The Origins of Optimal Betting
Developed in 1956 by John Kelly, a researcher at Bell Labs, the Kelly Criterion is a mathematical formula that determines the optimal size of a series of bets to maximize the long-term growth rate of wealth.
It was quickly adopted by legendary gamblers (like Ed Thorp to beat blackjack and roulette) and subsequently by quantitative hedge funds. It remains the gold standard for defining risk when both the probability of winning and the payout odds are known.
In prediction markets like KeyCandle, where payouts are fixed and you can estimate your historical win rate, the Kelly formula is perfectly applicable to optimizing position sizing.
The Kelly Formula Explained
The basic formula is: f* = (bp - q) / b
Where: f* is the fraction of the bankroll to bet. 'b' is the net odds received on the wager (the payout multiplier minus 1; e.g., if payout is 1.8x, b = 0.8). 'p' is the probability of winning (your historical win rate, e.g., 0.60 for 60%). 'q' is the probability of losing (1 - p, e.g., 0.40).
If we plug in a 60% win rate and a 1.8x payout: f* = (0.8 * 0.60 - 0.40) / 0.8 = (0.48 - 0.40) / 0.8 = 0.08 / 0.8 = 0.10. The Kelly formula suggests you should risk 10% of your bankroll on this setup to mathematically maximize long-term growth.
The Danger of "Full Kelly"
While mathematically optimal for maximizing growth to infinity, betting "Full Kelly" (the exact percentage spit out by the formula) creates an violently volatile equity curve in the short term.
If the formula suggests a 10% bet, a streak of three losses will result in a near 30% drawdown of the total bankroll. Most human traders cannot psychologically survive a 30% drawdown without altering their strategy or going on tilt.
Furthermore, the Kelly Criterion assumes you know your win rate with 100% accuracy. But historical win rates in trading are estimates subject to variance. If you overestimate your edge and bet Full Kelly, the mathematics dramatically accelerate your path to ruin.
The Professional Standard: Half-Kelly
Because of the psychological brutality and estimation errors inherent in Full Kelly, almost all professional practitioners use "Fractional Kelly," most commonly Half-Kelly or Quarter-Kelly.
By dividing the Kelly output by 2 (e.g., betting 5% instead of 10%), you sacrifice about 25% of the theoretical maximum growth rate, but you reduce the mathematical variance (drawdowns and volatility) by a massive 75%.
Half-Kelly provides the holy grail of risk management: the vast majority of the upside compounding power with a fraction of the psychological pain of drawdowns.
Applying Kelly as a Reality Check
The most powerful use of the Kelly Criterion for retail traders is as an objective boundary map. If Kelly says your optimal risk is 4%, and you are currently slapping 10% on your predictions, you are mathematically guaranteed to eventually destroy your account through variance, regardless of how good your charting is.
Conversely, if Kelly says your mathematical optimum is 5%, but you are only risking 0.5% out of fear, you mathematically leaving vast amounts of compounding growth on the table.
Do the math. Take your last 100 predictions. Calculate your win rate. Know your payout structure. Run the formula. If the result is negative, you do not have an edge; sizing down is irrelevant, you must fix the strategy. If it is positive, you now have the mathematical architecture to size for maximum long-term success.