Winning at weather isn't about predicting the weather better than meteorologists — it's about reading the market's overreaction to forecast noise. The strategy in three rules.
First: weight models by station bias. NOAA HRRR is great for short-range US temperature but consistently warm-biased in coastal cities. Apply the bias correction and you instantly outperform the consensus.
Second: trade the disagreement, not the consensus. When GFS and ECMWF diverge by more than 2°C on tomorrow's high, the market misprices the uncertainty. Take the side of the historically more accurate model for that station.
Third: hard-cap correlated exposure. A heat dome over Europe is one trade, not six. Position sizing per region prevents catastrophic drawdowns that wipe out months of edge.
