MAPE vs WAPE Analyzer

Compare MAPE and WAPE forecast accuracy metrics side by side on the same data. See exactly how low-volume intervals inflate MAPE and why WAPE gives an operationally meaningful picture.

This tool computes two competing accuracy metrics side by side to demonstrate their divergence. MAPE (Mean Absolute Percentage Error) calculates |actual - forecast| / forecast for each interval independently, then averages all intervals equally. WAPE (Weighted Absolute Percentage Error) sums all absolute errors across intervals and divides by the sum of all forecasts: sum(|actual - forecast|) / sum(forecast). Both measure forecast accuracy, but they weight intervals differently.

The critical assumption is that low-volume intervals inherently produce high percentage errors due to the MIV floor (sqrt(2/(pi*n))). An interval forecasting 8 calls that receives 12 shows a 50% error — but that 4-call miss has negligible operational impact. MAPE treats this interval identically to a 500-call interval, inflating the reported error. WAPE weights by volume, so high-volume intervals (which drive staffing decisions) dominate the calculation. The gap between MAPE and WAPE directly quantifies how much low-volume noise distorts your reported accuracy.

Practitioners should report WAPE for operational decision-making and stakeholder communication — it reflects the error that actually affects staffing and service levels. MAPE is only useful for diagnosing whether low-volume intervals need attention (e.g., should they be consolidated?). If MAPE exceeds WAPE by more than 5-8 percentage points, the forecast isn't broken; the measurement methodology is. The MIV floor displayed shows the theoretical best-case — any accuracy metric below this threshold is unachievable and shouldn't be targeted.