Learn
The Kelly Criterion Guide
Optimal bet sizing, explained without the calculus
The formula
Why full Kelly is too aggressive in practice
Kelly with multiple simultaneous bets
A worked Kelly example
Frequently asked
Does Kelly work if I don't know my true probability?
Not perfectly. Kelly is only optimal if the probability estimate is correct. If your estimate is off by even a few percentage points, full Kelly dramatically overshoots. That's why pros always use fractional Kelly - it's tolerant of estimation error.
What's the difference between half-Kelly and quarter-Kelly?
Half-Kelly stakes 50% of the full Kelly amount; quarter-Kelly stakes 25%. Half-Kelly captures ~75% of full Kelly's long-run growth with ~50% of the variance; quarter-Kelly captures ~43% of growth with ~25% of variance. Most pros use 0.25–0.5 Kelly depending on model confidence.
Can I use Kelly for parlays and accumulators?
Yes, but carefully. For a parlay, compute the joint probability (multiply the individual probabilities if they're independent) and the joint odds (multiply the individual odds), then apply Kelly to the result. Correlated parlays need a joint-probability estimate, not a multiplication.
What happens if Kelly says to bet more than I have?
Cap the stake at your bankroll. Kelly should never ask you to bet more than 100% of the bankroll - if it does, your edge estimate is almost certainly wrong.
Does BetsPlug auto-calculate Kelly?
Members get a Kelly sizing tool inside the dashboard that factors in correlation between simultaneous bets. The free preview shows confidence only; you can manually apply quarter-Kelly if you want a rough starting point.