Almost 70% of the retail-style accounts studied on Polymarket’s international platform finished below break-even, with the group collectively losing $338.9 million.
Galaxy Research’s October 1 report analysed approximately 2.9 million accounts, using public blockchain settlement data curated by Stork. The findings highlight the gap between prediction markets’ usefulness as forecasting tools and participants’ ability to profit from them.
What the study found
The report’s executive summary identifies six findings:
- Most accounts lost money. Overall, 69.2% finished below break-even.
- The group’s aggregate result was negative. Across the studied population, gains and losses combined produced a $338.9 million loss.
- Losses were associated with lower participation. After a loss, 15.2% of accounts did not trade again within 30 days, compared with 6.1% after a win.
- Specialisation produced uneven results. About 44.1% concentrated more than 60% of their activity in one topic. Sports represented 47% of specialists, but only 25.1% of sports specialists were profitable. Technology and science had the highest profitability rate, 41.2%, although the sample was smaller.
- Profitable traders generally took larger positions. This was an observed relationship, not evidence that increasing a stake improves the chance of success.
- The difference in typical position size was modest. Profitable traders had a median position of $13.96, versus $10 for unprofitable traders.
Holding positions longer showed no consistent relationship with profitability once traders were compared by activity level.
Related: Internal videos show how Polymarket trained creators on fake bets
What “retail” means here
Galaxy used 50 orders per active day as its cutoff to filter out accounts likely to be automated. The classification measures trading pace, rather than confirming whether an account belongs to a small individual investor or a professional.
Wallets also do not equal unique people. Someone who stopped using one address could have continued trading through another, limiting what inactivity tells us about people leaving the platform.
The findings apply to Polymarket’s international platform, not its separate US exchange.
Large aggregate losses, small typical losses
The median account lost approximately $3, showing why the total loss should not be mistaken for the typical participant’s experience. Larger losses were concentrated among a smaller portion of accounts.
The study shows that most accounts finished behind, but it does not establish a winning strategy. Topic familiarity, larger stakes and longer holding periods should not be treated as guarantees of profit.
Disclosure: This article does not represent investment advice. The content is for informational and educational purposes only.
