On election night November 3, 2026, prediction markets experienced unprecedented trading activity, with millions of dollars wagered on midterm election outcomes across platforms Kalshi and Polymarket [1]. Trading volume on midterm-related contracts more than tripled from the last week of October into midterms week, with Kalshi exceeding $3 billion in total midterm contract volume and Polymarket's two exchanges reporting nearly $700 million in volume for midterm contracts [1]. Major trades included at least $1 million wagered on a Kalshi market regarding Texas Governor Greg Abbott's reelection after 8 p.m., though Abbott ultimately lost to Democrat Gina Hinojosa by more than 700,000 votes according to the Associated Press [1]. Additionally, more than $1.4 million was traded on Kalshi after 11 p.m. on whether Republicans would secure control of the House of Representatives [1].
Accuracy analysis revealed mixed results for prediction markets. According to a CNBC analysis, more than 87% of all election markets across Kalshi and Polymarket gave the right candidate at least a 90% chance of winning their race 24 hours before election night, with many midterm races being uncompetitive, which boosted the overall hit rate [1]. However, Koleman Strumpf, a professor of economics at Wake Forest University, stated that "prediction markets did pretty well" but "didn't outperform the best," referring to traditional forecast polls [1]. Notably, prediction markets and gambling platforms do not use methodologies employed by traditional political polling and therefore are not substitutes for political polls [1].
Following the midterms, a significant dispute emerged regarding credit for accurate predictions. Both Polymarket CEO Shayne Coplan and Kalshi CEO Tarek Mansour credited prediction markets for correctly predicting races on social media platform X [1]. However, Nate Silver, a political forecaster and founder of the Silver Bulletin newsletter, disputed these claims, arguing that forecast models influenced prediction market movement [1]. Silver wrote on Silver Bulletin that "prediction market probabilities strongly converged toward [our models] in individual Senate and gubernatorial races" and stated he was tracking this closely "precisely because I envisioned that we might wind up in this sort of debate where prediction markets are low-key stealing valor from the models" [1]. Silver joined Polymarket as an advisor in 2024 [1].
CONCLUSION
Prediction markets demonstrated significant trading activity and reasonable accuracy on election night 2026, with billions in volume across major platforms, though their performance matched rather than exceeded traditional polling methods. The dispute between prediction market operators and forecasters like Nate Silver highlights ongoing questions about whether markets are independently predictive or primarily reflect existing forecast models, suggesting caution in attributing prediction market success to superior forecasting methodology.