CropProphet Modeler: Historical Corn, Soybean, Winter Wheat Yield & Production Forecasts for USA

CropProphet features 5 separate but related data sets to enable strategy calibration, quantitative model performance evaluation and portfolio alpha generation. Each data set will provide historical forecast information for: • corn, soybeans, and winter wheat, • national yield and production, and • daily time resolutions. The data sets are: Real-time The real-time, daily updated yield/production forecasts produced from May 1 to October 2nd of each year. This data is used during the current crop season. Point-in-Time The actual forecasts released by CropProphet from 2014 to 2019. It is important to recognize the, as stated in the summary, that the CropPropeht model input data and/or modeling methodology is updated each year. Overall accuracy estimates of CropProphet based on our Point-in-Time data should be treated carefully. Rolling A 10 year “out-of-sample” information-based forecast of yield/production. This simulates the Point-in-Time data but for the current operational forecast model. This data dataset facilitates a quantitative understanding of the performance of the current model. Cross Validated A prior 34-year data set based on “leave one out” model building, training, and prediction. This is the fairest means to evaluate the accuracy of the CropProphet forecasting performance. In-Sample A 15-year yield/production data set created using the same input data used to train the operational forecast model. This model helps to confirm the proper behavior of the cross-validated and out-of-sample Validates the behavior of the model based on prior years.