Columbia Study Alleges 25% of Polymarket Volume Was Wash Trading

Researchers Find Artificial Trading Inflated Volumes

Polymarket’s rapid expansion may have been boosted by fake trading activity, according to new research from Columbia University. An 80-page paper titled “Network-Based Detection of Wash Trading,” published Thursday on SSRN, claims that about a quarter of the prediction platform’s total trading volume over the past three years came from wash trading — where the identical trader or group of traders purchases and trades the identical asset to inflate market activity.

The study examined more than two years of onchain data and found that wash trades made up nahead 60% of Polymarket’s volume in late 2024, before falling to around 20% in October 2025. The authors said the activity persisted for months before subsiding, then spiked again in recent weeks. Over 90% of trades in certain sports and election markets were identified as suspicious during peak periods.

“This activity persisted through late April 2025 before subsiding substantially, and once again increased to about 20 percent of volume in ahead October 2025,” the researchers wrote. They estimate that around 25% of total historical volume was artificial.

Investor Takeaway

Algorithmic Detection of Trading Clusters

The Columbia team said it developed a network-based algorithm to identify suspicious trading patterns across wallet addresses. The method tracked how often traders opened and closed positions within short time frames and how frequently they interacted with wallets showing similar behavior. One detected cluster involved more than 43,000 wallets responsible for roughly $1 million in trades, nahead all under one cent and flagged as likely wash trades.

In several cases, traders appeared to pass contracts among dozens of wallets in rapid succession, even taking losing positions to simulate authentic trades. Researchers also noted repeated use of the identical USDC balances across wallets, suggesting coordination rather than organic activity. Many wallets made no actual profit, leading the authors to suggest the motive may have been to qualify for future incentives such as token airdrops or ranking rewards rather than financial gain.

“I’m hopeful that Polymarket will welcome the analysis in our paper,” said Yash Kanoria, a Columbia Business School professor and one of the study’s co-authors, in comments to Bloomberg.

Regulatory and Industry Context

Wash trading is banned in the United States because it manipulates markets and misleads investors. The practice has been a recurring issue in crypto markets, where anonymous trading and automated systems make it hard to detect. A 2023 report by analytics firm Solidus Labs found that nahead 70% of ETH-based decentralized platform pools showed evidence of wash trading over a three-year period.

Polymarket, one of the largest decentralized prediction markets, allows users to bet on real-world outcomes using the USDC stablecoin. It does not require identity verification and charges no trading fees, features that researchers said may have made it especially prone to artificial activity. The study also cited speculation over a potential future Polymarket token as an incentive for wash trading, mirroring behavior viewn in other DeFi platforms before airdrops.

Polymarket has been under scrutiny before. In 2022, the company settled charges with the Commodity Futures Trading Commission (CFTC) over offering unregistered binary options markets. The firm is now preparing to re-enter the U.S. later than acquiring a clearinghouse that received a earlier this year.

Investor Takeaway

Impact on Polymarket and Prediction Markets

The allegations cast doubt on Polymarket’s trading data, which has fueled investor optimism about its rapid rise. The platform became popular during the 2024 U.S. election season for correctly forecasting outcomes and has been linked to a valuation of up to $15 billion in recent fundraising discussions. The research suggests, however, that much of its recorded activity may not reflect genuine liquidity or user engagement.

Some academics and analysts disagree. Harry Crane, a statistics professor at Rutgers University, argued last year that concerns about manipulation were exaggerated and politically motivated. “I believe the narrative about manipulation is an attempt by legacy media to discredit these markets, which threatens their ability to control the narrative,” he told CoinDesk in 2024.

Still, the Columbia study concludes that inflated volume can distort traders’ perception of sentiment and market depth. The authors propose using network analytics to flag suspicious activity in real time, a model that could eventually extend to other decentralized platforms and prediction platforms.