#Gate广场五月交易分享 The Truth About Prediction Markets: It's Not Crowd Wisdom, but 3% Informed Traders Driving Prices

Prediction markets are renowned for their high accuracy over the long term, and the common attribution is to "crowd wisdom"—the aggregation of information from many participants, with biases offsetting each other to ultimately form an effective price. However, the latest top-tier SSRN study based on full transaction data from Polymarket overturns this traditional view: the accuracy of prediction markets is not derived from the masses, but is driven by approximately 3% of specialized traders.

1. Core Conclusion: The True Operating Mechanism of Prediction Markets
As the world's largest prediction market platform, Polymarket's trading volume surged from $3.3 million in 2023 to nearly $2 billion in 2025, covering nearly 100k event types across politics, economics, sports, and crypto assets. The study uses comprehensive blockchain account and transaction data to stratify traders and examine the sources of price discovery, arriving at three key conclusions:
1. Only 3.14% of accounts have sustained profitability
Using symbol randomization tests to distinguish skill from luck, all traders are categorized into five groups: Market Makers (0.1%), Skilled Winners (3.14%), Lucky Winners (29%), Unlucky Losers (61.4%), and Unskilled Losers (6.41%). Only skilled winners demonstrate statistically significant sustained profitability; profits of lucky winners cannot be extended out-of-sample.
2. Skilled winners dominate all price discovery
Order flow from skilled traders significantly predicts future prices and final outcomes, continuously pushing prices toward the true result and improving pricing accuracy; while the majority of traders contribute most of the trading volume, they do not provide informational value, with some groups persistently trading in the opposite direction.
3. Profits are highly concentrated, losses fund gains
Skilled winners and market makers together account for less than 3.5% of accounts but earn over 30% of market profits; 67% of ordinary traders bear all losses, with the public providing liquidity and profits to a small informed minority.

2. Crowd Wisdom vs. Informed Minority: Theory and Empirical Evidence
Traditional crowd wisdom theory assumes all traders possess independent information, with no excess returns or sustained advantage; in contrast, informed trading theory posits that only a small minority hold effective information and profit continuously from the majority. The data fully supports the latter:
Profit Persistence: Skilled traders within the training set, 44% remain skilled in the test set, far exceeding the 10% level of active public funds; 51% of unskilled losers remain unprofitable, and 60% of luck-based winners turn into losers outside the sample.
Price Prediction Power: Only orders from skilled winners can significantly predict next-period price movements and final outcomes; lucky winners have no predictive ability, and unskilled losers persistently trade in reverse.
Pricing Improvement: The higher the trading volume from skilled traders, the smaller the pricing error, with the strongest effect near the announcement of results; public traders, on the other hand, tend to inflate pricing deviations.

3. Key Empirical Question: Who Reacts to Information First?
The study focuses on scenarios where information is clearly realized: FOMC decisions and corporate earnings announcements, testing which trader groups respond first to the news.
Results show: Only skilled traders react immediately at the moment of news release, trading in the direction of the surprise; other groups show no consistent response, with some even trading in the opposite direction. This demonstrates that skilled winners possess stronger information processing and rapid execution capabilities, making them the sole carriers of information into prices.

4. The Real Impact of Insider Trading
Prediction markets are often suspected of being dominated by insider trading due to weak regulation, anonymity, and contract precision. The study identifies suspicious accounts through "short-term account opening, concentrated large positions, and stopping trading after events," finding:
Insider trading can have a significant impact in single instances, with predictive power markedly higher than that of ordinary skilled traders;
However, such cases are limited to a very small number of events and do not constitute the core force behind overall market price discovery;
Overall pricing efficiency remains supported by sustained skilled traders, not by sporadic insider trading.

5. Investment Implications for Individuals and Institutions
1. Abandon the simplistic notion that "public consensus = correctness"
The effectiveness of prediction market prices does not mean the public's judgment is accurate; rather, a small number of professionals inject information into prices, so blindly following market sentiment is unwise.
2. Trading volume ≠ informational content
High trading activity and popularity of an asset do not necessarily reflect genuine information; often, they are just noise capital gathering.
3. Sustained profits are a matter of skill, not luck
Realizable, replicable gains come from informational advantages, processing efficiency, and disciplined execution; large profits from a single event are likely due to luck and should not be generalized.
4. For institutions: tracking informed order flow is more valuable
Whether for macro expectations, event-driven strategies, or policy interpretation, following the behavior of a small number of skilled traders is more effective than aggregating public opinions.

6. Study Summary
The high accuracy of prediction markets is not a product of crowd wisdom but results from 3% informed traders continuously setting prices, with 97% of the public providing liquidity. Price efficiency does not imply rational participation; the core source of market efficiency is the informational advantage and professionalism of a small elite.