Because this resets each interval, watching it climb toward the expected full count for the interval length tells the admin, before the interval closes, whether enough valid data has actually been gathered for the pending Alert TCA state to be a meaningful decision rather than one based on a sparse handful of samples. For example, seeing only 2 out of an expected 12 five-minute samples processed partway through an hour-long interval would warn the admin the eventual alert decision may rest on thin data.