I am going to have to see some evidence. Going back to 2018 chart above, I was previously pointing to wrong weeks by one week. So GA did not have 2 losses going into CFP only 1. So week 14 is next to last week of ranking.
LSU, OSU and Clemson in top 3 spots. Number 4 is GA with 1 loss (hidden by pointer). Number 5 & 6 have 1 loss. One might say teams 4,5 and 6 are virtually tied at 1 loss and maybe that spot is not locked in. Then week 15 hits and 2 of the 3 teams lose and this leaves 1 team with 1 loss. They go to the final forth spot. I don’t see 2018 as challenging the theory. Team 4 did change but was it really team 4? Yes, so a conference game did make a difference in the top 4.
But remember I also said that in the case of indecisiveness that the conference game should be used to determine the winner. That is written right in the rules. And that is the one exception I support for using conference game to determine a single spot in the top 4. They did it that year and I think it was right to do. And it fits the theory.
For 2021 there are zero barbs to deviate from theory as all 4 teams deserved to be there in the minds of the committee IMO based on their records going into final week AND given there were no spots in question that required the championship to determine if the team belongs. That is my opinion and one I think the committee shares. The results of championships will always be used to reorder the top 4 teams into spots #1-4.