Kelly criterion calculator.
The Kelly criterion answers the question every bettor gets wrong: not what to bet, but how much. Enter the odds, your own estimate of the chance, and your bankroll, and this free calculator returns the stake that grows your money fastest over time, plus the safer half-Kelly figure and the edge you are actually betting on. If you have no edge, it tells you to bet nothing.
Size your bet
LiveYour win chance is your own honest estimate, not the bookmaker’s. The whole method depends on it being accurate.
Kelly stake (full)
10.0%
that is 100.00 of your bankroll
A negative edge shows a 0% stake, which means do not bet. Full Kelly is aggressive; most bettors use half. Calculations run in your browser; nothing is stored. For entertainment, not betting advice.
The short answer
What is the Kelly criterion?
The Kelly criterion is a formula for working out how much of your bankroll to stake on a bet, based on your edge. It stakes more when your advantage is bigger and less when it is smaller, and over the long run it grows a bankroll faster than any other staking plan, as long as your estimate of the odds is accurate.
How to use this calculator
Enter the odds on offer, your own honest estimate of the chance the bet wins, and your total bankroll. The calculator returns the full Kelly stake as a percentage and an amount, the safer half-Kelly figure, and the edge you are betting on. The edge is the key: it is the gap between your estimate and the price.
If your win chance only matches the implied probability of the odds, your edge is zero and Kelly tells you to stake nothing. The tool only suggests a bet when you genuinely rate the outcome more likely than the price does.
Where it comes from
Kelly beyond betting.
The formula was not invented for sport. Where it came from explains both its power and its limits.
John Kelly Jr, a physicist at Bell Labs, published the criterion in 1956 as a result in information theory: a way to size bets when you hold an information advantage. It was taken up by professional gamblers, most famously the blackjack card-counter and mathematician Edward Thorp, and later by investors, where sizing positions to your edge is now a standard idea.
It works anywhere you face the same situation again and again: a favourable bet, a bankroll, and a choice of how much to risk each time. Sports betting fits that mould exactly, which is why the method carries over so cleanly from the card table to the bet slip.
There is one crucial difference. A card counter can calculate their edge precisely; a sports bettor can only estimate it. That uncertainty is the entire reason to bet a fraction of Kelly and to treat its output as a ceiling, never a target. The maths is borrowed from a world of known probabilities and applied to one of educated guesses, so a margin for error is not optional.
The formula
No black box.
Here is the math.
Kelly is your edge divided by the net odds. Get the edge right and the rest is arithmetic.
# Kelly fraction (p = your win chance)
f = (decimal odds × p − 1) / (decimal odds − 1)
= edge / net odds
# Stake
stake = f × bankroll
# Half Kelly (lower variance)
stake = (f / 2) × bankrollWorked through
The formula in practice.
Three quick examples on a 1000 bankroll show how the stake follows the edge, not the odds.
Even money, a real edge. Odds of 2.00 and you rate the chance 55%. Your edge is 2.00 times 0.55 minus 1, which is 0.10, a 10% edge. Kelly is 0.10 divided by 1.00, so 10% of bankroll: a 100 full Kelly stake, or 50 at half Kelly.
Bigger odds, same stake. Odds of 3.00 and you rate it 40%. The edge is 3.00 times 0.40 minus 1, which is 0.20. Kelly is 0.20 divided by 2.00, again 10%. A bigger edge at longer odds can land on the same stake, which is the point: Kelly balances edge against risk.
No edge, no bet. Odds of 2.00 and you rate it 50%. The edge is exactly zero, so Kelly is 0% and the calculator tells you not to bet. If you rated it below 50%, the edge is negative and the answer is the same: leave it.
How fast it scales
How much Kelly tells you to stake.
At even money (odds of 2.00), here is what Kelly stakes for different reads on the same bet. Notice how quickly it climbs, which is exactly why full Kelly is so aggressive.
| Your win chance | Full Kelly stake | Half Kelly stake |
|---|---|---|
| 50% | 0% (no bet) | 0% |
| 52% | 4% | 2% |
| 55% | 10% | 5% |
| 60% | 20% | 10% |
| 65% | 30% | 15% |
| 70% | 40% | 20% |
A 70% read at even money has full Kelly betting 40% of your entire bankroll on one event. Even if you are right on average, a run of losses at that size is brutal. That is the case for betting a fraction of Kelly, covered next.
Turn it down
Full Kelly vs half Kelly.
Full Kelly grows a bankroll fastest in theory, but the swings are violent. Most serious bettors deliberately stake a fraction of it.
Full Kelly is built to maximise long-run growth, but it assumes perfect information and an infinite series of bets. In the real world it produces stomach-churning drawdowns: losing half your bankroll on a bad run is entirely normal at full Kelly, and most people cannot stick to a plan through that.
Half Kelly is the popular compromise. By staking half the recommended amount you give up only a small slice of long-run growth, around a quarter, but you cut the size of your swings dramatically. Some go further to quarter Kelly. The calculator shows the half-Kelly figure alongside the full one so you can choose your comfort level.
The logic is simple: the growth you lose by going conservative is small, and the protection you gain is large. When your edge estimate might be a little off, and it always might be, betting a fraction of Kelly is the sensible hedge.
The catch
Kelly only works if your number is right.
The formula is only as good as the win probability you feed it. Overestimate your edge and full Kelly will systematically over-bet you toward ruin.
This is the part most calculators skip. Kelly assumes your probability is the true probability. But your estimate is a guess, and bettors are famously overconfident. If you think a bet is 60% when it is really 53%, Kelly will tell you to stake far too much, and the maths that is supposed to protect your bankroll will instead drain it.
Because the penalty for over-betting is much worse than the penalty for under-betting, erring small is always the safer mistake. That is the second reason fractional Kelly is wise: it builds in a margin for the fact that your edge is uncertain. Use the implied probability behind the price as a reality check: if your number is not clearly above the price’s implied chance, you do not have a proven edge, and Kelly has nothing to size.
A real bet
A worked value bet.
Say a team is priced at +150, which is decimal 2.50, an implied 40% chance. You have done your homework and genuinely rate them 50%.
Your edge is 2.50 times 0.50 minus 1, which is 0.25, a 25% edge, because you think a 40%-priced outcome is really a coin flip. Kelly is 0.25 divided by 1.50, which is 16.7% of your bankroll. On a 1000 bankroll that is a 167 full Kelly stake.
That is a huge bet on a single game, and it shows the danger in one line: it is only correct if your 50% read is genuinely right. Drop your estimate to 45% and Kelly falls to about 8%, half the stake. This is why most bettors would take the half-Kelly figure here, around 83, and why a disciplined process for estimating probabilities matters more than the formula itself.
The hard part
How to estimate your win probability.
The formula is the easy bit. The number it depends on, your true chance of winning, is the hard bit, and the one that decides whether Kelly helps or hurts.
Start from the market, not your gut. The no-vig price across the major bookmakers is a remarkably strong estimate of the true probability, because it reflects huge amounts of money and information. Strip the margin from the best available odds and treat that as your baseline; your own number should differ from it only when you have a specific, defensible reason.
Build your estimate from data where you can: head-to-head records, recent form, injuries, a model if you run one. Then sanity-check it against that no-vig baseline. If your figure is wildly different, the burden is on you to explain why you know something the whole market does not.
Finally, calibrate. Keep a record of the probabilities you assign and what actually happens. If the things you call 60% come in around 60% of the time, your estimates are honest and Kelly can work. If they land closer to 50%, you are overconfident, and the fix is to shade every estimate toward the market until your record earns the right to do otherwise.
Where it goes wrong
Common mistakes with Kelly.
The formula is simple; misusing it is easy. These four errors undo most of its benefit.
Using the bookmaker’s probability
Feed in the implied probability from the odds and your edge is zero, or negative once the margin is counted, so Kelly rightly says bet nothing. Kelly needs your own estimate, and it has to beat the market’s price to produce any stake at all.
Overestimating your edge
The most expensive mistake. A few points of overconfidence in your win rate become a much larger overbet, and the damage compounds across a losing run. When in doubt, shade your probability down, not up.
Using full Kelly on parlays
Kelly assumes independent, repeated bets. Parlays and correlated markets carry far higher variance, so standard Kelly badly overstakes them. Size multiples down hard, or keep Kelly for single bets only.
Not updating your bankroll
Kelly is a percentage of your current bankroll. Recalculate as it rises and falls; staking off an old, higher balance after a losing run quietly turns a safe plan into an aggressive one.
Before you bet
An honest word on Kelly.
The Kelly criterion is the soundest staking method there is, but it is a tool for sizing bets you already believe are good value, not a way to find good bets. It cannot create an edge you do not have, and fed a wrong probability it will happily point you off a cliff. Treat it as a discipline: bet a fraction of it, keep your estimates honest, and never stake money you cannot afford to lose.
Most people who bet do not have a real, repeatable edge, and for them the right Kelly stake is zero. If betting has stopped being fun or feels out of control, free confidential help is at BeGambleAware.org.
Questions
Kelly criterion questions.
What is the Kelly criterion?
+
The Kelly criterion is a formula for sizing a bet based on your edge. It stakes a larger share of your bankroll when your advantage is bigger and less when it is smaller, and over the long run it grows a bankroll faster than any other staking plan, provided your estimate of the chances is accurate.
What is the Kelly criterion formula?
+
The Kelly fraction is (decimal odds times your win probability, minus 1) divided by (decimal odds minus 1). In plain terms it is your edge divided by the net odds. Multiply the result by your bankroll for the stake. At odds of 2.00 with a 55% chance, that is a 10% stake.
What is the difference between full and half Kelly?
+
Full Kelly stakes the exact amount the formula gives, which maximises long-run growth but produces large swings. Half Kelly stakes half of that. It gives up only a small amount of growth while cutting the size of your drawdowns sharply, which is why most experienced bettors use it.
Should I use full Kelly for sports betting?
+
Usually not. Full Kelly assumes you know the true probability exactly, which you never do in sports. Most bettors use half or quarter Kelly to protect against overestimating their edge and to keep the swings survivable. The smaller stake is the price of being human about your own accuracy.
What happens if I have no edge?
+
If your win chance only matches the implied probability of the odds, your edge is zero and Kelly stakes nothing. If you rate the outcome less likely than the price implies, the edge is negative and the calculator again says bet nothing. Kelly only suggests a stake when you have a genuine positive edge.
Can the Kelly criterion lose money?
+
Yes. Kelly only grows a bankroll if your probability estimates are accurate. If you overestimate your edge, it over-bets and can drain your bankroll faster than flat staking would. It also has big swings even when you are right, which is why fractional Kelly and honest estimates matter.
How big should my bankroll be?
+
Your bankroll should be money set aside only for betting that you can afford to lose entirely. Kelly stakes are a percentage of it, so the figure matters less than treating it as a fixed, separate pot and recalculating your stakes as it rises and falls.
Why do professionals use fractional Kelly?
+
Because their edge is never certain. Fractional Kelly, usually a half or a quarter, sacrifices a little long-run growth for a large reduction in risk, and it builds in a buffer for the fact that estimated probabilities are imperfect. The downside of over-betting is far worse than the downside of under-betting.
What is quarter Kelly?
+
Quarter Kelly stakes a quarter of the full Kelly amount. It is even more conservative than half Kelly, giving up a little more long-run growth for much smaller swings. Cautious bettors and anyone unsure of their edge often prefer it, and you can scale to it from the full and half figures shown.
Can I use the Kelly criterion for accumulators?
+
Be careful. Kelly assumes independent, repeated single bets, while an accumulator or parlay combines several outcomes into one high-variance bet. Standard Kelly overstakes them badly. If you must, use the combined odds and your combined probability, then stake only a small fraction of the result.
Where did the Kelly criterion come from?
+
It was published in 1956 by John Kelly Jr, a physicist at Bell Labs, as a result in information theory. Professional gamblers such as Edward Thorp applied it to blackjack, and investors later adopted it for sizing positions. The same logic, sizing a bet to your edge, carries straight over to sports betting.
Is the Kelly criterion better than flat staking?
+
For maximising long-run growth with a genuine edge, yes: Kelly presses harder when the edge is bigger. But it swings more and punishes a wrong estimate severely. Flat staking is simpler and steadier, and fractional Kelly is the middle ground many bettors settle on.
About the developer
Jean Borg
Jean builds and maintains every calculator on freecalculators.pro from Malta, with a focus on tools that are fast, free and show their working. This Kelly criterion calculator uses the standard Kelly formula and the figures are verified for accuracy. It is provided for education, not as betting or financial advice. Please bet responsibly. Page last updated June 2026.