Note: Further Notes on Lagging Forecasts

In Gardner [Gardner, E. S. 1984. The strange case of the lagging forecasts. Interfaces14(3) 47–50.], I pointed out that erroneous versions of exponential smoothing models have been repeated in the literature for the last 20 years. Some 23 errors were cited in journal articles and textbooks. All of the errors were in linear-trend models and caused the forecasts to lag the data.

Since Gardner [Gardner, E. S. 1984. The strange case of the lagging forecasts. Interfaces14(3) 47–50.] appeared, I have received a number of letters requesting information on other model formulations for exponential smoothing. Gardner [Gardner, E. S. 1985. Exponential smoothing: The state of the art. J. Forecasting4(1) 1–38.] is my attempt to provide a complete reference on the subject. This paper includes 29 models along with details on starting values, parameter choice, and forecast monitoring.