One trend I noticed that's mattered in the NBA (past tense because I no longer bet the sport anything like regularly) is teams playing in conference vs out of conference, and when in conference whether it's a divisional or non-divisional contest. This is in the context of totals. One season (to pick an example from long ago, couldn't give you the precise year) I noticed the Wash. Wizards seemed to play very loose ball against all other teams (ppg higher rather than lower, a predominant over record) but when it came to playing divisional opponents, they'd get serious on D and produced a significant Under record. So if you looked at their season O/U record as a whole, it looked pretty much even and you'd not think they had a bias towards either total result. But that aggregated figure lied. Since all 4 major sports have that basic structure (2 leagues/conferences, then divs inside of leagues/conferences) I'd posit it would pay to pay attention to a teams ATS/SU/Totals record parsed out among those 3 categories, since this delineation isn't purely based off of betting lines (aka "a teams ATS record when a game is lined over 200 pts for the total kinda BS) but rather is grounded in reality: divisional rivalries still mean something to teams whose seasons might not otherwsie be going well, which means it's harder to pick for when they give a shit.