Short and sweet: Forecast accuracy

Forecast accuracy is defined as the deviation of a forecast from the actual values that are later determined. Forecast accuracy is expressed via a forecast error , which measures the extent to which a forecast deviates from the subsequent actual values.

Forecast accuracy, forecast error and forecast period

A forecast accuracy and a forecast error always refer to a specific forecast period, e.g. a three-month forecast. In practice, forecast errors are often measured for different forecast periods. By tracking the forecast accuracy, it is hoped that, on the one hand, the forecast can be continuously improved and, on the other hand, any systematic forecast bias in the form of forecast values that are continuously too high or too low can be identified. If there is a systematic forecasting error, this can be used to correct future forecasts.

There are numerous different formulas for measuring the forecast error, the advantages and disadvantages of which are hotly debated in practice.

Our tip for forecast accuracy

In many cases, forecast accuracy as an assessment criterion for the quality of a forecast falls short of the mark. No matter how much effort is put into preparing forecasts, there is always a residual amount of chaotic market behavior and therefore a forecasting error that cannot be reduced any further.

In the majority of articles, there is generally no systematic forecasting error. Such systematic forecasterrors are more common in human forecasts . If powerful machine forecasts are used (statistical statements; AI methods), the algorithms themselves reduce the error as much as possible. Forecast accuracy is therefore only for information purposes, but is not a regulatory criterion for making forecasts continuously more accurate.

In practice, we even observe the danger that the focus on forecasting accuracy leads to ever greater mathematical sophistication without any practical benefit in most cases; chaotic behavior cannot be squeezed into a single prediction value. On the other hand, since the breadth of chaotic market behavior can be measured, safety stocks can and must absorb the remaining forecast inaccuracy.

 For this reason, it is important tohave the system calculate safety stocks and not to set them manually.

Picture of Prof. Dr. Andreas Kemmner

Prof. Dr. Andreas Kemmner

Talk to us!

We are there for you personally and will be happy to advise you individually on our services and solutions.