Forecast Errors

Mean Percentage Error (MPE)

Average of % errors. Shows Bias in % form

Mean Error (ME)

Average of the error terms. Good measure of bias but not accuracy. Negative answer means tendency to over estimate

Thiel's U-Statistic

Compares the forecast technique used to the naive forecast error. The smaller the U the better the technique

Weighted Mean Absolute % Error (WMAPE)

MAPE weighted by magnitude of demand

Root Mean Squared Error (RMSE)

Square root of MSE. Similar to finding Standard Deviation

Mean Absolute Percentage Error (MAPE)

Shows relative accuracy. The average % a forecast is off. Cant calculate if D is ever zero

Mean Squared Error (MSE)

Square each error term and average them. Similar to finding variance

Mean Absolute Error (MAE)

Taking the absolute value of the error terms and averaging them better shows accuracy