Weighted absolute percentage error: total absolute error divided by total actual volume, Σ|actual − forecast| / Σ|actual|. Unlike MAPE it doesn't blow up on small or zero actuals, because each error counts in proportion to volume.
Card 1 of 259. Answer: Weighted absolute percentage error: total absolute error divided by total actual volume, Σ|actual − forecast| / Σ|actual|. Unlike MAPE it doesn't blow up on small or zero actuals, because each error counts in proportion to volume.
About · Quizzes · History of AI · To suggest new questions or report an error send me a dm.
WAPE is the sum of absolute errors divided by the sum of actual values. It equals the mean absolute error divided by the mean of the actuals, which is why it is also called the MAD/mean ratio. Retail and supply chain teams use it because it answers a business question: what share of total units did we get wrong?
Example: a store forecasts three products whose actual sales are 1,000, 50 and 0 units, with forecasts of 900, 60 and 5. WAPE = (100 + 10 + 5) / 1,050 ≈ 11%. MAPE can't be computed at all, because one actual is 0.
Its weakness is the flip side of its strength: high-volume items dominate it, so terrible forecasts for long-tail products barely move the number. Report it by segment as well as overall. And check the definition before comparing numbers across teams, since "weighted MAPE" sometimes means weighting by revenue or margin rather than units.
