Polymarket Hits 78.5% Earnings Accuracy vs 43.7% for Wall Street
A study shows prediction markets outperformed analyst consensus by 35 points, driven by real-money incentives and lack of corporate bias.
Key points
- Polymarket bettors achieved 78.5% accuracy in predicting earnings, significantly outperforming the 43.7% rate of Wall Street analysts.
- The 35-point gap is attributed to real-money incentives in prediction markets versus the biased, negotiated nature of sell-side consensus estimates.
- Companies often lower analyst expectations before quarters close, a practice known as the walk-down, which artificially inflates the likelihood of a beat.
Bettors on the Polymarket platform called quarterly earnings results correctly 78.5% of the time, a significant margin over the 43.7% accuracy rate achieved by Wall Street analyst consensus. This 35-point gap highlights a structural inefficiency in traditional forecasting models, where real-money stakes drive sharper probability estimates than negotiated sell-side projections.
The disparity stems from the incentive structures governing each group. Prediction market participants aggregate diverse signals from supply chains, employees, and customers into continuous price updates. In contrast, sell-side analysts face pressure to maintain optimistic forecasts to preserve access to management, often resulting in estimates that are deliberately lowered by companies to ensure a perceived beat.
Incentives drive accuracy gaps
Companies frequently engage in a practice known as the walk-down, steering analysts toward lower expectations in the weeks before a quarter closes. This engineering of an easy bar to clear makes the consensus figure a biased, negotiated number rather than a neutral baseline. Prediction markets avoid this bias because contracts settle at fixed values based on actual outcomes, forcing participants to price in true probability without corporate influence.
According to Yahoo Finance, the study by Daniel Rabetti and colleagues found that this crowd-sourced approach is particularly effective for firms with company-specific earnings drivers. The real-time nature of prediction market pricing allows for rapid adjustment to new information, whereas consensus estimates often remain stale between analyst refreshes, reducing their predictive power for investors.
Sample limitations and scope
The findings are based on a six-month sample limited to firms with active prediction markets. This constraint skews the universe toward larger, more liquid, and widely followed names, which may not represent the full breadth of the stock market. Despite this narrow scope, the magnitude of the accuracy difference remains substantial enough to warrant attention from anyone relying on traditional earnings headlines.
Forward guidance remains key
Forward guidance retains significant value in earnings evaluation because it is the one metric management has not yet had the opportunity to manipulate downward in the immediate pre-report period. While the study underscores the limitations of consensus estimates, it does not invalidate the importance of management’s outlook for future quarters, which serves as a critical input for long-term valuation models.
