Percentage Error Calculator
Instant Results ⚡
Find out how far off an estimated or measured value is from the exact value. Used in science, engineering, and everyday calculations.
Find out how far off an estimated or measured value is from the exact value. Used in science, engineering, and everyday calculations.
Percentage error measures how inaccurate an estimated or measured value is compared to the true or accepted value. It is widely used in science experiments, engineering tolerances, and quality control to express how close a result is to the expected value. A smaller percentage error means a more accurate measurement.
Follow these steps:
Formula:
|Estimated − Exact| ÷ |Exact| × 100
The absolute value is used because percentage error expresses the size of the error, not its direction. A separate signed result is also shown so you can see whether the estimate was too high or too low.
Example 1
Measured value: 98 — True value: 100
|98 − 100| ÷ 100 × 100 = 2% error (underestimate)
Example 2
Measured value: 9.5 — True value: 10
|9.5 − 10| ÷ 10 × 100 = 5% error (underestimate)
Example 3
Measured value: 47 — True value: 50
|47 − 50| ÷ 50 × 100 = 6% error (underestimate)
What is the difference between percentage error and percentage difference?
Percentage error compares a measured or estimated value to a known exact value —
there is always a "correct" answer. Percentage difference compares two values with no
defined correct one, using their average as the base.
Can percentage error be negative?
The standard formula uses absolute values so the result is always positive. However,
a signed version can tell you the direction — negative means the estimate was too low,
positive means it was too high. This calculator shows both.
What is a good percentage error?
It depends on the context. In many school science experiments, under 5% is considered
acceptable. In precision engineering or pharmaceutical manufacturing, errors must often
be below 1% or even 0.1%.
What if the exact value is zero?
Percentage error is undefined when the exact value is zero, since division by zero is
not possible. An alternative error metric would be needed in that case.