Analysis Reveals Polymarket's Top Traders Are Poor Forecasters, Not Skilled Predictors

Summary

A detailed statistical analysis of Polymarket's top 100 profit-generating traders has uncovered a counterintuitive finding: the platform's biggest earners are actually the worst-calibrated forecasters. Using rigorous statistical methods including Brier scores and rank correlations, researchers found a strong positive correlation (+0.608) between poor calibration and high profits—meaning worse forecasters consistently earn more money. The analysis examined 3,651 resolved positions across 95 wallets totaling $72.2M in realized profit, finding that 21 traders performed worse than random chance yet still ranked among the platform's top earners. The research suggests that Polymarket's leaderboard is driven not by forecasting skill but by concentrated directional bets, particularly on the 2024 election market, which accounted for 39% of aggregate profits among the top 100 traders.

• Evidence of potential coordinated insider trading persists beyond the identified 'Theo' wallets, with 8 additional wallets showing same-day sweep patterns around election day
• Polymarket's profit leaderboard may reflect access to information or market manipulation rather than superior forecasting ability

Editorial Opinion

This analysis exposes a critical flaw in using realized profit as a measure of forecasting skill on prediction markets. The finding that worse-calibrated traders earn substantially more suggests Polymarket may be vulnerable to informed trading and market manipulation rather than serving as a pure mechanism for aggregating accurate forecasts. The concentration in the 2024 election market particularly raises questions about whether the platform is functioning as intended—rewarding accuracy—or whether it has become a casino where directional bets and potential information advantages dominate. For Polymarket's credibility as a forecasting tool, these structural issues demand urgent attention.