What this page tracks, and why only this. The research behind it tested free weather forecasts against prices for gas, power, crops and energy equities, from daily rules to machine-learning models, and found no edge: markets already price what the public forecasts say, and day-to-day weather moves prices far less than their own noise. What persisted in causal tests was confined to the rare, extreme tail of slow climate states (El Niño above all) and to the CPE framework’s own admission rules, on few episodes and without a positive return in a simple rule; it is a candidate, not a result. This page follows only that, as frozen CPE rows scored forward. Every signal in the table was frozen on from data through 2012 and is scored forward only. The daily weather-to-demand panels, which are still scored for the registered experiment, are on the archive page.