I think the FPI is a terrible ranking. Oregon is 6th in the current ranking and Ohio State is 3rd, it even has us at #4. There is no world where that should be the case. USC is 18. I have no idea what goes into that ranking, but it has always had extremely questionable results. My guess is it puts way too much weight into roster talent.
Computers will do weird things, you can't judge them entirely in isolation. The list I posted showed a pattern of antiquated computers over-rating teams with better winning records (Georgia below Indiana and SMU) but the quirks otherwise are part of the feature.
You absolutely need something to blend them together into something more palatable and even then I think there should be a human element. Having said that, if (modern) computers start saying something it is worth looking at why they might be saying that. Humans could be missing something the data reveals.
Sagarin also has Ohio State over Oregon (which does not take rosters into the equations at all). I'm not going to jump too far into the fray there, but Oregon has played a fairly soft schedule as well so Oregon is arguably over-rated but of course one can question Ohio State being over them.
But, look at this ranking:
Calculated by aggregating computer generated college football power rankings from around the web and determining each team’s average ranking.
powerrankingsguru.com
We can disagree on the positions, but what's wrong with the teams in the top 12? Is anything wrong with it? That's probably the best list of top 12 teams if you ignore the positions that I've seen.
So, it's easy to look at computers and forget for a second they are computers and just go man I don't agree with that position, but if you zoom out a little you go wait they are getting a lot of things right. For instance even if we agree Alabama shouldn't be 4th or 5th, they also don't belong below teams like Boise State either.
If I take a computer ranking like that though and then I just blend it with the AP poll for instance (I also believe there should be a human element) , suddenly I get something like this:
#1 Texas
#2 Oregon
#3 Notre Dame
#4 Ohio State
#5 Georgia
Penn State
#7 Tennessee
#8 Alabama
#9 Indiana
SMU
#11 Ole Miss
#12 South Carolina
#13 Miami
Now, we can still argue a bit over those rankings, but it injects a good bit of logic to the rankings. You can see where it fixes some problems with the polls and values SoS over simply winning games. While I would personally swap a couple places around I think that might be the most accurate ranking relative to actually considering SoS I have seen. Perfect? Nope, but a step in the right direction...