0/4 of the CFP top 4 seeds made it to the semifinals
1/4 of the CFP top 4 seeds DIDN'T get thumped in the quarterfinals

That looks like a pretty crappy result to me.

Also, BamaCoug's observations about the overall FPI ratings suggest some serious flaws as well.

However, Indy Coug's link to the prediction tracker website suggests nearly all of the rating systems have similar overall performance, so it's possible predicting college football outcomes is just hard, and nobody has found a significantly better way to do it, in which case it's much easier to argue that the system is just not that good, and they couldn't make it any better even if they wanted to.

See my response to his post for some caveats, though. It may still be possible for a rating system to skew ratings for a few teams at the expense of a few other teams, or a few games at the expense of a few other games, and still look pretty good overall.