And shows absolutely nothing. That’s not how it works (or how it should work), at all.
Of course, none of this matters. It’s not being used for anything right now, it is only attempting to give the best info it can based on an extremely limited dataset. But your specific UW vs Wazzu example is terrible.