84% of Polymarket traders are losing money. That says a lot about who prediction markets are really built for.

The most interesting part of the latest Polymarket data isn’t just that most people lose.

It’s how uneven the edge really is.

An April 2026 analysis covering 2.5 million wallets found that 84.1% of participants were unprofitable. Only 2% made more than $1,000 over their full trading history, and just 840 wallets cleared $100,000 in cumulative profit.

That’s not a small retail underperformance story. That’s a market structure story.

The wallets making serious money are concentrated around arbitrage, market making, and execution speed. In other words, the advantage seems to sit with participants who treat these markets like infrastructure, not opinion.

That matters, because prediction markets are often sold as a cleaner way to express views on real-world outcomes. But if most retail traders are consistently late, emotional, or trading without a structural edge, then the market starts looking less like “collective intelligence” and more like another venue where faster players extract value from slower ones.

Real question is whether prediction markets are actually useful for retail traders, or whether they mostly reward people who already have tools and execution advantages the average user doesn’t.