difficulty 2/5
MAPE
The average error expressed as a percentage of the true value — scale-free and intuitive, and unusable when the target can approach zero.
Intuition
Being 10 units off matters differently depending on what you were predicting. On a true value of 20 it is a disaster; on 10,000 it is noise. MAE cannot tell those apart, because it treats every unit of error identically.
MAPE divides each error by the true value first, turning it into a relative error. The result is a percentage that means the same thing at any scale — and that division is also the source of every problem it has.
Formal definition
How it works
Where it's used
Why & when
Worked example
Pitfalls
Check yourself
3 questions
The questions on this page come with access. They are graded on the server, so the answers stay there until you have given yours.
On the cheat sheet: Regression metrics and losses →
Sources
- scikit-learn — sklearn.metrics.mean_absolute_percentage_errordocs · retrieved 2026-08-30https://scikit-learn.org/stable/modules/generated/sklearn.metrics.mean_absolute_percentage_error.html
- scikit-learn User Guide — Metrics and scoringdocs · retrieved 2026-10-04https://scikit-learn.org/stable/modules/model_evaluation.html
- Forecasting: Principles and Practice, 3rd ed.book · retrieved 2026-10-06https://otexts.com/fpp3/accuracy.html
- Forecasting: theory and practice (2022)paper · retrieved 2026-10-06https://ar5iv.labs.arxiv.org/html/2012.03854