A way of computing the quality of a collection of measured deviations, or errors, compared to the ideal case. An error metric is expressed as a single number, which allows for comparison between deviations.