Calculators

Weighted Absolute Percentage Error Calculator

Calculate WAPE from actual and predicted pairs privately in your browser, including zero-actual observations.

Measure weighted absolute percentage error (WAPE) from paired actual and predicted values. Your data stays in your browser.

Enter one pair per line: actual, predicted. Separate the two values with a comma, semicolon, or space. Zero actuals are supported when at least one actual is nonzero. Errors on zero-actual rows remain in the numerator. Signed numbers are accepted. Nonzero magnitudes must be between 1e-100 and 1e100. Up to 10000 pairs; scientific notation supported; omit thousands separators. Use consistent units.

Your prediction error results will appear here.

Measure total forecast error relative to total actual magnitude.

WAPE = 100 × Σ|predicted − actual| / Σ|actual|. Unlike MAPE, it divides total absolute error by total absolute actual values rather than averaging row percentages. For actual/predicted pairs (100, 90), (200, 220), and (50, 50), WAPE = 100 × 30 / 350 ≈ 8.57142857143%. Zero-actual rows contribute their prediction errors; the result is undefined if every actual is zero, even when predictions are also zero. Signed values are accepted using absolute actuals in the denominator. Lower is better; WAPE can exceed 100% and is not an accuracy score. Large actual magnitudes have more influence, so compare forecasts on the same observations and units. Enter up to 10000 pairs, one per line, separated by a comma, semicolon, or space. Scientific notation is accepted; omit thousands separators. Nonzero magnitudes must be from 1e-100 to 1e100. Scaled arithmetic and compensated sums reduce rounding error; results show up to 12 significant digits. All inputs stay in your browser.

Frequently Asked Questions

Everything you need to know about this tool, how it works, and privacy.

How is WAPE calculated?

Divide the sum of absolute prediction errors by the sum of absolute actual values, then multiply by 100. Errors of 10 and 20 with actuals of 100 and 200 give 10%.

How does WAPE differ from MAPE?

MAPE averages individual percentage errors. WAPE uses a single denominator: total absolute actual values. It therefore supports individual zero actuals and gives larger actual magnitudes more influence.

What happens when actual values are zero?

Zero-actual rows still contribute prediction errors. If all actual values are zero, WAPE is undefined and the calculator displays an error.

Can WAPE exceed 100%?

Yes. WAPE is unbounded above. Lower values indicate smaller errors; subtracting it from 100 does not produce a general accuracy score.

Is a separate guide needed?

The formula, worked example, input limits and zero handling are explained on this page. Calculations run entirely in your browser.