Guide / staking

Kelly criterion explained

The Kelly criterion is a staking method that turns estimated edge into a suggested bet size. It matters because finding value is only half the problem in betting. The other half is deciding how much of the bankroll to risk when the edge appears.

What the Kelly criterion is

The Kelly criterion is a formula used to size a bet when the bettor believes the offered odds are better than the fair odds. In plain language, it asks: if you really have an edge, how much of your bankroll should you stake?

This makes Kelly a natural companion to value betting, expected value, and implied probability. Those pages help estimate whether an edge exists. Kelly helps turn that estimate into a sizing decision.

The basic Kelly formula

In its classic betting form, the Kelly fraction is:

Kelly fraction = (bp - q) / b

Here, b is the net decimal payout minus 1, p is your estimated probability of winning, and q is 1 minus p. The result is the fraction of bankroll the formula suggests staking.

For decimal odds, readers often rewrite the problem as: estimate your fair probability first, then compare it to the offered price. If there is no real edge, Kelly should not tell you to bet at all.

A simple Kelly example

Suppose a sportsbook offers decimal odds of 2.20. That means the net payout b is 1.20. If you estimate the true win chance at 50%, then p = 0.50 and q = 0.50.

The Kelly result becomes:

(1.20 x 0.50 - 0.50) / 1.20 = 0.0833

That would suggest staking roughly 8.3% of bankroll at full Kelly. The number looks clean, but the practical warning is obvious: if your probability estimate is even slightly wrong, the staking advice can change dramatically.

Kelly is only as good as the edge estimate behind it. Bad probability estimates create bad staking advice.

Why many bettors use half Kelly or smaller

Full Kelly can be aggressive in real-world betting, especially when estimates are uncertain and variance is high. That is why many practical bettors prefer half Kelly, quarter Kelly, or even looser flat-stake systems.

Fractional Kelly keeps the same direction as the formula while reducing exposure to estimation error and drawdown stress. In plain language, it accepts that real betting models are usually noisier than neat textbook examples.

What practical Kelly use really looks like

Kelly works best when the bettor has a disciplined way to estimate edge and a reason to trust that estimate over time. Without that, the formula can become a confidence amplifier rather than a bankroll tool.

That is why Kelly belongs more naturally next to closing line value and process-tracking pages than next to promotional betting content. It is a staking framework for readers who are already thinking about long-run edge, not a shortcut to smarter gambling.

Common Kelly mistakes

  • Treating a shaky opinion as a precise probability estimate.
  • Using full Kelly with no tolerance for drawdowns.
  • Forgetting that edge estimation error is often larger than the formula suggests.
  • Using Kelly on combo bets where pricing clarity is weaker.
  • Thinking Kelly creates edge instead of only sizing an existing one.

Where to go next on WikiOne