According to Gizmodo, Columbia University researchers analyzed three years of trading data from predictions market Polymarket and found that approximately 25% of all trading volume consists of artificial “wash trades.” The study identified networks of traders buying and selling the same contracts to create fake volume, with researchers flagging 1.26 million wallets—nearly 15% of all accounts on the platform—as potentially participating in these manipulative activities. At peak times, particularly in December 2024, the researchers estimate fake trading volume spiked to as high as 60% of all activity. The platform’s use of cryptocurrency stablecoins may facilitate these transactions, though researchers don’t accuse Polymarket itself of direct involvement. Polymarket declined to comment on these findings.
The wash trading reality
Here’s the thing about wash trading—it’s basically financial theater. Traders are just moving money between their own accounts or coordinated networks to create the illusion of genuine market interest. The Columbia researchers developed an algorithm that spotted accounts only interacting with small groups of other accounts, regularly buying what their partners were selling and vice versa. And when you’ve got 1.26 million wallets potentially involved, we’re talking about systematic manipulation, not just a few bad apples.
What’s particularly concerning is how this distorts the very purpose of prediction markets. These platforms are supposed to aggregate collective intelligence about real-world events, but when a quarter of the activity is fake, can we really trust the “wisdom of the crowd”? The research paper suggests this isn’t just occasional manipulation—it’s baked into the platform’s DNA.
The crypto complication
Now, the researchers make an interesting point about why this might be so prevalent on Polymarket specifically. Using cryptocurrency stablecoins as the medium of exchange removes many of the friction points and oversight that traditional financial markets have. There’s no settlement period, lower transaction costs, and less regulatory scrutiny. Basically, it’s easier to play games when you’re operating in crypto-land.
But let’s be real—this isn’t just a crypto problem. Wash trading has plagued financial markets for centuries. The difference is that traditional markets have developed sophisticated surveillance systems and regulatory frameworks to detect and punish this behavior. Prediction markets operating in the crypto space? Not so much.
The social media hype factor
While Polymarket might not be directly orchestrating the wash trading, they’re certainly not shy about pumping volume through other means. During the New York mayoral election, their social media team went into overdrive, posting about “surges of whales” betting on Andrew Cuomo and highlighting ballot irregularities. Their November 4 tweet asking “Do they know something we don’t?” and follow-up posts were clearly designed to generate trading activity.
Is this market manipulation? Technically probably not. But it’s definitely shameless hype-building. And when you combine aggressive social media promotion with a platform where a quarter of the volume is fake, you’ve got a recipe for misleading retail traders who think they’re participating in genuine price discovery.
Broader implications
So what does this mean for the future of prediction markets? If nearly a quarter of trading is artificial at baseline—and spikes to 60% during certain periods—can these platforms ever be trusted as accurate gauges of public sentiment? I’m skeptical.
The researchers have basically shown that prediction markets have a massive authenticity problem. And until platforms like Polymarket implement serious detection and prevention measures, we should probably take their “collective intelligence” with a large grain of salt. After all, when the house itself might not be rigged but a quarter of the players are cheating, does it really matter?
