What Is the Kelly Criterion?
The Kelly Criterion is a mathematical staking framework designed around the relationship between an estimated probability and the payoff offered. In its basic binary form, it outputs a fraction of bankroll rather than a fixed monetary amount.
Why Is Probability Error So Important?
Kelly assumes the probability input represents a real edge. If the estimate is too high, the formula can recommend a stake that is much too large. The risk is especially serious when a model is poorly calibrated or the apparent edge is small relative to estimation uncertainty.
Worked Example
With decimal odds of 2.20 and a 50% win estimate, b equals 1.20 and q equals 0.50. Kelly gives (1.20 × 0.50 − 0.50) ÷ 1.20 = 0.0833, or about 8.33% of bankroll. On a 1,000 bankroll, full Kelly is about 83.33.
What Is Fractional Kelly?
Half Kelly and quarter Kelly apply only a fraction of the full calculated stake. This reduces exposure and volatility, but it does not fix a bad probability estimate. The tool shows these values so users can see how strongly staking changes when a more conservative fraction is used.
What Happens When Kelly Is Negative?
A negative Kelly fraction means the entered probability and odds do not produce a positive edge under the formula. This tool floors the displayed stake at zero rather than suggesting a negative wager. That is a mathematical output, not an instruction about any specific market.
Does Kelly Guarantee Bankroll Growth?
No. Kelly is based on repeated decisions under assumed probabilities. Real sequences can experience substantial drawdowns, and estimated probabilities can be wrong. Full Kelly can be psychologically and financially aggressive even when assumptions are reasonable.
How Should Kelly Be Used With EV?
Both frameworks depend on the same probability-versus-price relationship. EV asks whether the simplified expectation is positive; Kelly additionally scales exposure according to edge and payoff. If the probability estimate is unreliable, both outputs become unreliable.
What Are the Tool’s Limits?
The formula shown here is the simple binary version. Correlated bets, multiple simultaneous positions, changing bankrolls, commission and complex payoff structures require more advanced treatment. Use the output for education and scenario analysis rather than automatic staking.
How Should You Interpret the Output?
The output from Kelly Criterion Calculator for Betting should be read as the mathematical result of the values entered, not as a forecast that an individual wager will win. If an input is an estimate, especially a probability estimate, test a range of plausible values rather than relying on one precise number. A small change in assumptions can materially change expected value, Kelly stake size or other derived outputs.
What Should You Record Before Using the Result?
Record the market or game, the odds or payout, the stake or bankroll, the probability estimate where applicable, and the time the inputs were observed. This makes the calculation reproducible and gives you something concrete to review after the outcome. The useful question is whether the assumptions and arithmetic were reasonable at the time, not whether one result happened to be a win.
What Is the Most Important Limitation to Keep in Mind?
The framework on this page supports a better-defined decision, but it cannot remove uncertainty. Keep the original inputs, assumptions and stake rules visible, and avoid changing the interpretation simply because the latest result was favourable or unfavourable. Where a probability, model output or operator feature is estimated or time-sensitive, recheck it before acting. The purpose of the guide is to make reasoning easier to inspect, compare and review, not to create certainty where none exists.