What the Kelly Formula Actually Solves
Imagine you have $1,000, a horse race, and a 40% chance the favorite will win at 2.5-to-1 odds. The core problem? How much of that bankroll to wager to maximize growth without going bust.
Step-by-Step Calculation
Step one: Convert odds to decimal. 2.5-to-1 becomes 3.5 decimal. Step two: Identify win probability (p) = 0.40, lose probability (q) = 0.60. Step three: Plug into Kelly: f = (bp – q) / b, where b = odds – 1 = 2.5.
Do the math. (2.5 × 0.40) – 0.60 = 1.00 – 0.60 = 0.40. Divide by 2.5 → 0.16. That’s 16% of your $1,000, i.e., $160. Simple.
Why Not Bet the Whole Stake?
Betting $1,000 would double your money if you win, but a 60% chance of losing it all erodes long-term growth. Kelly’s 16% keeps variance in check, letting compounding work its magic.
Applying the Formula in Real Time
Look: You see a race, odds shift to 3.0-to-1, confidence climbs to 45%. Recalculate fast. b = 3.0, (3.0 × 0.45) – 0.55 = 1.35 – 0.55 = 0.80. Divide by 3.0 → 0.267, so now you wager 26.7% of the current bankroll.
Notice the bankroll updates after each race, whether you win or lose. The Kelly stake adapts, never static.
Partial Kelly for Safety Nets
Here is the deal: Full Kelly is aggressive. Many pros halve it — 50% Kelly — cutting volatility dramatically. In our example, half Kelly would be 8% of $1,000, i.e., $80. Still profitable, far less risky.
Common Pitfalls
Don’t feed the formula garbage. If your win probability estimate is off by even 5%, the optimal fraction swings wildly. Always hedge with solid data, not gut feeling.
Bottom Line
Use the Kelly criterion to turn a single race into a disciplined bankroll strategy. Plug in the odds, the win probability, and let the math dictate the stake. And here is why: it maximizes exponential growth while keeping ruin at bay. For a concrete walkthrough, check out this kelly criterion worked example.
Start small, adjust constantly, and never chase losses. That’s it.