This article develops a procedure for weighting historical loss cost experience based on longer time-series weather information. Using a fractional logit model and out-of-sample competitions, weather variables are selected to construct an index that allows proper assessment of the relative probability of weather events that drive production losses and to construct proper “weather weights” that are used in averaging historical loss cost data. A variable-width binning approach with equal probabilities is determined as the best approach for classifying each year in the shorter historical loss cost data used for rating. When the weather-weighting approach described above is applied, we find that the weather-weighted average loss costs at the national level are different from the average loss costs without weather weighting for all crops examined.