Free toolskelly
The Kelly criterion converts an edge into a stake: the fraction of bankroll that maximizes long-run compound growth if your win probability is exactly right. It is a sizing formula, not a betting system — when the edge is zero or negative, Kelly's answer is to bet nothing.
Kelly optimizes growth ONLY if this number is right. Estimate error is why fractional Kelly exists.
Kelly fraction
Full Kelly
5.50%
f* = (bp − q) / b — % of bankroll at full Kelly
Half Kelly stake
$27.50
2.75% of your $1000.00 bankroll
Verdict
Sized
Positive expectation at your probability — sized, not endorsed
Full Kelly maximizes expected log-growth under a probability you know exactly. In practice your estimate carries error, and overbetting an error is far more damaging than underbetting — which is why half or quarter Kelly is the standard practical choice.
With net decimal odds b (decimal − 1), win probability p, and q = 1 − p:
Why fractional? Full Kelly assumes zero error in p. Real estimates carry error, and Kelly punishes overbetting asymmetrically — staking twice the true optimum reduces long-run growth to zero. Betting half or a quarter of Kelly gives up a little theoretical growth for a lot of protection against being wrong, which is why it is the standard practical choice.
Where should p come from? Not from hope. Compute the market's own fair probability with the no-vig calculator first — if your estimate barely differs from the de-vigged market, Kelly's honest answer is a very small bet, or none. And if you gamble, keep it entertainment-sized: responsible-gambling resources here.
More free tools
These calculators run the exact helpers the Blitzen terminal uses on every board. See them applied to real markets — and how we grade everything against the close.