treated the same as the difference between playing BYU and Ohio State? I'm pretty sure that's not how Sagarin does things, although his formula isn't public. He's not dumb enough to treat those gaps the same.

Furthermore, the teams that are at issue here do make a difference. Clemson beat UNC (Sagarin's #53 team) 21-20. Does Clemson win that game if it's against a mid-level P12 team (Washington State, Arizona State, Stanford, UCLA, who are the #5-8 teams in the P12), all of which Sagarin rates higher than UNC?

What about Georgia losing to #44 South Carolina? If that game had been against Purdue, Maryland, Rutgers, Illinois, etc., that's a win, even though looking at that South Carolina game going in you wouldn't think it would matter to Georgia whether it was playing South Carolina or NC State. Georgia will win either way, right?

Penn State narrowly beat Sagarin #47 Pitt. What if that were a slightly better opponent?

Baylor beat Rice by 8, Iowa State by 2, Texas Tech by 3, and West Virginia by 3. You replace Rice, Texas Tech, or West Virginia with a mid-tier P12 team and Baylor is no longer undefeated.

If Utah had played one of those mediocre (not even the bad ones) ACC teams of B12 teams in the 40-60 range instead of #28 USC, it would be undefeated.

This is where the difference is; there's a big difference between playing a slate of teams in the 30-50 range as opposed to teams in the 50-70 range, even though all of those games should be wins for a top-10 team. These teams' relative rankings are not a result of playing one FCS team as opposed to a bad G5 team. The bulk of the difference is cumulative and based on playing a bunch of games where you only have an 80% chance of winning rather than a 90% chance.