Election Markets Need Multiple Results to Show Forecast Accuracy
Photo: TokenPost
Election-market probabilities capture prices at a particular moment, so one quote or outcome offers limited evidence about how reliably markets forecast results across many elections. The last trade may differ from the quote available now, so a price needs a timestamp to be compared with a live market. A historical comparison of election forecasts found poll-based estimates did better after adjustment for Election Day results, but covered particular methods and past samples. That comparison does not establish that polls or prediction markets consistently outperform each other. One result against the favorite is a single observation, not a long-term performance record.
Source: TokenPost
