An examination of the accuracy of judgmental extrapolation of time series
Recent studies of forecasting accuracy have neglected to include the most commonly used technique of judgmental extrapolative forecasting, perhaps reflecting the widely held view that this approach is inferior to statistical techniques. This paper reports on a study of the accuracy of judgmental extrapolation of time series. Three alternative judgmental forecasting approaches were used and their accuracy compared to that of the quantitatively-based forecasts developed by experts in the course of a competition. The paper concludes that judgmental extrapolation is on average no less accurate than statistical forecasting and, in a number of subgroups of the time series, was the most accurate. Further, one of the judgmental approaches appeared to provide more robust forecasts than the statistical techniques.