Mean Absolute Error: the plain-English number
Precision and recall apply to classification, where predictions fall into categories. Regression models, which predict continuous numbers like house prices or delivery times, need a different kind of metric entirely, one that measures how far off a prediction was, not whether a category was right or wrong. Mean Absolute Error, or MAE, is the simplest place to start. It takes the absolute difference between each prediction and the true value, then averages those differences across all predictions.
The strength of MAE is that it stays in the same units as what you're predicting, and it's easy to explain to a non-technical audience. If a model predicting house prices has an MAE of $15,000, that sentence needs no further translation, it means the model's predictions are off by $15,000 on average. Every error, big or small, contributes to that average in direct proportion to its size.