Crowd discussion predicts trading activity and price moves in prediction markets

The Conversation Turns First: Crowd Discussion and Price Reversals in Prediction Markets

Social and Information Networks

Summary

Prediction markets let people trade on guesses about future events and also talk about them publicly. The authors studied if what people say in those discussions helps predict how much trading will happen and if prices will change. They found that paying attention to comments almost matches past trading history in predicting heavy future trading, especially in short-term markets. The tone and stance in comments give clues about the direction of buying, but changes in market leadership are mostly predicted by the current market status rather than comment content. The results suggest that discussions and trading reflect shared information but do not prove that comments cause market changes.

What this means in practice

  • For financial analysts: Enhance prediction models by incorporating discussion-derived signals to better forecast trading volume and price shifts in prediction markets.
  • For social media monitoring teams: Monitor toxic and sentiment features in online discussions to predict crowd behavior changes relevant to market or event outcomes.

Authors

Omneya Sultan, Fred Morstatter

Abstract

Prediction markets combine trading with public discussion of the same events. We examine whether comment-derived signals predict subsequent activity, buying direction, and changes in the leading outcome. A correlation sweep across 79 non-political Polymarket markets guides six classification experiments comparing comment features, trading features, and their combinations. On live blocks containing comments, attention nearly matches trading history in predicting heavy trading within 18 hours (PR-AUC 0.786 versus 0.790, against prevalence 0.606), with its relative advantage concentrated in short markets. Comment content carries directional information: toxicity ranks future buying direction above chance in 43 of 53 scored markets, while adding attention, sentiment, and stance to flow history increases ROC-AUC from 0.780 to 0.788. Leadership changes are predicted primarily by market state. Stance shifts against the leader before reversals in 23 of 27 evaluable markets, but adds no clear improvement in individual-block forecasting. These results distinguish attention from directional support and price uncertainty. They establish predictive associations consistent with discussion and trading responding to shared information, without identifying a causal effect of comments on markets.
View PDFOpen arXiv