Are prediction-market odds actually accurate?
Often more accurate than pundits or polls — here's the evidence, and the calibration to expect.
They have a real track record
These aren't just vibes. The Iowa Electronic Markets, run by the University of Iowa since 1988, have forecast election vote shares with an average error of around 1.5 percentage points — frequently closer than the final pre-election polls. Reviewing the evidence across elections, sports and business, economists Justin Wolfers and Eric Zitzewitz found that market forecasts typically beat moderately sophisticated benchmarks.
Why money makes a forecast honest
When people put real stakes behind a view, they're paid to be right and they lose when they're wrong — which filters out cheap talk and wishful thinking. The result is that the price tends toward a stake-weighted average of everyone's beliefs, so it reflects information rather than the loudest opinion.
What “well-calibrated” means
A forecast is well-calibrated when things it calls 70% likely actually happen about 70% of the time. Across many markets, prediction-market prices clear that bar reasonably well. PredictionHub leans on this: we publish a cross-venue consensus and let you track your own calibration in My Predictions.
But not magic
Accuracy depends on enough people trading enough money. Thin, quiet markets are noisier, prices can lag breaking news, and no forecast is a guarantee — see “When to distrust the odds.”
Sources
Keep reading