The Kelly criterion answers the question that comes after finding value: how much should I stake? Its formula — edge divided by odds-minus-one — commits a fraction of your bankroll proportional to how big your advantage is and how likely the bet lands.
Say you have 2.20 on a coin flip that is truly 50/50: fair odds 2.00, so you hold a real edge. Kelly says stake (0.5 × 2.20 − 1) ÷ (2.20 − 1) ≈ 8.3% of your bankroll. Better odds or higher probability push the fraction up; a smaller edge shrinks it towards zero.
Two lessons hide in the formula. First: when your edge is zero, Kelly says stake zero. Betting without an edge is not entertainment with upside — it is scheduled donation, and no staking plan, martingale or "system" changes that. Second: even with a real edge, Kelly stakes are startlingly small and full Kelly is a rollercoaster. Most professionals use a quarter or half Kelly, accepting slower growth for survivable swings.
We publish edge sizes rather than stake advice — but Kelly is the honest lens for reading them: a 3% edge justifies a small, disciplined bet, not a bankroll-defining plunge. Anyone who tells you otherwise is selling variance as certainty.
See how we apply this on every match in our methodology. Methodology →
The Kelly criterion says edge divided by (odds − 1): a 5% edge at odds of 2.00 suggests 5% of bankroll. In practice edges are estimated with error, so most disciplined bettors use a fraction of Kelly — half or a quarter — or flat stakes of about 1% until the edge is proven.
Only if your edge estimates are accurate. On our own 152 real value bets, half-Kelly and flat 1% staking produced similar returns, while full Kelly amplified estimation errors into larger drawdowns. Flat staking is the safe default while a strategy is still being validated.
See these ideas at work in our public track record — every call logged before kick-off, losses included.