No stat isolates for the performance of the other players on the field. You might as well argue that YPA, completion percentage, or any other stat. Even in baseball you can't fully isolate a hitter's or pitcher's performance over what is contributed by opposing players and noise. The fact that you can't isolate a player's skill from noise and the contribution of other players' skills will always be true and pointing that out is not in any way profound or interesting. It's just an obvious statement that has no bearing on anything I've said on this thread, yet you keep bleating it.
As for SP+ underrating Herman's teams in prior seasons, I'm not aware of Sushi's post but I don't think SP+ has a history of underrating Texas. For instance, S&P+ had Texas going 8-4 in it's pre-season 2017 projected record. So, relative to SP+'s preseason expectations, during Tom's tenure at Texas it overrated his team one season, underrated it another, and just about nailed it in the third.
https://www.sbnation.com/college-football/2017/5/17/15634960/texas-longhorns-football-2017-preview-schedule-roster
More importantly though, the fact that Tom had overperformed relative to SP+ in his time at UH was no basis for arguing that you should expect Tom to continue to perform relative to that metric. It was probably just noise!