There are a few confounding factors to consider here though. If you count the number of teams in and out of Power 5 confs which are off the line, and do it on both sides of the line, you see a trend that is telling. The split is something like 3:1 for Power 5 teams underachieving, and the opposite 1:3 for overachieving. That sure seems like a product of the test or process, and not a random result which was truly based on how well a team did with what it had. I am guessing this is more do to the vagary of recruiting much more than Massey's rankings. To be sure, there are non random influences about rankings which could help produce skewed results (like the fact that Power 5 teams are much more likely to be ranked when they don't deserve it, compared to non-Power 5 teams). However, rankings are probably even more skewed. Recruiting rankings are largely marketing efforts to make money from the fans of Power 5 schools, and they are certainly extremely relative to the schools involved. I mean they start identifying good players in large part by how many, and which, Power 5 school is looking at the kid. A players ranking even moves based on who offers him a scholi, and which teams he signs with. Other than the Top 100 or so kids though, the rest is probably a complete coin flip, and the ratings are actually more determined by the interest of a Power 5 school, than anything else. That means the Power 5 schools are actually overrated for their recruiting rankings every year (especially the bottom half of the Power 5, where most of the underachievers are), which creates a systematic underachievement in that graph.
It's similar to that graph that came out sever years ago trying to show trends on how teams move up the rankings from where they start. Of course it showed that major programs were not moving as well as lesser known programs. This was obvious once you thought about it, because the major schools started the year ranked high, and had no where to move up the rankings. Of course this graph is far superior to that old one, but this graph also seems to have confounding factors biased against Power 5 schools as well. Just something to keep in mind. You can still laugh at Tenn, but keep in mind that kid with a 3 or 4 star on their teams may really just be a 2 star if he went to TCU, and TCU looks all the better for it.