Forecast accuracy calculator
How good was the forecast, really?
Paste two columns from your own data and get the three numbers that matter: the average error, the volume-weighted error, and whether you lean high or low. Bias is the one that quietly costs money.
All free toolsTwo columns, one row per interval — forecast first, then actual. Paste straight from a spreadsheet; tabs, commas and spaces all work.
5 rows read
The formula
How this is calculated.
- 1MAPE = mean of |forecast − actual| ÷ actual
- 2WAPE = Σ|forecast − actual| ÷ Σ actual
- 3Bias = (Σ forecast − Σ actual) ÷ Σ actual
Before you trust it
What this model assumes.
- MAPE treats a quiet interval and a peak interval as equally important, and it cannot handle intervals with zero actuals — those rows are skipped. WAPE weights by volume, which is usually the fairer measure of a day.
- A low error with a consistent bias is worse than it looks: it means you are reliably wrong in one direction, which compounds across a roster.
This runs entirely in your browser — nothing you type is sent to us. It is arithmetic on your own figures, not a benchmark or a recommendation.
In the platform
Where WFMPro does this for you.
These tools answer one question at a time, from numbers you type in. WFMPro runs the same maths continuously against your live data.
Run this on your own data with us.
A working session on your forecasts, schedules and shrinkage — and where WFMPro would take it from there.