"There are lies, there is a terrible lie, then there are statistics and weather forecast" Folk wisdom Initially, we note that the prediction is quite unsteady region where it is difficult to rely on any or accuracy. I once talk with a young man, enthusiastically going on 1C 7.7 operational and managerial accounting. When I asked about the possibility of prediction, it showed a separate module, which they say has long been used by financiers of the enterprise. But when we compared to past periods of these forecasts with actual sales data, it appears that prediction accuracy is around 30% is about the same as if we were asked to make such predictions of random people from the street. The average accuracy of coincidence with reality in such forecasts made almost at random, was to approximately the same range of validity.
1 Another interesting story on this topic once told colleagues at Seminar on sales. Its essence is that one of Moscow's post-graduate students on economic matters has decided to collect all the sales forecasting methods and analyze them for – which is still better to use. Collect it collected, the benefit of using the Internet now make it much easier. But here's the result was sad. The main conclusion arrived at by the hero of this story was that "not worth powder and shot." Ie None of the presently known methods sales forecasting does not have a noticeable predictive value. And really, if you look from a technical point of view, these methods and techniques, you will find that they are based on is not some supernatural formulas and algorithms, but rather simple manipulations to the average values. For example, anti-aliasing in the calculation of trends is none other than the calculation of the average for neighboring points in the diagram.
