Polymarket data shows top traders capturing most platform profits

A Wall Street Journal report from earlier in 2026 put a number on something most participants in prediction markets already suspected: sixty-seven percent of all profits on Polymarket were flowing to one-tenth of one percent of accounts. That is not a healthy market. That is a toll road with one exit ramp.

The comparison that keeps coming back to me is computer-assisted wagering on US horse racing tracks. Syndicates with superior data processing and order execution advantages drained liquidity from those venues systematically, until the tracks faced a genuine question about whether casual participants would keep showing up. Some stopped showing up. The markets that survived did so because the operators made structural changes, not because they trusted the dynamic to self-correct.

Prediction markets are running the same dynamic right now, and the industry is largely discussing it as a regulatory classification problem. Whether a Kalshi contract is a futures instrument or a gambling product is a real legal question, but it is not the operational question that matters most here. The operational question is whether a market where a fractional sliver of accounts captures the majority of returns can sustain the retail participation that justifies its existence. The answer from horse racing is: not indefinitely.

The counterargument is that this is what efficient markets look like. Informed traders push prices toward truth; uninformed traders pay for the signal. That is accurate as far as it goes, but it assumes the uninformed traders understand the arrangement they are entering. On Polymarket, on Kalshi, on any of the platforms currently marketing themselves to the Gen Z cohort as a smarter alternative to stocks, the pitch is participation in price discovery. The pitch is not "you are providing liquidity to better-informed accounts at a consistent loss." Both things can be true simultaneously. Only one of them appears in the marketing.

Concentration of this magnitude in a relatively young market also produces a specific structural risk that the debate around it has not caught. When the profitable accounts are few enough to count, the market's accuracy becomes dependent on their continued participation. If regulatory pressure, platform changes, or the simple economics of scale push those accounts elsewhere, the price signal degrades fast. The market looks efficient until the day it does not, and there is no early warning system for that transition.

The platforms know this. The smarter operators are already thinking about maker-taker structures, information asymmetry disclosure, and retail experience design. Whether they move fast enough to address the concentration problem before regulators use it as a cudgel is the question I would be pricing right now, and the direction is not favorable to the status quo.