Kalshi Edge Calculator

Why Half-Kelly?

The Kelly Criterion finds the mathematically optimal fraction of your bankroll to bet for maximum long-run growth. But full Kelly is extremely aggressive - it assumes your probability estimate is perfect. It's not.

Half-Kelly captures most of the growth benefit while cutting the risk of ruin roughly in half. Most professional bettors use between quarter-Kelly and half-Kelly. Never bet more than your max cap - the cap is your defense against overconfidence.

How to Estimate True Probability

Your edge is only as good as your probability estimate. For weather markets: start with the NWS point forecast as a base rate, then adjust for known systematic biases (e.g., NWS runs warm in Las Vegas, marine layer suppresses LAX highs).

Calibration matters more than any single bet. Track your predictions vs outcomes. If you say 70% and it happens 50% of the time, your 70% estimates are wrong - and your edge calculations are fiction.

Use historical base rates: how often does Vegas hit a specific temp bracket in April? Add real-time signals (METAR obs, hourly forecasts) as the day approaches. The later in the day, the more information you have - and the less uncertainty.

Why This Matters

A prediction market price is a probability estimate made by the crowd. If the crowd is wrong - and you're right - that gap is money. But the crowd is often well-calibrated. Edge exists when you have information or analytical methods the crowd doesn't.

Positive EV doesn't mean you'll always win. It means that over many bets, you come out ahead. Consistent process beats any single outcome. Log every bet, track your calibration, and cut positions that aren't generating real edge.