Approximate and exact corrections of the bias in Croston’s method when forecasting lumpy demand: empirical evaluation

Croston’s method was developed to forecast intermittent demand, employing separate exponential smoothing estimates of the average demand size and the average interval between demand occurrences. Syntetos and Boylan reported an error in Croston’s mathematical derivation of expected demand, and proposed an approximate correction now usually referred to as the Syntetos-Boylan approximation. Subsequently, Shale, Boylan, and Johnston derived the expected bias in Croston’s method and proposed an ‘exact’ correction factor. Both approximate and exact corrections have been derived analytically. In the current study, we empirically investigate whether or not there are actually significant improvements in terms of statistical forecast accuracy as well as inventory control performance obtained by applying the approximate or exact correction.