New research suggests that the vast majority of users on Polymarket are not profitable. According to on-chain analyst Andrey Sergeenkov, 84.1% of traders are currently in the red.
The study analyzed around 2.5 million wallet addresses using blockchain data from Polygon via Dune Analytics. It found that only 2% of users have ever made more than $1,000 on the platform, while just 0.033%, roughly 840 wallets, have earned over $100,000.
The study refutes the assumption that consumers can receive a consistent source of income through prediction markets. There is a chance less than 1% that any user will consistently earn $5,000 a month. Among the highest-level earners (6,600 average monthly profit of $5,000 or more), only 2.6% remained active for more than 1 year.
According to this report many of the users make trades shortly before leaving the site, meaning that their average profitability drops over time. According to a different 2025 analysis of 124 million trades, about 70% are unprofitable.
This research was released at the same time that Polymarket continues to grow, as evidenced by its recent partnership with Major League Baseball and its position among the largest prediction market sites worldwide. Data from Token Terminal indicates that $9.8 billion worth of trades were made on Polymarket over a 30 day period (the second most-traded prediction market behind Kalshi).
Polymarket has also launched several new features, some with referral programs to help attract users, and Sergeenkov warned about potential losses to ordinary retail traders due in part to lack of user education.
The rest of the prediction market industry has experienced explosive growth, with significant increases in volumes traded from 2024 to 2025. This growth has drawn even more regulator attention than before, especially within the United States, as various regulatory bodies examine how these markets should be categorized and regulated.
Polymarket has also introduced a new proprietary stablecoin, Polymarket USD, as part of a wider infrastructure upgrade. While its crowd-based forecasting model is often praised for accuracy, the data highlights a stark contrast between market predictions and individual trader outcomes.
