BET SIZING • EDGE • GROWTH
Kelly Criterion Calculator
Size bets as a fraction of bankroll that maximizes expected long-run growth from your estimated edge.
FULL KELLY FRACTION—
—
—
Method, example and assumptions
Formula: with win probability p and net payoff ratio b (profit per 1 unit risked), q = 1 − p and the Kelly fraction is f* = (b·p − q) / b. It is the fraction of bankroll that maximizes expected long-run growth when the inputs are correct.
Worked example: p = 55% (0.55) and b = 2 give q = 0.45, so f* = (2 × 0.55 − 0.45) / 2 = 0.325 → Full Kelly 32.5% of bankroll, Half Kelly 16.25%, Quarter Kelly 8.125%. Expected value per 1 unit risked = 2 × 0.55 − 0.45 = 0.65.
Assumptions: p and b are estimates. Overestimating the edge inflates the suggested bet, and full Kelly assumes the exact fraction can be compounded without costs, minimum sizes or sizing constraints.
Method reviewed August 23, 2026.
How to use the Kelly criterion
The Kelly criterion sizes a bet as the fraction of bankroll that maximizes expected long-run growth given an estimated edge. Enter win probability and net payoff ratio; the calculator returns full, half and quarter Kelly fractions plus the expected value per unit risked.
Full Kelly has high variance. Many traders bet a fraction of it because fractional Kelly keeps most of the long-run growth with much smaller drawdowns.
Frequently asked questions
What is the Kelly criterion?
It sizes a bet as the fraction of bankroll that maximizes expected long-run growth given an estimated edge.
Why do traders use half or quarter Kelly?
Full Kelly has high variance; fractional Kelly keeps most of the growth with much smaller drawdowns.
What if the result is negative?
A negative Kelly means the inputs imply no positive edge; the model suggests not betting at that size.