Only 3.14% of Polymarket Traders Drive the Market's Accuracy, LBS + Yale Study Finds
An LBS + Yale study of 1.72M Polymarket accounts finds just 3.14% are skilled winners — they capture 30%+ of all gains, while 67% of accounts absorb every aggregate loss in the system across 98,906 events and $13.76B in trading volume.
A new working paper from the London Business School and Yale has put hard numbers on what experienced traders have long suspected about prediction markets: a tiny minority does the price discovery, and everyone else funds it.
The data
The paper, "Prediction Market Accuracy: Crowd Wisdom or Informed Minority?", analyzed 98,906 events, 210,322 markets, and $13.76 billion in trading volume across 1.72 million accounts on Polymarket between 2023 and 2025. Authors Roberto Gomez-Cram, Yunhan Guo and Howard Kung (LBS) plus Theis Ingerslev Jensen (Yale) defined "skilled winners" as traders whose profits persisted out of sample, who participated in an average of 79 markets each, and who consistently positioned with the eventual outcome.
The 3% drives the price
By that bar, exactly 3.14% of accounts qualified. Market makers and skilled takers together represented under 3.5% of accounts but captured more than 30% of total gains. On the other side, 67% of accounts were classified as unskilled losers and absorbed the entirety of aggregate losses. The "wisdom of crowds" thesis — that prediction-market accuracy comes from the collective — does not hold up: accuracy is being produced by a small skilled minority and funded by everyone else.
The Bottom Line
This is the cleanest academic confirmation yet that prediction markets behave more like poker tables than aggregation oracles. For Polymarket and Kalshi, the implication cuts two ways: the product is genuinely useful for institutional users who want forward-looking signal, but the retail user funnel is structurally a losing trade. Watch for renewed regulatory pressure (the same week as the CFTC-NY lawsuit) and for tokenized prediction-market vaults that promise retail exposure to the 3% — that is the obvious product gap this paper just opened.
Daily Web3 × AI intel, straight to your inbox. Subscribe to BlockAI News.
How we report: This article cites primary sources, regulatory filings, and on-chain data where available. BlockAI News uses AI tools to assist with research and first-draft generation; every article is reviewed and edited by a human editor before publication. Read our full How We Report page, Editorial Policy, AI Use Policy, and Corrections Policy.
