How it works

The Kelly Criterion answers one question: what fraction of your bankroll should you stake to maximize long-run growth? Bet too little and you leave growth on the table; bet too much and variance eventually ruins you. Kelly is the exact balance point.

# b = net decimal odds (decimal odds minus 1)
# p = your estimated win probability (0 to 1), q = 1 - p

f* = (b × p - q) / b = p - q / b

# if f* ≤ 0 the bet has no edge, so stake 0
stake = f* × bankroll

A few things worth being explicit about:

Kelly is only as good as your probability estimate. The formula assumes p is accurate. If your p is too high, full Kelly overbets badly. This is the single biggest reason to size down.
Half-Kelly and Quarter-Kelly keep most of the growth of full Kelly with far less variance and drawdown. Halving the fraction roughly quarters the variance while giving up only a small slice of long-run growth. That is why they are the common defaults.
A zero or negative fraction means no edge. When your win probability is below the odds' implied probability, Kelly says stake nothing. The calculator shows $0 rather than a negative bet.
Edge vs price is your probability minus the implied probability of the odds (1 / decimal odds). Positive edge is a necessary condition for a positive Kelly stake.

An honest caveat

This tool does not tell you your true win probability, and that is the hard part. It takes your estimate at face value. If you are guessing at p, treat full Kelly as an upper bound and lean toward Quarter-Kelly. Real edges come from a well-calibrated probability, which usually means devigged sharp lines, not gut feel.

Where the probability comes from. This is the same math ParlayAPI's /v1/ev and /v1/devig endpoints run at scale across 45+ sources, turning real bookmaker prices into fair, devigged win probabilities you can feed straight into a sizing model like this one.

See the API docs →

Kelly Criterion Calculator

Your bet

Recommended stake

How it works

An honest caveat