Experienced sports bettors know that finding an edge is only part of the equation. How much you wager when you believe you have an advantage can matter just as much as identifying the opportunity in the first place.
That is where the Kelly Criterion comes in.
Let’s look at what the Kelly Criterion is and how the strategy provides a framework to maximizing long-term bankroll growth while accounting for the possibility of losing.
What is the Kelly Criterion in Sports Betting?
The Kelly Criterion is a mathematical approach to bankroll management that calculates how much of your bankroll you should risk based on your estimated probability of winning and the odds available in the market. Rather than betting the same amount on every wager, Kelly adjusts bet size according to the strength of your perceived edge.
In other words: the larger your edge, the larger the percentage of your bankroll you can theoretically justify wagering.
When your edge is small, Kelly recommends a smaller bet. When there is no positive expected value, Kelly recommends no bet at all.
The standard Kelly formula is:
Kelly % = (bp − q) / b
Where:
- B is net decimal odds received on the wager
- P is your estimated probability of winning
- Q is your estimated probability of losing, or 1 − p
The result represents the percentage of your current bankroll that a full Kelly strategy would recommend wagering.
An Example of the Kelly Criterion in Sports Betting
Suppose you have a $10,000 bankroll and are considering a wager priced at +120.
At +120, a $100 winning wager generates $120 in profit, meaning b = 1.20.
After building your model and evaluating the market, you estimate that the wager has a 50% probability of winning.
That gives you:
- p = 0.50
- q = 0.50
- b = 1.20
Plug those numbers into the formula:
Kelly % = [(1.20 × 0.50) − 0.50] / 1.20 which equals 0.0833, or approximately 8.33%.
Under a full Kelly strategy, the recommended wager would therefore be approximately 8.33% of your bankroll, or $833 on a $10,000 bankroll.
The market price of +120 implies a break-even probability of approximately 45.45%. If your 50% estimate is accurate, you have identified a meaningful theoretical edge.
The key phrase, however, is if your estimate is accurate. Later, we will dive into how to estimate your edge, and some common mistakes with this strategy.
Full Kelly vs. Fractional Kelly Betting
Full Kelly theoretically maximizes long-term bankroll growth when your probability estimates are correct but in real life and sports, that assumption creates a major challenge because no model estimates probabilities perfectly.
Because even sophisticated bettors face uncertainty, many choose fractional Kelly instead.
Fractional Kelly simply wagers a portion of the amount recommended by the full Kelly calculation.
Using our $833 full-Kelly wager as an example, you can determine that a half Kelly would generate an approximate bet size of $417, and a quarter would mean a bet of $208.
Half Kelly would risk approximately 4.17% of the bankroll, while quarter Kelly would risk approximately 2.08%.
Fractional Kelly sacrifices some theoretical growth in exchange for lower volatility and less sensitivity to errors in probability estimates, but that tradeoff can be worth it in sports betting, where edges are uncertain and losing streaks are inevitable even when a bettor makes positive-expected-value wagers.
The Benefits of the Kelly Criterion
Kelly betting is sometimes misunderstood as an aggressive strategy because full Kelly can recommend large wagers. In practice, the framework actually encourages the opposite behavior because it forces bettors to connect bet size to measurable edge rather than confidence, emotion or recent results.
A bettor using a disciplined Kelly-based system can’t just say, “I love this bet, so I’ll bet twice as much.” The bettor must quantify the expected probability of winning, compare that probability with the market price, and size the wager accordingly.
Fractional Kelly adds another layer of protection by recognizing that the bettor’s probability estimate itself contains uncertainty.
This can help professional and sophisticated bettors scale wagers according to perceived edge, avoid risking the same amount on opportunities of different quality, reduce arbitrary changes in bet size, adjust wager sizes automatically as bankrolls rise or fall, control exposure during inevitable periods of variance, and avoid wagering if they can’t identify positive expected value
Both full Kelly and fractional Kelly are systematic ways to decide how much risk may be justified by a particular edge.
How Do You Estimate Your Edge?
Speaking of edge, estimating your edge is the hardest part of using the Kelly Criterion effectively.
The formula itself is simple, but determining “p,” your true probability of winning, isn’t.
Maybe a sportsbook offers -110 on a particular side. The break-even probability at -110 is approximately 52.38%.
If your model estimates that the team actually has a 55% chance of covering, you believe you have an edge over the offered price.
But where does that 55% come from?
Advanced bettors may develop probability estimates using things like historical performance and larger statistical samples, player and team efficiency metrics, injuries and lineup changes, matchup-specific data, and more.
Using all this information, the goal is to determine whether the probability implied by the available odds differs enough from your own estimate to create a positive expected value.
Market Price vs. Your Estimated Probability
Imagine that you make a team’s fair probability of winning 60%.
Fair decimal odds for a 60% probability would be:
1 / 0.60 = 1.67
That corresponds to roughly -150 American odds.
If the market offers that team at -130, you’re getting a better price than your estimate of fair value suggests.
If the market instead offers -170, your model may still predict that the team is more likely to win than lose, but that doesn’t make it a profitable wager. The price has moved beyond the fair value indicated by your estimate.
That distinction between predicting winners and identifying value is fundamental to using Kelly correctly.
Common Mistakes When Using the Kelly Criterion
Speaking of using Kelly correctly, let’s look at some common mistakes bettors face when using this strategy and how you can avoid them.
False Precision
The Kelly Criterion is only as useful as the probabilities entered into it.
If you believe a wager wins 58% of the time when its true probability is closer to 52%, the formula may recommend risking significantly more money than your actual edge justifies. This is known as false precision.
A sophisticated model can produce a probability such as 56.7%, but the number of decimal places doesn’t necessarily mean that the underlying estimate is correct. Bettors still need to account for model error, changing information and uncertainty.
Fractional Kelly can help reduce the consequences of those errors, but it cannot transform an inaccurate model into a profitable betting strategy.
Overbetting Your Bankroll
If a model says the optimal wager is 3% of your bankroll, wagering 6% doesn’t mean accepting twice as much risk for twice as much potential growth. Overbetting can substantially increase volatility while reducing the long-term growth characteristics that make Kelly attractive in the first place.
Repeatedly betting beyond your calculated edge can put your bankroll at risk and the problem gets even worse when bettors increase wager sizes after losses. A bettor who abandons bankroll rules to chase a losing streak is no longer following a Kelly strategy, even if the original wagers were sized using the formula.
Not Changing Your Bet Size Based on Your Bankroll
Because Kelly calculates wagers as a percentage of bankroll, the dollar amount you risk should adjust as your bankroll changes.
Let’s say quarter Kelly tells you to wager 2% of your bankroll.
With a $20,000 bankroll, that would be $400. If you lose and your bankroll falls to $15,000, 2% is now $300. If your bankroll grows to $25,000, 2% is now $500.
This dynamic changing helps reduce exposure during drawdowns and allows wager sizes to scale upward as the bankroll grows.
Using Kelly Criterion With Circa Sports
A Kelly-based strategy depends heavily on price because even a strong projection can become a poor wager when the odds move far enough.
This makes line evaluation especially important.
When betting through Circa Sports, bettors can compare the available price with their own fair-value estimates before calculating an appropriate wager size. Rather than deciding how much they want to win and working backward, disciplined bettors can start with probability, identify whether the available odds offer an edge and then determine how much of their bankroll that edge justifies risking.
For advanced bettors, that process might look like this:
- Build an independent probability estimate.
- Convert the available Circa Sports odds into break-even probability.
- Compare the market probability with your estimated fair probability.
- Pass on the wager if there is no sufficient edge.
- Calculate the full Kelly recommendation when an edge exists.
- Apply a predetermined fraction, such as half or quarter Kelly.
- Size the wager using your current bankroll.
- Record the bet, price and eventual closing line.
For sophisticated sports bettors, Kelly is one of the most useful frameworks for thinking about bet sizing. Full Kelly offers a mathematically aggressive approach to maximizing theoretical long-term growth, while fractional Kelly gives bettors a way to account for the uncertainty inherent in real-world sports markets.
Whichever approach you use, the underlying principles remain the same: identify value, quantify your edge, respect your bankroll and let the numbers, rather than emotion, determine your exposure.
With disciplined bankroll management and careful evaluation of the odds available at Circa Sports, bettors can use the Kelly Criterion as one more tool for making systematic, price-conscious betting decisions over the long term. Download the Circa | Sports® app today to get started.