Kelly Criterion Calculator
See what the Kelly criterion — a formula used by professional traders — suggests for your position size.
Worked Example (Illustrative Only)
These examples demonstrate how the formulas work using hypothetical assumptions.
The contract is priced at 25¢ on Kalshi. You check the NWS forecast and believe there's a 40% chance. Your Kalshi balance is $500.
The market thinks there's only a 25% chance NYC hits 90°F, but the forecast suggests 40%. That's a 15-point edge on a cheap contract with 3:1 payout odds. Full Kelly says risk 20% of your $500 balance, but Half Kelly — the safer choice — suggests $50. If you're right, 200 shares at 25¢ each would pay $150 profit. If you're wrong, you lose $50 — just 10% of your bankroll.
Kelly Criterion Explained
Everything you need to know about using the Kelly criterion for prediction market position sizing.
What is the Kelly criterion?
The Kelly criterion is a mathematical formula that calculates the theoretically optimal fraction of your bankroll to risk on a single bet. It was developed by John Kelly at Bell Labs in 1956 and is widely used by professional gamblers, traders, and hedge fund managers. The formula balances two competing goals: betting enough to grow your bankroll when you have an edge, but not so much that a loss wipes you out.
How does the Kelly formula work for prediction markets?
In prediction markets, the Kelly formula compares your personal probability estimate to the market price. If you think an event has a 75% chance of happening and the YES contract costs 65¢, you have a 10-point edge. Kelly uses that edge and the payout odds to calculate what fraction of your balance to risk. The bigger your edge and the better the odds, the more Kelly tells you to bet. If you have no edge (your estimate equals the market price), Kelly says don't bet at all.
What is Half Kelly, and why should I use it?
Half Kelly means betting exactly half of what the full Kelly formula recommends. Most professional traders use Half Kelly or less because the full Kelly amount assumes your probability estimate is perfectly accurate — which it almost never is. Half Kelly sacrifices about 25% of the theoretical growth rate but cuts your variance (the size of your swings) roughly in half. It's a much smoother ride for nearly the same long-term result.
What happens if I bet more than Kelly recommends?
Betting more than the full Kelly amount actually reduces your expected long-term growth rate, even if you have a genuine edge. This is called "overbetting" and it's one of the most common mistakes in prediction markets. At twice the Kelly amount, your expected growth drops to zero. Beyond that, you're mathematically expected to go broke over time. This is why the calculator shows a warning when your position size exceeds 10–15% of your balance.
Why does the calculator say 'no edge detected'?
The calculator shows 'no edge' when your probability estimate is at or below the market's implied probability. For example, if the contract costs 65¢ (implying 65% probability) and you estimate 60%, you don't have an edge on the YES side — the market already prices it higher than you believe. In this case, you might have an edge on the NO side instead. The calculator will suggest flipping to NO when this happens.
Should I always bet the Kelly amount?
No. The Kelly criterion is a theoretical upper bound, not a target. In practice, your probability estimates contain uncertainty, markets can move against you, and your bankroll may need to survive many bets. Most practitioners use Quarter Kelly to Half Kelly. The right fraction depends on how confident you are in your probability estimate, how many simultaneous positions you hold, and your personal risk tolerance.
Does Kelly account for platform fees?
This calculator includes Kalshi and Polymarket fee math in its calculations. Fees reduce your effective payout, which means Kelly will recommend a smaller position than it would without fees. At low edge values (1–3 points), fees can eat most or all of your expected profit, causing Kelly to recommend very small or zero-size positions. This is actually useful information — it tells you the trade isn't worth making after costs.
How is Kelly different from expected value?
Expected value tells you whether a trade is profitable on average. Kelly tells you how much to risk on that trade. A trade can have positive expected value but still be a bad idea if you bet too much of your bankroll on it. Think of EV as the 'should I trade?' question and Kelly as the 'how much?' question. They work together — use the EV calculator first to confirm an edge exists, then use Kelly to size the position.