A forecast is only useful if you know how wrong it tends to be. The forecast error calculator takes the actual values and forecasts of six periods and gives the mean absolute percentage error (MAPE), the mean absolute error (MAE), the root mean squared error (RMSE), the mean squared error, the bias, the weighted absolute percentage error (WAPE) and the tracking signal.
How the calculation works
- Enter the actual value and the forecast for each of the six periods, in the same units.
- The error of a period is actual − forecast. A positive error means the forecast was too low.
- MAE averages the absolute errors, RMSE gives more weight to large errors, and MAPE expresses the average error as a percentage of the actual value.
- The bias is the mean signed error, and the tracking signal is the sum of errors ÷ MAE. A tracking signal outside ±4 points to a forecast that is consistently too high or too low.
Example
The errors of the default series are −5, +5, −4, +4, +5 and −10. The MAE is 33 ÷ 6 = 5.5, the MSE is 207 ÷ 6 = 34.5 and the RMSE is 5.87. The MAPE is 4.07 % and the WAPE is 4.06 % (33 ÷ 813). The bias is −0.83 and the tracking signal is −5 ÷ 5.5 = −0.91, well inside ±4, so the forecast is accurate and not persistently biased.
Why it matters
Each metric answers a different question: how big the typical miss is, how bad the worst misses are, how large the error is in relative terms and whether the forecast leans one way. Comparing them tells you whether to fix the model, the data or the inventory buffer.
Good to know
Six periods are a small sample, so treat the result as an indication and not a proof. MAPE is undefined when an actual value is zero, so enter a small number instead. The tracking signal here is computed over the whole window, while in practice it is updated every period.
Related tools
The demand forecast calculator builds the forecast, the exponential moving average calculator smooths a series, and the percent error calculator handles a single measurement.