Criterion Choice
What one criterion preferred, and whether the range of counts it was given decided that.
A preference at the smallest count fitted is usually a real answer, and a warning on it would be noise. The case worth reporting is narrower: the preference sits at the smallest count while the criterion is still improving at the largest. On a sample of 10,000 production lead times searched over one to six components, BIC preferred one component while still falling at six, and carried to twelve it crossed below the one-component value at ten. The range decided that answer, and isUnfinished says so.
Parameters
the criterion's name
the count it preferred, or null when no count had a comparable value
where that count sits in the range fitted
whether the criterion improved from the second-largest to the largest count with a comparable value, in its own direction; false with fewer than two such counts