EN
Main aim of the article was comparison of sales forecasts effectiveness in a company. In the first stage time series representing sales were selected with respect to presence of seasonality and consistency with gamma distribution. Two types of statistical tests were used: F test (seasonality) and nonparametric Locke tests (compatibility with gamma distribution). In the next step ex post forecast were computed by means of linear trend model with seasonality and gamma distribution (on the level of median of theoretical distribution). Effectiveness of forecasts was compared on the basis of chosen coefficients of forecasts errors. Forecasts obtained be means of gamma distribution turned out to be better in all considered aspects