Originally Posted by saevel25
But what if those other teams are beating teams that are terrible? Then that means those teams are not as good as their record indicates right? How many levels of regression do you need to get a complete picture.
Beating a mediocre team doesn't mean you're worse than another team that beats a good team. There are really only two outcomes in College Football that tell you a lot about the team in question:
1. Losing to a bad team, and
2. Beating a good team.
If you have teams from different conferences with similar records, and both have one loss to a good team (i.e. Stanford, Clemson, and Missouri), then you can't really know which team is better until they play each other or have a sufficient number of games against common teams.
And I think there's an inherent conference bias in favor of SEC teams and traditional powerhouses in the rankings (though Mizzou's ranking doesn't reflect it). Look at Aggy vs. Oklahoma State. Aggy is ranked higher with two losses than Okie State with only one loss. Who has Aggy beaten? Ole Miss? Vandy? These are not great teams. Okie St has at least beaten one "ranked" team (Texas Tech), but Tech's ranking was crap too. They made it to #12 based on six straight wins against scrub teams, then lost three straight. But Okie St still gets credit for a win against a #12 team because that's what they were ranked at the time, even though the ranking didn't reflect the team at all.
Mizz should be ranked higher than they are. They won at UGA, beat Florida, and their only loss was in OT against a very good SoCar team. I think they're lower than they should be because of where they started and the fact that they're Missouri, not Texas or Oklahoma or Alabama or Florida.