When Polymarket users wagered that a company's earnings would fall short of expectations, their accuracy rate reached 44%, more than double the historical baseline of 18%. When traders expressed high confidence that earnings would exceed expectations, the accuracy rate climbed to 90%, surpassing the industry average of 80%. Yin Luo, head of quantitative research at Wolfe Research, attributed the high accuracy to the crowdsourced model: "Investors placing bets on Polymarket may be more diversified than market consensus expectations based on sell-side analysts' (financial firms' research teams) predictions."
A working paper updated in early April by researchers from London Business School and Yale University found that these emerging prediction platforms integrate new information faster than analysts while avoiding certain inherent biases in Wall Street earnings forecasts. Researchers noted that participants bet with real money, and many users participating in earnings predictions are exceptionally professional; insider trading may also be a factor influencing prediction accuracy. Despite the platforms' potential, event contracts (derivatives linked to earnings outcomes) currently represent a small fraction of trading volume on Polymarket and competitor Kalshi. According to data tracked on Dune Analytics, Polymarket's earnings prediction trading volume was only $795,315 in the most recent week, accounting for just 0.03% of total platform volume. While financial institutions are investing heavily in prediction platforms, the sector remains in early stages.