📊 Kelly Criterion Calculator
Find the optimal stake size for your bankroll
What This Kelly Criterion Calculator Does
You have an estimated edge on a bet, a bankroll, and one practical question: how much should you risk? This Kelly Criterion calculator gives you the mathematically optimal answer using the formula John Kelly Jr. developed at Bell Labs in 1956. He wrote it for telephone signal noise. It turned out to be the cleanest answer to “how much of my money should I put on this”, and professional gamblers, hedge funds, and serious sports bettors have been using it ever since.
The formula is brutally honest. No edge means it returns zero or negative — don’t bet. A real edge means it might recommend stakes that look reckless to a casual punter. That’s why most experienced bettors use a fraction of full Kelly. The variance at 100% is rough on the bankroll and rougher on the soul.
The Kelly Formula
The single-bet version reads:
Kelly % = (b × p − q) ÷ b
- b = decimal odds minus 1 (the net win per unit staked)
- p = your estimated probability of winning (0 to 1)
- q = 1 − p (probability of losing)
Positive result is the fraction of bankroll to stake. Zero or negative means the bet doesn’t have positive expected value at your estimate, and Kelly says skip. Read it as a stop-light, not a target.
A Real Example
Lakers vs Warriors tonight. The market has Lakers at 2.10. Based on your model — recent form, injury reports, rest days, home court — you estimate Lakers win 60% of the time. The market implies 47.6% (1 ÷ 2.10). The 12.4% gap is your edge.
Plug those numbers into the Kelly Criterion calculator with a €1,000 bankroll:
b = 2.10 − 1 = 1.10p = 0.60q = 0.40Kelly % = (1.10 × 0.60 − 0.40) ÷ 1.10 = 0.236 = 23.6%Stake = €1,000 × 0.236 = €236
Full Kelly says €236 on a single basketball game. That’s a sharp number for one match. Half Kelly recommends €118. Quarter Kelly drops it to €59. Most professional bettors I know stake somewhere between Quarter and Half Kelly, almost never full.
Reading the Output
The Kelly Criterion calculator returns four numbers worth understanding before you size up.
Kelly %
Optimal fraction of bankroll to bet. Always between -100% and 100%. Negative or zero means don’t bet at all — at your estimate, the wager has negative expected value. Treat that as the calculator’s stop sign.
Stake Amount
Bankroll multiplied by Kelly %. The actual euro figure to wager if you trust full Kelly. Most users should mentally halve or quarter this for fractional Kelly.
Edge
Your estimated probability minus the market’s implied probability. The bigger this gap, the more aggressive Kelly gets. Below zero means you’re betting at a loss before any bookmaker margin is even applied.
Expected Value
The average return per unit staked over many repetitions of the same bet. A 12% edge at 2.10 odds returns roughly €0.13 per €1 staked over time. Sounds small. Compounded across hundreds of bets a year, it’s serious money.
Half Kelly, Quarter Kelly, and Why Pros Use Them
Full Kelly maximizes long-term bankroll growth if your probability estimates are perfectly calibrated. They aren’t. Yours, mine, nobody’s are.
A small error in probability compounds badly when you stake aggressively. Think 60% but actually 53%? Full Kelly punishes that miscalibration with bankroll drawdowns of 50% or worse during normal variance. Statistically expected. Emotionally devastating.
The fractional Kelly trade-offs:
- Half Kelly — about 75% of optimal long-term growth, ~25% lower variance
- Quarter Kelly — about 50% of optimal growth, much smoother ride
- One-Eighth Kelly — used when probability estimates are very uncertain
The math says Half Kelly captures most of the upside while cutting the worst drawdowns substantially. That’s why nearly every serious sports betting operation runs at fractional Kelly. Pure flat staking is rare, but full Kelly is rarer.
What Kelly Doesn’t Tell You
Your Probability Estimate Is Probably Wrong
Kelly is hypersensitive to the probability you input. The Kelly Criterion calculator gives a precise answer based on whatever you type. If your estimates are systematically off, the formula amplifies that error. Garbage in, garbage out — but with leverage.
Variance Is Brutal Even at Full Kelly
Even with perfectly calibrated estimates, full Kelly produces drawdowns of 50% or more on bad runs. Standard probability theory shows this is mathematically expected, not a fluke. You’ll watch your bankroll get cut in half during normal swings.
Closing Line Value Is the Real Test
If your bets consistently get worse odds at the bookmaker right before kickoff than the price you took, you have real edge. If they get better, you don’t — you’re betting against sharper money. Without consistent CLV, the Kelly Criterion calculator is computing optimal stakes from imaginary edge.
Kelly Doesn’t Solve Account Limits
Bookmakers limit accounts that bet sharp sizes consistently. Kelly recommends bigger stakes when your edge is bigger — exactly when bookies notice. Plan account management before you scale up. Surebet calculator opportunities can offset some of the variance, but won’t fix the limit problem.
Bankroll Definition Matters
The formula assumes a fixed bankroll that grows or shrinks based purely on betting outcomes. If you’re depositing fresh funds or withdrawing winnings, recalibrate. A €1,000 starting bankroll that grows to €1,500 should now use €1,500 as the input. Kelly stakes scale with the live number, not the original.
Frequently Asked Questions
What is the Kelly Criterion in betting?
A formula that calculates the mathematically optimal stake size for any bet where you have a probability estimate and decimal odds. It maximizes long-term bankroll growth, but only if your probability estimates are accurate. Most professional bettors use a fraction of full Kelly to reduce variance.
Why use fractional Kelly instead of full Kelly?
Full Kelly produces 50% or larger bankroll drawdowns even with perfect probability estimates. Half Kelly captures about 75% of the long-term growth with significantly less variance. Quarter Kelly cuts variance even more at the cost of slower growth. The trade-off almost always favours fractional Kelly for human bettors.
Can the Kelly Criterion calculator give a negative result?
Yes. If your estimated probability is below the market’s implied probability, the Kelly Criterion calculator returns a negative percentage. That means the bet has negative expected value at your estimate. Treat zero or negative as a clear skip signal, not a small bet.
What if my probability estimate is wrong?
Kelly is unforgiving when estimates are off. Small errors compound badly because the formula scales bet size with edge. This is exactly why fractional Kelly exists. If you have not tracked your edge against closing lines for at least 200 bets, start with Quarter Kelly minimum.
Do professional bettors actually use the Kelly Criterion?
Most do not use full Kelly. Quarter to Half Kelly is the standard for serious bettors and sports betting funds. Some run dynamic Kelly that adjusts based on confidence in the estimate. Pure flat staking is rare among full-time professionals because it leaves long-term growth on the table.
Does Kelly work for parlays and accumulators?
The basic Kelly Criterion calculator on this page handles single bets only. Multi-event Kelly for parlays exists but is more complex and almost always recommends much smaller stakes than naively applying single-bet Kelly to the parlay odds. As a rough rule, treat any parlay edge with extreme suspicion.
What’s the difference between Kelly and flat staking?
Flat staking bets the same amount on every wager regardless of edge. Kelly scales bet size with edge — bigger when you are more confident, zero when you have none. Over time, Kelly grows bankroll faster but with much higher short-term variance. Flat staking is simpler and easier on the nerves.
