Introduced on day 26
Interval coverage
The share of realised outcomes that fall inside a stated prediction interval; a calibrated 80% interval covers about 80% of outcomes in the long run.
In detail
Coverage is measured on data that were not used to build the interval, at the horizon used for the decision, and ideally by period and segment. It must be read together with width (sharpness): a very wide interval always covers. An empirical coverage far from the nominal level, with enough observations that the sampling error is small, signals a broken assumption.