Both the empirical and normal distributions were calculated to determine which would be the best model to use to maximize order quantity. The empirical distribution is helpful because it shows/reflects the historical forecasting capability. It does, however, have many shortcomings. First of all, it only predicts a limited amount of possible outcomes-this is because it is a discrete distribution. Also, the historical data is limited to only the wines from the 2004 catalog, which is only 45 items, thus adding to the limitations of possible outcomes. Each A/F ratio was calculated in relation to the critical ratio for each wine and the percentile ranking. These A/F ratios were then applied to the forecast quantity and the profit maximizing order quantity was determined. In my opinion, in order for the empirical model to be the most useful, there would have to be more items to include and analyze.