly regardless of how precisely you execute the math.
- Use Kelly as a guide, not a rigid rule. Pair it with line shopping, expected value analysis, and honest self-assessment of your handicapping accuracy.