Kelly Criterion Calculator
How To Use This Calculator
The Kelly Criterion helps you determine the optimal bet size to maximize long-term growth while managing risk. Enter your odds, win probability, and bankroll to see exactly how much to wager.
Input the odds from your sportsbook.
Your estimated chance of winning (use a no-vig calculator to find this).
We recommend 0.25-0.5 to reduce variance.
Enter your total betting bankroll.
What is the Kelly Criterion?
The Kelly Criterion is a mathematical formula developed by John L. Kelly Jr. in 1956. It calculates the optimal bet size based on your edge and bankroll to maximize long-term growth.
Unlike flat betting, Kelly sizing adjusts your wager based on your advantage - betting more when you have a bigger edge and less when your edge is smaller. This protects your bankroll while maximizing growth potential.
f* = (bp - q) / bWhy Use the Kelly Criterion?
Optimal long-term compound returns
Never risk too much on a single bet
Bet more when advantage is higher
Step-by-Step Kelly Calculation
You have a $5,000 bankroll. You find a bet at +150 (2.50) odds with a devigged fair probability of 45%. What should you bet?
Kelly Criterion vs Flat Betting
The core difference between Kelly and flat betting comes down to one question: do you trust your probability estimates enough to size bets accordingly? Kelly Criterion sizes each bet proportional to your edge. When you have a large edge, you bet more. When the edge is thin, you bet less. Flat betting ignores edge size entirely and wagers the same fixed amount on every play.
Kelly staking maximizes the long-term growth rate of your bankroll. Over 100 bets with a consistent 5% edge at +100 odds, a Kelly bettor would stake roughly 5% of their current bankroll on each bet and compound gains aggressively. A flat bettor wagering 2% of their starting bankroll on the same 100 bets would grow more slowly but experience smoother results. Drawdowns of 30-40% are common with full Kelly, while a flat bettor's drawdowns stay more predictable.
For practical purposes: use Kelly-based staking if you have a proven model or reliable probability estimates and can stomach the variance. Use flat betting if you are still developing your handicapping, working with small samples, or prefer simplicity. Many serious bettors start with flat staking, then graduate to fractional Kelly once they have confidence in their edge calculations.
One important nuance: Kelly assumes you know your true win probability. If your estimates are off by even a few percentage points, Kelly can oversize bets and accelerate losses. Flat betting is more forgiving of estimation errors. This is why the majority of professional sports bettors use fractional Kelly rather than full Kelly.
Fractional Kelly Explained
Full Kelly is mathematically optimal in theory but too aggressive for real-world sports betting. It assumes your probability estimates are perfectly accurate, which they never are. A small overestimate of your edge leads to overbetting, and overbetting is the fastest way to destroy a bankroll. This is why fractional Kelly exists: you take the Kelly-recommended stake and multiply it by a fraction, typically between 0.25 and 0.50.
The most common fractions are quarter Kelly (25%), third Kelly (33%), and half Kelly (50%). If full Kelly says to bet 8% of your bankroll, half Kelly bets 4%, third Kelly bets 2.67%, and quarter Kelly bets 2%. Half Kelly achieves roughly 75% of the growth rate of full Kelly while cutting variance nearly in half. Quarter Kelly achieves about 50% of full Kelly's growth rate but makes the risk of a significant drawdown almost negligible.
Professional bettor consensus leans toward quarter to half Kelly. This range accounts for the inevitable inaccuracy in probability estimates, protects against correlated losses, and keeps drawdowns manageable. If you are new to Kelly staking, start at quarter Kelly. As you build a track record and gain confidence in your probability model, you can move toward third or half Kelly. Very few sharp bettors use full Kelly in practice.
Kelly Criterion for Multiple Simultaneous Bets
A common problem arises when you find several value bets at the same time. If you calculate full Kelly independently for five simultaneous bets and each recommends 6-8% of your bankroll, your total exposure could reach 30-40%. That level of exposure on concurrent bets creates serious risk, especially if the outcomes are correlated.
The theoretically correct solution is simultaneous Kelly, which optimizes stake allocation across all concurrent bets while accounting for correlations and total portfolio risk. A simpler approach is proportional reduction: calculate Kelly for each bet independently, sum the total exposure, and if it exceeds your desired limit (say 15-20% of bankroll), reduce each bet proportionally.
The most practical solution is the simplest: use fractional Kelly. If you habitually stake at quarter or third Kelly, your independent bet sizes are already small enough that even five or six simultaneous bets rarely create dangerous total exposure. Quarter Kelly on five concurrent bets with moderate edges typically totals 8-12% of your bankroll, which is a reasonable level of risk.
Common Mistakes to Avoid
Full Kelly assumes your probability estimate is perfectly accurate. Even small errors lead to significant overbetting. Always use fractional Kelly (0.25-0.5) to build in a margin of safety.
Kelly sizing is based on your current bankroll. If you're using yesterday's bankroll figure after a big loss, you're overbetting. Recalculate bankroll regularly — ideally before every bet.
Kelly only works for +EV bets. If your estimated probability is wrong and the bet is actually -EV, Kelly will amplify your losses. Ensure you have a genuine edge before sizing with Kelly.
If you have 10 Kelly-sized bets running at once, you're risking much more than Kelly intended. Reduce your Kelly fraction based on how many open bets you have, or use a simultaneous Kelly formula.
Frequently Asked Questions
What is fractional Kelly betting?
Fractional Kelly means using a percentage of the full Kelly recommendation. Half Kelly (0.5) or Quarter Kelly (0.25) reduces variance while still capturing most of the expected growth. Most professional bettors use fractional Kelly to protect against estimation errors in their edge calculations.
How do I find my win probability?
Use our No-Vig Calculator to remove the bookmaker's margin from sharp book odds. The resulting "fair" probability is your estimated win probability for Kelly calculations.
What if Kelly suggests betting 0%?
If the Kelly formula returns zero or negative, it means the bet has no edge or is negative EV. Don't place the bet - look for better opportunities where you have a genuine advantage.
Why not use full Kelly?
Full Kelly is mathematically optimal but very aggressive. Small errors in probability estimates can lead to overbetting. Fractional Kelly provides nearly the same growth with significantly less variance and lower risk of large drawdowns.
What happens if you overbet using Kelly?
Overbetting (using more than full Kelly) actually decreases long-term growth and dramatically increases risk of ruin. At 2x Kelly, your expected growth drops to zero. Beyond that, you are mathematically expected to go broke over time. This is why fractional Kelly is strongly recommended.
Can I use the Kelly Criterion for parlays?
The standard Kelly formula is for single bets. For parlays, calculate the combined edge and odds of all legs, then apply Kelly to the parlay as a whole. Because parlays have higher variance and compounding estimation errors, use a smaller Kelly fraction (0.1-0.25).
How often should I recalculate my Kelly bet size?
Recalculate before every bet. Since Kelly is based on your current bankroll, your recommended stake changes as your bankroll grows or shrinks. Many bettors update their bankroll figure daily or weekly.
What is fractional Kelly and why should I use it?
Fractional Kelly means taking the stake recommended by the Kelly formula and betting only a fraction of it, typically 25% to 50%. You should use it because full Kelly assumes your probability estimates are perfect, which they are not. Fractional Kelly sacrifices a small amount of theoretical growth in exchange for significantly lower variance and better protection against estimation errors.
How do I use Kelly Criterion for NBA or NFL betting?
The Kelly Criterion works the same regardless of sport. Enter the odds from your sportsbook and your estimated win probability. The challenge is accurately estimating that probability, which requires factoring in injuries, rest days, matchup data, and market movement. Many bettors derive their estimates by devigging sharp book odds like Pinnacle.
What happens if Kelly says to bet 0%?
A Kelly stake of 0% means you have no edge on the bet. Your estimated win probability, after accounting for the odds offered, produces zero or negative expected value. This is a signal to skip the bet entirely. If it returns a negative number, the sportsbook has the edge and you should not bet at all.
Can I use Kelly Criterion with a betting exchange?
Yes. Kelly works on exchanges like Betfair, Smarkets, or BetDaq. The key difference is that exchange odds already reflect close-to-true market probabilities, so edges tend to be smaller. When calculating Kelly for exchange bets, account for the exchange commission (typically 2-5%) in your effective odds before running the formula.
Kelly Criterion vs fixed percentage staking: which is better?
Fixed percentage staking is simpler but ignores edge size. You bet the same amount on a 1% edge as on a 10% edge, which is suboptimal. Kelly allocates more to bigger edges, maximizing bankroll growth over time. If you have reliable probability estimates, fractional Kelly will outperform fixed percentage staking in the long run.
Pro Tip: Start with Quarter Kelly
New to Kelly betting? Start with a 0.25 multiplier. This captures 75% of the expected growth of full Kelly while reducing variance by 50%. As you gain confidence in your edge calculations, you can gradually increase to half Kelly (0.5). Most professionals never go beyond half Kelly.