The Kelly Criterion: How Much Should You Bet?
Finding a good bet is only half the job. The other half is deciding how much to put on it. Bet too little and you leave growth on the table; bet too much and one bad run wipes you out. The Kelly Criterion is the math that answers the question.
What Kelly actually does
Kelly sizes each bet in proportion to your edge and the odds. A small edge means a small bet. A large edge means a larger one. It deliberately scales down as your advantage shrinks — which is exactly the discipline most bettors lack.
A quick example
Say your model gives a team a 55% chance to win, and the book is paying +100 (even money, b = 1). Plug it in: (1 × 0.55 − 0.45) ÷ 1 = 0.10. Kelly says stake 10% of your bankroll.
Why almost nobody bets full Kelly
Full Kelly maximizes long-run growth — but only if your win probabilities are exactly right. In real betting they never are. Your model is an estimate, and full Kelly punishes overconfidence brutally: a string of losses at 10% stakes can cut a bankroll in half fast.
That's why most disciplined bettors use half-Kelly or even quarter-Kelly. You give up a little theoretical growth in exchange for dramatically lower variance — smoother swings, far less risk of ruin, and a margin of safety against the fact that your edge estimate is fuzzy.
Kelly vs flat betting
Many bettors instead use flat unit sizing — the same amount on every play — which is simpler and perfectly fine. Kelly's advantage is that it presses harder on your best spots and pulls back on marginal ones. Its risk is that it demands an honest, accurate edge estimate. If you trust your numbers, fractional Kelly is powerful; if you don't, flat units keep you safe.
Want to size a bet to your edge? Try our Kelly calculator, and pair it with a fixed bankroll plan so no single bet can sink you.