Calculators

Mean Bias Error Calculator

Calculate mean bias error and mean absolute error from actual and predicted pairs privately in your browser.

Measure mean bias error from paired actual and predicted values using predicted minus actual. Positive means average overprediction; negative means average underprediction. 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 and predictions are supported. 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 the direction of average prediction error.

Mean bias error (MBE) = Σ(predicted − actual) / n. This page uses predicted minus actual: positive bias means average overprediction and negative bias means average underprediction. Some sources use the opposite sign, so check conventions before comparing. For actual/predicted pairs (100, 90), (200, 220), and (50, 50), MBE = (−10 + 20 + 0) / 3 = 3.33333333333 input units. Mean absolute error (MAE) is 10 input units for those pairs. Opposing errors cancel in MBE; zero bias does not imply perfect predictions. Use MAE alongside bias to assess error magnitude. Zeros and signed values are accepted. Enter up to 10000 pairs, one per line, with comma, semicolon, or whitespace separators. Omit thousands separators. Scientific notation is supported; nonzero magnitudes must be between 1e-100 and 1e100. Compensated summation reduces rounding error; results show up to 12 significant digits. All calculations stay in your browser.

Frequently Asked Questions

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

How is mean bias error calculated?

Average predicted minus actual across all pairs. Errors of −10, 20 and 0 give MBE of 10/3, or approximately 3.33333333333 input units.

What does the sign mean?

Positive MBE means average overprediction; negative MBE means average underprediction. This calculator uses predicted minus actual. Other sources may reverse the sign.

Does zero bias mean perfect predictions?

No. Errors of −10 and 10 cancel to zero bias, while MAE is 10. Use the reported MAE to measure absolute error magnitude.

Can actual values be zero or negative?

Yes. MBE has no percentage denominator, so zeros and signed values are accepted. Results retain the units of the inputs; use consistent units across rows.