There are 130+ FBS teams and only 12 games. 8-9 of which are in conference. It is not possible to create a SOS metric that can precisely rank the quality of teams based on who they play. You can only sort them broadly based on quality. Yeah, the top P4 teams and teams that look like those teams are really good, they look better than the other good and pretty good looking teams. This is the committee model in a nutshell, the pomp is just covering their decisions. You will note how the chosen metrics shift year to year and team to team. Same happens in BB but they use more computer rankings.
Everyone assumes the SEC teams are good because they play other SEC teams. The reality is the top SEC teams are good because they get the best players so they are inclined but not guaranteed to be among the best 5 or 6 at large teams. 2 or 3, probably. But, it's all assumptions.
The only thing we know about the SEC is that Georgia was the best team in it this year. They won the championship and beat the next closest contender twice. We assume Texas was next best based in their record, but they didn't play as many top half conference teams as Bama or Tenn. But they went undefeated in conference except for UGA, so we assume and rank them as the SEC #2.
You can do the same for every P4, and the committee discounts them for number of losses and the perception of their conference vs the others. Reputation and the eye test. It's all subjective.
The best system is based on conference championship with some at large teams to account for the small sample of games, the lack of similar schedules and the any given Saturday effects.
Well, that's what we currently have. A system that values championships and allows for others to play their way to the NCG. Just because the SEC and ESPN didn't like this year's result and CFB is more about hot takes than playing the games.
This will be the first year that we get an uncontestable deserving champion. Try to enjoy it.