Work out forecast accuracy instantly with clear inputs, formula shown and shareable results.
MAPE expresses error as a percentage of the actual value, which makes it comparable across series of different scales but explodes when actuals approach zero. MAE keeps the original units and treats every error equally, while RMSE squares errors and therefore punishes large misses harder. Bias is the signed average error and reveals systematic over- or under-forecasting that the absolute measures hide.
Accuracy metrics
MAPE = mean(|forecast - actual| / |actual|); MAE = mean(|error|); RMSE = sqrt(mean(error^2)); bias = sum(error) / sum(actual)
Report bias alongside one magnitude metric. A forecast with low MAPE but consistent positive bias is quietly accumulating excess inventory or capacity.
MASE, which scales error by a naive forecast's error, or weighted MAPE computed on totals rather than per period.