538 has done an analysis of how highly ranked teams early in the season fare later and what it means for the perception of a good win. The basic premise is highly ranked teams often falter after an early loss and thus should they be considered a quality win. 

This calls into question haw should those early wins be considered in determining the CFP selection. Should SOS be determined based on rankings when the game was played or once the overall records and final rankings come out? The selection committee values SOS but it isn’t clear what they view as SOS in their final decisions. They are supposed to ignore pre-season polls but do they contribute to a bias when deciding who are the best teams a team has beat?

538 does some analysis on how teams fare against teams highly ranked early vs late; and the picture is interesting. Bama, for example, is 7-2 in road games against teams that are in the top 10 when played and 4-7 vs teams the end up in the top 10. One question is does the loss result in a drop in performance or were the teams just not good enough to deserve the high ranking.

Interesting read: 

https://fivethirtyeight.com/features/good-wins-in-college-football-dont-...