It felt to me like you were asking whether there is an admission advantage beyond in-state admissions opportunities.
The recently cancelled Landscape product is a clue to how at least many colleges might think about this.
First, Landscape looked at the area context of both the HS attended and the residence of the applicant.
Second, it did so at a granular level, specifically the census tract level. That is because a lot of potentially relevant area data is reported at the census tract level.
So while I wouldn’t advise this in the first place, if someone WAS trying to “game” all this for college admissions purposes, you would want to game both high school and residence, and you would be gaming based on census tract data and not state-level data.