Blog
📈
Strategy6 min readApril 4, 2026

Portfolio Theory for Prediction Markets: Sizing Positions Like a Fund Manager

Kelly Criterion, correlation, and diversification applied to prediction markets. Stop sizing positions by gut feel and start sizing them by expected value.

Most prediction market traders size their positions by gut feel or arbitrary round numbers. Fund managers size positions using portfolio theory. The difference in returns over time is substantial.

The Kelly Criterion in Prediction Markets

Kelly Criterion gives you the theoretically optimal fraction of your bankroll to bet on a given position, given your estimated edge. The formula is: f = (bp - q) / b, where b is the net odds received, p is your estimated probability of winning, and q is 1 - p.

In practice, most experienced traders use "fractional Kelly" — betting 25-50% of the Kelly-recommended fraction. Full Kelly maximises long-run growth but produces terrifying drawdowns that most humans cannot psychologically tolerate.

📊

Half-Kelly is theoretically sound and psychologically survivable. Full Kelly is mathematically optimal but produces drawdowns that cause most traders to abandon the strategy at the worst possible time.

Correlation in Your Prediction Market Portfolio

If you hold positions in five different markets that all depend on the same underlying event (say, a single election outcome), you do not have five independent bets. You have one concentrated bet expressed across five markets. Correlated positions should be sized as if they are a single position.

  • Map out which of your positions share underlying exposures
  • Group correlated positions and apply Kelly sizing to the group, not individual positions
  • Genuinely uncorrelated markets (sports vs politics vs tech) are your real diversification
  • Markets that resolve at the same time create liquidity risk — you cannot easily exit multiple positions at once

Maximum Exposure Rules

Set a maximum exposure per market (e.g. no more than 10% of bankroll in any single market) and a maximum exposure per domain (e.g. no more than 30% in political markets). These rules feel constraining until the one time they save you from a catastrophic loss.

"The first rule of prediction market portfolio management: do not blow up. The second rule: do not blow up. After that, optimise for returns."

Adapted from market wisdom

#portfolio-theory#Kelly-Criterion#position-sizing#borro-market#prediction-market-bankroll#expected-value#risk-management-prediction-markets

More in Strategy