The Complete Guide to Kelly Criterion Bet Sizing on Polymarket
The Kelly Criterion is the gold standard for bankroll management among professional bettors, hedge fund managers, and prediction market traders. Originally derived in 1956 by John L. Kelly Jr. — a researcher at AT&T's Bell Labs studying signal noise in long-distance telephone lines — the formula was quickly adopted by gambler Edward Thorp, who used it to beat blackjack and later run one of the most successful quantitative hedge funds in history. Today, every serious Polymarket trader, sports bettor, and options trader uses some variant of Kelly to decide how much capital to allocate to each opportunity.
Why optimal bet sizing matters more than picking winners
Most beginners obsess over win rate. They believe that if they can just identify enough winning bets, profits will follow automatically. In reality, two traders with identical edges can have wildly different outcomes simply because of how they size their bets. Bet too small and you leave compounding growth on the table. Bet too large and a string of normal losing variance can wipe you out before your edge has time to materialize. The Kelly Criterion solves this by telling you the exact fraction of bankroll that maximizes the long-run logarithmic growth of your wealth.
The Kelly formula explained step by step
The full Kelly formula is f* = (bp − q) / b, where f* is the fraction of your bankroll to wager, b is the net decimal odds (decimal odds minus one), p is your estimated probability of winning, and q is the probability of losing (1 − p). The numerator bp − q is your expected value per dollar staked. The denominator b normalizes that EV by the payoff multiplier. When EV is negative, the Kelly fraction is negative — which the calculator clamps to zero, since you should simply pass on the bet.
Worked example using a Polymarket contract
Suppose a Polymarket contract on a US election outcome trades at 42¢, implying a 42% market probability. Your own model — perhaps based on the QuantFox leaderboard consensus, polling averages, and recent betting flow — suggests the true probability is closer to 55%. The decimal odds at 42¢ are 1 / 0.42 ≈ 2.38, so b = 1.38. Plugging in: (1.38 × 0.55 − 0.45) / 1.38 = (0.759 − 0.45) / 1.38 ≈ 22.4%. Full Kelly says wager 22.4% of your bankroll. On a $10,000 bankroll, that's a $2,240 position. Most professional bettors would actually deploy quarter Kelly here, putting in $560 — capturing roughly three-quarters of the long-run growth with a fraction of the variance.
Why fractional Kelly is almost always the right answer
Full Kelly is mathematically optimal only if your probability estimates are perfectly accurate. In the real world, your edge estimates are noisy, your win probability is itself an estimate with error bars, and correlated bets can stack risk in ways the simple formula does not capture. Quarter Kelly (multiplier 0.25) has become the de facto standard because it preserves about 75% of the optimal growth rate while cutting drawdowns by more than half. Half Kelly is reasonable if you have very high confidence in your estimates. Full Kelly is only appropriate for situations like card counting where the true probability is mathematically known.
Common mistakes traders make with Kelly sizing
The single biggest mistake is overestimating your edge. If you input 65% win probability when the true probability is 55%, you'll oversize every bet and eventually go broke even though you have a real edge. Treat your win probability conservatively — if you think it's 60%, plug in 55% and let the buffer protect you. The second common mistake is ignoring correlation: betting Kelly-sized stakes on five different "Trump wins" markets is effectively a 5× concentrated bet, not five independent ones. The third is failing to recompute Kelly as your bankroll changes. After a winning streak, your bankroll is larger, so the dollar value of each Kelly bet should grow proportionally; after a losing streak, it should shrink.
Kelly Criterion vs flat betting and unit systems
Many bettors stick with flat betting (always 1% or 2% of bankroll) because it's simpler. Flat betting is a fine starting point and will dramatically outperform random sizing. But it leaves money on the table when you have a large edge and exposes you to too much risk when your edge is small or speculative. Kelly captures the intuition that bigger edges deserve bigger bets and small edges deserve small bets — the larger the gap between your estimated probability and the market price, the larger the recommended position. For traders following high-conviction Polymarket whales or algo trade signals, Kelly sizing makes the math behind that intuition explicit.
When to override the calculator
The Kelly Criterion assumes you can place exactly one bet at a time on a binary outcome with a known payoff. In practice, you may want to cap any single position at 5% of bankroll regardless of what Kelly suggests, especially on illiquid Polymarket contracts where exiting early may incur slippage. You should also reduce position size when implied volatility is high, when news could break before resolution, or when you're trading a market that resolves far in the future. Treat the Kelly output as the theoretical ceiling, not the mandatory bet size.