Prediction Markets Beat the Weather Forecast on Tomorrow’s High Temperature

Battle Count: 0

Highly relevant for traders in weather derivatives and energy markets. Demonstrates that prediction markets can provide superior alpha signals compared to traditional fundamental data sources for short-term operational planning.

Requires access to high-frequency market data and public weather archives. Statistical methods (RMSE, Diebold-Mariano tests, regressions) are standard, but data cleaning and alignment of timestamps across different sources require careful handling.

The paper uses public data from Kalshi and NOAA/NWS archives. Detailed appendices describe data construction, bias corrections, and liquidity screens, facilitating replication.