Book
Distilling the Wisdom of Crowds: Prediction Markets vs. Prediction Polls
by Pavel Atanasov, Phillip Rescober, Eric Stone, Samuel Swift, Emile Servan-Schreiber, Philip Tetlock, Lyle Ungar, Barbara Mellers · 2017 · 1 reading card
1 card
Distilling the Wisdom of Crowds: Prediction Markets vs. Prediction Polls · 2017
A market aggregates through price, a poll through an algorithm; the good algorithm has decay, track-record weights and recalibration.
The classic case for markets comes from the Science article (2008) signed by Kenneth Arrow and a large group of co-authors, economists and legal scholars: prediction markets can produce forecasts with lower error than conventional methods. Their example: in the final week before US presidential elections from 1988 to 2000, the Iowa Electronic Markets erred by 1.5 percentage points on average, against 2.1 for the final Gallup poll. The price gathers the information of those willing to bet on it and updates continuously. Atanasov and colleagues then pitted the two methods against each other with random assignment: more than 2,400 participants, 261 events, two seasons of the geopolitical tournament. The market beat the simple mean of the polls. But team polls aggregated with three ingredients — older forecasts count less, those with a good record count more, the aggregate is recalibrated — beat the market, with the largest advantage at the start of long questions, when the market is still thin. For an organisation the conclusion is practical: you do not need a market to aggregate well; you need numerical forecasts, a track record and an algorithm. The 2026 landscape is legally murky: in the US the status of event contracts is disputed between the federal regulator and the states, and on 25 September 2026 a federal appeals court (the Sixth Circuit) ruled that states may apply gambling laws to sports contracts (CNBC). The course neither evaluates nor recommends platforms.
“In both seasons of the tournament, prices from the prediction market were more accurate than the simple mean of forecasts from prediction polls. However, team prediction polls outperformed prediction markets when forecasts were statistically aggregated using temporal decay, differential weighting based on past performance, and recalibration.”