How is MAD calculated?

• A. MAD = ∑(|A - F|)
• B. MAD = ∑(|A - F|) / n
• C. MAD = ∑(A - F) / n
• D. MAD = ∑(|A - F|)² / n

Answer: (B)

MAD measures the average size of forecast errors by taking the absolute difference between what actually happened and what was forecast, and then averaging those differences. In practice, you compute the absolute error for each period, sum them, and divide by the number of periods. The absolute value is crucial because it prevents positive and negative errors from canceling each other out, giving a true average magnitude in the same units as the data.

That’s why the correct form is the sum of absolute errors divided by the number of observations. If you didn’t divide by n, you’d have the total error instead of the average. If you didn’t take absolute values, positive and negative errors could cancel. If you squared the errors, you’d be measuring a different metric (variance-related) rather than MAD.