This article uses an exponential smoothing method to forecast the time series of temporary sales of an opencast rock and raw materials mine. Six models of foresight were developed: three seasonal additive models (with linear, exponential and fading trend) and three seasonal multiplicative models (with linear, exponential and fading trend). The exponential smoothing method can be used for current enterprise management, not just mining. This method can be used to make relevant decisions based on the verified forecasts. It is designed for short-term forecasting-even several times during the day based on current changing data. This method is a useful tool for forecasting time series for not only sales. It can be used to forecast inventory, receivables, etc. However, despite the great progress in predictive methods of the future, which is particularly aided by computer technics, the forecast of the Economist is fraught with greater or lesser errors, and it is therefore necessary to verify developed models. The quality of the forecast should be determined by its relevance, which is determined by means of ex post errors (expired forecast errors). Furthermore, the quality should also be monitored and the forecasts should be corrected if necessary. To develop forecasts in the six models mentioned, the STATISTICA program, which provides a transparent and quite rapid foreca - sting of the use of the exponential smoothing method in twelve possible variants. STATISTICA also allows to verify the developed model by drawing an adjustment chart of this model with actual time series, verifying ex post errors, and creating a histogram of the rest of the model. This article also carried out verifications of the models developed by designating the errors of expired forecasts (ex post errors), as well as verification, on the basis of the histogram, whether the rest of the developed models have a normal distribution. For this purpose the Shapiro-Wilk test was used.
B. Trzaskuś-Żak (Sun,) studied this question.