Who Wins and Who Loses In Prediction Markets? Evidence from Polymarket

84 Pages

Posted: 7 Apr 2026 Last revised: 1 May 2026 Date Written: March 18, 2026AbstractWe study trading gains and losses on Polymarket, the world’s largest prediction market platform. Our comprehensive dataset spans 2022 to 2026, covering more than 2.4 million users, $67 billion in volume, and 588 million trades. We document a striking profit concentration: the top 1% of users capture 76.5% of all trading gains. Despite this concentration, prices are well-calibrated and aggregate information efficiently. Gains flow almost entirely to sophisticated traders who trade using limit orders at prices that are advantageous relative to realized outcomes. Losses, by contrast, are associated primarily with liquidity-taking; longshot betting (trading at extreme prices) is a descriptive marker of losers but plays a smaller role once we control for activity scale. For roughly one in five losers, the lower-bound cost of taking liquidity alone is enough to flip their PnL from negative to positive. Monthly performance is only weakly persistent, however, so this cross-sectional pattern may reflect heterogeneous trading conditions or selection rather than persistent skill. Our results suggest that the informational benefits of prediction markets come at a cost to unsophisticated participants.

Keywords: prediction market, market microstructure, profit concentrationJEL Classification: G12, G14, G15, D47, D82 Suggested Citation: Suggested Citation