A large-scale analysis of closed markets on Polymarket from 2021 to May 2026 reveals a troubling picture: about 70% of all completed contracts failed to accumulate a trading volume of even $10,000. This is not just a statistic — it is a diagnosis of an entire segment of the prediction industry, where real liquidity and interest are concentrated only in a few dozen major events.
Less than 10% of closed markets showed a volume ranging from $100,000 to $1 million. Moreover, nearly 45,000 contracts — almost 5% of the total — did not record a single trade. These are the platform's "dead souls," which only create an illusion of activity.
Bots as the main beneficiaries of the low-liquidity segment
The key finding of the study is that in the segment with a turnover below $10,000, bots exclusively dominate. Analysts found that over 80% of all volume in such markets is generated by wallets making more than 50 transactions per day or over 1,000 in total. These algorithms earned about $1.2 million on "micro-markets," but their main profit is concentrated in larger contracts.
In the range from $1 million to $10 million, bots have already earned about $50.5 million, accounting for 38% of their total profit. Markets with a volume exceeding $10 million added another $35.1 million. Thus, bots profit from every trade and actively migrate to the most liquid pools, leaving the periphery lifeless.
The World Cup as a catalyst, but not a panacea
The sharp increase in volumes on prediction platforms in June 2026, linked to the FIFA World Cup, only confirmed this trend. From June 1, the weekly volume on major platforms surged from $65 million to $5.4 billion, peaking at $5.6 billion in the week of June 22–28. The main contribution came from Kalshi, but even this spike did not change the fundamental market structure: tens of thousands of small contracts remain illiquid, and all money flows into a few top events.
A similar situation is observed on Kalshi, where many low-liquidity markets have also been recorded. The only difference lies in the calculation methodology: on Dune for Kalshi, only one side of the trade is counted, which somewhat distorts the picture but does not negate the overall conclusion.
Expert opinion: Polymarket and Kalshi demonstrate the classic "long tail" effect, where 90% of content provides no value. Until the platforms solve the problem of spam markets and implement mechanisms for organic liquidity attraction, real volume and profit will remain the domain of only a few high-profile events. Bots here are not the evil, but a consequence of the lack of quality user demand for niche predictions.