What Is the Kelly Criterion?
The Kelly Criterion is a mathematical formula developed by John L. Kelly Jr. in 1956. Originally designed for information theory, it was quickly adopted by gamblers and investors to determine the optimal fraction of a bankroll to stake on any given bet in order to maximize long-term growth while minimizing the risk of ruin.
In sports betting, it answers one of the most important questions: not just whether to bet, but how much to bet.
The Kelly Formula
The formula is straightforward:
f* = (bp – q) ÷ b
Where:
- f* = the fraction of your bankroll to wager
- b = the net odds received on the bet (decimal odds – 1)
- p = your estimated probability of winning
- q = your estimated probability of losing (1 – p)
A Worked Example
Suppose you're betting on a football match at decimal odds of 2.10, and your research leads you to estimate a 55% chance of winning.
- b = 2.10 – 1 = 1.10
- p = 0.55
- q = 0.45
f* = (1.10 × 0.55 – 0.45) ÷ 1.10 = (0.605 – 0.45) ÷ 1.10 = 0.155 ÷ 1.10 ≈ 0.141
Kelly says to bet approximately 14.1% of your bankroll on this opportunity. If your bankroll is $1,000, that's a $141 stake.
Why Full Kelly Can Be Dangerous
While mathematically optimal, full Kelly betting is extremely aggressive and carries significant volatility risk. Key problems include:
- Probability estimates are rarely perfect. Even a small overestimation of your edge leads to overbetting and potential large drawdowns.
- Large swings are uncomfortable. Full Kelly can result in losing 30–40% of your bankroll on a bad run, making it psychologically difficult to stay disciplined.
- It assumes perfectly accurate probabilities, which almost no bettor can consistently achieve.
Fractional Kelly: The Practical Solution
Most professional bettors use Half Kelly or Quarter Kelly — simply betting 50% or 25% of the Kelly-suggested amount. This dramatically reduces variance while preserving most of the long-term growth benefits.
| Kelly Version | Stake Multiplier | Risk Level | Best For |
|---|---|---|---|
| Full Kelly | 100% | High | Mathematically optimal (rarely used) |
| Half Kelly | 50% | Moderate | Experienced bettors with accurate models |
| Quarter Kelly | 25% | Low | Conservative approach, beginners |
When to Use Kelly (and When Not To)
Kelly works best when:
- You have a well-researched, quantified edge on the bet.
- You are betting a large enough sample to realize long-term growth.
- You're disciplined enough to follow the numbers without emotional override.
Kelly is less suitable when:
- Your probability estimates are highly uncertain.
- You're betting recreationally rather than seriously tracking your edge.
- You can't handle large bankroll swings emotionally.
The Takeaway
The Kelly Criterion is a powerful tool, but it must be applied with humility. Your probability estimate is only as good as your research. Used thoughtfully — especially in its fractional form — Kelly provides a principled, mathematically grounded approach to bet sizing that far outperforms gut-feel staking over the long term.