Listing text predicts Polymarket volume but not later dispute risk well
Original title: What Prediction-Market Contracts Reveal at Launch: Trading Volume and Dispute Risk
On 802,863 resolved markets, log-volume R² is 0.747, while dispute ranking reaches AUC 0.78 at a 0.31% base rate.
Researchers use only information available when a Polymarket contract is listed to predict lifetime trading volume and later settlement disputes across 802,863 resolved markets from January 2023 to May 2026. Contract text and metadata predict log volume well out of sample, mostly from market topic rather than neural text models. Accuracy for whether a market reaches $1,000 drops sharply when tested on later periods with new question forms. Dispute scores rank risk better than chance but are too imprecise for individual forecasts.
- Out-of-sample R² for log volume is 0.747; topic adds 0.289 over a 0.220 baseline.
- Wording and resolution-rule design add 0.027 combined.
- Same-period AUC 0.940 for reaching $1,000 falls to 0.577 on later quarters.
- 76.9% of later test markets used question forms absent from training; familiar forms still reach 0.778.
- Dispute average precision is 0.012 at a 0.31% base rate.
From the original article
1 min read at arxiv.orgUsing a near-census of 802,863 resolved markets listed on Polymarket between January 2023 and May 2026, and only information available at listing (the market's title and description, named resolution source, and contract metadata), we predict lifetime trading volume and whether settlement is later disputed. Contract information predicts trading volume well (out-of-sample $R^2 = 0.
By Ahmed Mahrous · arxiv.org
A short excerpt; the article and its images belong to the publisher. Summary, key facts and assessment above are AI-written from the article and can be wrong.