03-11-2016 05:12 PM - edited 03-11-2016 05:15 PM
This discussion: https://communities.sas.com/t5/Forecasting-and-Econometrics/out-of-sample-range-and-holdout-sample/m... might be of interest.
To answer your questions: the MAPE calculations only differ in the data which is considered. If your hold out MAPE is smaller than your fit MAPE this can indicate that your model seems to perform better on data which was not used for parameter estimation, which is a good thing typically. But you may want to verify this behaviour by adding out of sample data.