What is the average of the absolute values of the forecast errors called?

The average of the absolute values of the forecast errors is referred to as the mean absolute error (MAE). This measure is calculated by taking the absolute differences between the actual values and the predicted values (forecast errors), summing these absolute differences, and then dividing by the number of observations.

The reason this metric is valuable is that it provides a straightforward way to assess the accuracy of a forecasting model, as it treats errors uniformly, regardless of their direction (overestimations versus underestimations). This uniformity eliminates any cancellation that may occur with positive and negative errors when calculating the average without taking the absolute value. As such, MAE gives a clear indication of how much, on average, the predictions deviate from actual outcomes, making it an essential tool for performance evaluation in forecasting models.

Other measures mentioned, like mean squared error or root mean squared error, involve squaring the errors and therefore emphasize larger discrepancies more than smaller ones. While they also serve useful purposes, they do not reflect the average of the absolute values of the errors directly, making mean absolute error the correct answer in this context.