M-competition — A closer look at NAIVE2 and median APE: A note

Abstract Two observations regarding the M-Competition are presented. First, the seasonal indices that were used in the NAIVE2 method were not calculated using the exact procedures that were defined in the M-Competition paper. Second the median absolute percentage error comparative measure was not computed as one might expect it to have been and was not documented as such. The resolution of these matters might enhance the usefulness of the M-Competition study.