Forecast error is the gap between the actual value and the forecast value. The MAE (mean absolute error) is the average of the gaps in absolute value and indicates the typical size of errors. The bias is the mean of signed errors: a positive bias means demand is under-forecast, a negative bias that it is overestimated.
Example
Actual 100, 120, 110 and forecast 90, 130, 100: errors 10, -10, 10; MAE = 10 and bias = 3.3.