Overview

The Kelly Criterion is a bankroll-sizing formula derived by John Kelly at Bell Labs in 1956. It estimates the fraction of bankroll suggested by an edge input, but it is only as trustworthy as the probability estimate behind it.

The Formula

f* = (bp - q) / b

Where f* is the bankroll fraction, b is decimal odds minus 1, p is estimated win probability, and q = 1 - p.

When To Use It

Use Kelly after you have a defensible probability estimate and want a sizing reference. Many disciplined bettors prefer fractional Kelly because model error and correlated bets can make full Kelly too aggressive.

Worked Example

A +110 price with an estimated 52% win probability suggests a full-Kelly stake near 8.4% of bankroll. A half- or quarter-Kelly version can reduce volatility and model-error risk.

Common Mistakes

• Overtrusting a small edge estimate.
• Ignoring correlation between bets.
• Updating bankroll too frequently during variance.
• Using market-implied probability as the edge input.