Because of the lack of any empirical reference with which to compare predictions, validation of long run linear programming models is extremely difficult. This paper reports two procedures for combining time series data with results from a long run linear programming supply model in order to make verifiable short run predictions.