How it works
Absolute error = |measured−reference|. Relative error = absolute error / |reference|; percentage error is 100×relative error.
Input definitions
- Measured value
- Enter the value in the labeled unit.
- Nonzero reference value
- Enter the value in the labeled unit.
A worked example
- Measured value
- 98
- Nonzero reference value
- 100
Assumptions & limits
Nonzero reference, both values in the same units. Measurement uncertainty is not estimated.
Sources & review
Maintained by AnsRio. Methods and examples use automated checks; this does not imply independent professional certification.
Method updated 2026-10-17. Content checked against the stated method and examples. Our verification approach.
Common questions
Can I calculate percentage error when the reference is zero?
No. Relative and percentage errors divide by the absolute reference and are undefined at zero. Absolute error alone can still be meaningful in that situation.
Why can statistical results differ between tools?
Denominator, interpolation and boundary conventions can differ. Use the method and assumptions on this page to compare the same definition; displayed decimals may also be rounded.
How does the Absolute and relative error tool work?
Absolute error = |measured−reference|. Relative error = absolute error / |reference|; percentage error is 100×relative error.
What are the assumptions and limits of Absolute and relative error?
Nonzero reference, both values in the same units. Measurement uncertainty is not estimated.