Enter the sum of all absolute errors and the number of results into the calculator (Summary tab), or switch to From Data to calculate common error metrics from Actual vs. Predicted values. The calculator will evaluate the Average Error.

Average Error Calculator

Calculate mean absolute error (MAE) from actual and predicted values, or use the helper tab to solve for MAE, total absolute error, or number of results.

From Data
Formula Helper

Calculate Average Error from Two Lists

Enter matching actual and predicted values separated by commas, spaces, or new lines.

Related Calculators

Average Error (MAE) Formula

AE = SAE / n

Variables:

  • AE is the Average Error (mean absolute error, MAE)
  • SAE is the sum of absolute errors
  • n is the number of results

To calculate Average Error (MAE), divide the sum of the absolute errors by the number of results.

How to Calculate Average Error?

The following steps outline how to calculate the Average Error (MAE).

  1. First, determine the sum of the absolute errors.
  2. Next, determine the number of results.
  3. Next, gather the formula from above = AE = SAE / n.
  4. Finally, calculate the Average Error.
  5. After inserting the variables and calculating the result, check your answer with the calculator above.

Example Problem : 

Use the following variables as an example problem to test your knowledge.

sum of absolute errors = 575

number of results = 30

Average Error = 575 / 30 = 19.1667

FAQ

What is an “error” (residual) in statistics/forecasting?
An error (often called a residual) is the difference between an observed/actual value and a predicted or estimated value.

Why is calculating average error important?
Calculating average error helps summarize how accurate estimates or predictions are. It is commonly used to compare models and track performance over time.

Can average error be negative?
It depends on the metric. Absolute-based metrics like MAE, MAPE, RMSE, and MSE are always non-negative. Signed metrics like ME (mean error) or MPE (mean percentage error) can be negative.

How does the number of results (n) affect the average error?
Increasing the number of results usually makes the computed average error more stable (less sensitive to any single observation). However, it does not guarantee that the average error value will decrease.