Skip to content
All calculators

Kelly criterion calculator

The stake that maximises long-run growth — and the reason almost nobody should use all of it.

What you think the real chance is.

What the contract costs, in cents.

Most practitioners use a quarter or a half.

Optional — turns the fraction into an amount.

The result appears below and the address bar becomes a link you can share.

Result

Stake at 0.25× KellyShare of bankroll
6.8 %
That is
$68.18
Full Kelly
27.3 %
Edge per $1 staked
$0.33

Kelly assumes your probability is correct. It is an estimate, and the formula punishes overconfidence hard — which is why almost nobody sizes at full Kelly.

What the formula actually says

For a contract that pays $1, the Kelly stake simplifies to the edge divided by the profit: (your probability − price) ÷ (1 − price). If your estimate matches the market, the edge is zero and so is the stake. If the market is above your estimate, the formula returns a negative number, which means the bet is not there — not that you should size it small.

Kelly maximises the long-run growth rate of a bankroll. That is a specific objective, and it is not the same as maximising expected profit, which would tell you to bet everything on any positive edge and go broke on the first loss.

Why full Kelly is the wrong default

The formula assumes your probability is correct. It is not — it is an estimate, and Kelly is brutally sensitive to that. Overestimate your edge by a factor of two and full Kelly does not halve your growth, it can turn it negative. Fractional Kelly exists because the cost of betting too little is slow growth, while the cost of betting too much is ruin, and those two errors are not symmetric.

There is a second reason specific to prediction markets: they are not repeated draws from the same urn. Kelly's guarantee comes from the long run, and a market that resolves once does not offer one. Treat the number here as an upper bound to size down from, not a target.

Common questions

What is the Kelly criterion?

A formula for the stake that maximises the long-run growth rate of a bankroll, given a probability estimate and the odds on offer. For a binary contract that pays $1, it reduces to (probability − price) ÷ (1 − price).

Why does it say not to bet?

Because the market price is at or above your own probability estimate. Kelly returns zero or negative in that case, which means there is no edge to size — the answer is to pass, not to bet smaller.

Should I use full Kelly?

Almost certainly not. Full Kelly assumes your probability is exactly right, and it punishes overconfidence severely. A quarter or a half is the common practical choice, and it costs far less growth than the tail risk it removes.

Read more

These calculators do arithmetic on numbers you supply. They are not advice, and nothing here is a recommendation to take a position.