for example, Portland opened +1, I stepped away for a few minutes and they were at -1.5, a little similar to Utah yesterday.
typically teams that cross the 'mendoza line' (from +1 to -1 etc) do well, or did ATS. I don't do much database stuff, but it might be interesting if it was easy to exclude those teams (or maybe the opposite).
What you are pointing out does make sense IMO. Typically the betting public loves to jump on 'easy winners', a short spread on a good team. I bet a fair amount of those SU winners are home dogs.
I'm not 100% but if I remember correctly teams that move from dog to favorite are 4-1 this year, if that makes more sense. So it doesn't happen all that much, and usually early on in the year far more frequently than later, but if you have the ability to manipulate a db (open vs close) it may (or may not) increase the yield.