Percent Error
Calculator
Calculate the percentage error between your observed (measured) value and the true (accepted) value. Includes absolute error, relative error, signed error, and a step-by-step breakdown.
6%
PERCENT ERROR
0.6
ABSOLUTE ERROR
0.06
RELATIVE ERROR
Enter Your Values
Percent Error = |Observed − True| / |True| × 100Percent Error
6%
Acceptable in many contexts0.6
Absolute Error
−6%
Signed Error
6%
Percent Error
0.6
Absolute Error
0.06
Relative Error
Underestimate ↓
Direction
Step-by-Step Solution
Write values
Observed = 9.4 | True = 10Calculate absolute error
|9.4 − 10| = |-0.6| = 0.6Calculate relative error
0.6 / |10| = 0.06Convert to percent
0.06 × 100 = 6%Accuracy
100% − 6% = 94%Related Error Metrics Explained
Absolute Error
|Observed − True|0.6
The raw magnitude of error, in the same units as the measurement.
Relative Error
|Obs − True| / |True|0.06
Error as a fraction of the true value. Unitless.
Percent Error
Relative Error × 1006%
Relative error expressed as a percentage.
Signed Percent Error
(Obs − True) / True × 100−6%
Positive = overestimate. Negative = underestimate.
Percent Difference
|V1−V2| / avg × 100See Percentage Calc
For comparing two measurements with no known true value — try the percent difference mode on our percentage calculator.
Accuracy
100% − Percent Error94%
How close the measurement is to the true value.
What Is Percent Error?
Percent error (also called percentage error) measures how far an observed (measured or experimental) value deviates from the true (accepted or theoretical) value, expressed as a percentage. It's a staple of science labs, engineering tolerance checks, and statistics, wherever you need to quantify how accurate a measurement actually was — not just how far off it was in raw units, but how far off it was relative to the true value.
The Percent Error Formula
Standard (unsigned): Percent Error = |Observed − True| / |True| × 100Signed (shows direction): Signed % Error = (Observed − True) / |True| × 100Related: Absolute Error = |Observed − True| | Accuracy = 100% − Percent ErrorSigned vs. unsigned: the unsigned version is the standard one taught in school and the one reported as “the” percent error — it strips away direction so a measurement that's 6% too high and one that's 6% too low both read as “6% percent error.” Reach for the signed version when the direction of the mistake matters for diagnosing the cause — for example, if every trial in a series comes out consistently high, that points to a systematic bias (like a scale reading too heavy) rather than random chance.
— FAQ