We report the results of the first large-scale, long-term, experimental test between two crowd sourcing methods – prediction markets and prediction polls. More than 2,400 participants made forecasts on 261 events over two seasons of a geopolitical prediction tournament. Some forecasters traded...

Reliable detection of forecasting skill is slow and resource-intensive. Forecasters need to answer dozens of questions, which may take months or years to resolve. To accelerate this process, we propose the Full Accuracy Score (FAS), a measure that combines the strengths of ground-truth-based...