How this calculation works
Choosing the average as the denominator makes the calculation symmetric. To measure growth relative to a baseline, use percentage change.
100 × |a − b| / ((a + b)/2)A worked example
80 and 120 differ by 40, which is 40% of their average of 100.
Good to know
Use positive values in the same units. Swapping them gives the same result.
Reference: OpenStax Prealgebra (opens in a new tab) ↗
Use the average as the reference
For measurements 80 and 120, the gap is |120 − 80| = 40. Their average is (80 + 120)/2 = 100. Dividing the gap by that average and multiplying by 100 gives 40%. The formula treats the two measurements symmetrically, so changing their order preserves the result.
The diagram places both measurements equally far from their midpoint. Each is 20 units from 100 and the full gap is 40 units. The percentage uses that full gap. Using only one of the two 20-unit distances would halve the answer and describe deviation from the average instead.
Choose a denominator that matches the comparison
This comparison suits two positive measurements when neither has been designated as the baseline. If 80 is specifically an earlier value and 120 is a later value, growth relative to the earlier value is 40/80 × 100 = 50%. A decline from 120 to 80 is 40/120 × 100 ≈ 33.33%. Both differ from the symmetric 40%.
The Absolute change in units calculator returns the 40-unit gap directly. Symmetric percentage difference divides that gap by the average, so use values measured in the same units. Positive input requirements keep the denominator positive. For positive values the result is below 200%, approaching that bound as one reading becomes very small relative to the other.
A few questions, answered
Do equal positive values give zero?
Yes. Their gap is zero and their average is positive, so the percentage difference is zero.
Is this the same as percentage error?
Percentage error usually divides by a designated reference or accepted value. This calculation divides by the average of both entered measurements.