This data treatment controls for the irregularity that P5 teams appear on an Indy schedule. If the point is to try to measure how BYU measures up to P5s *if it were a P5* (even without the funding), you gotta try to replicate a regular P5 scheduling scenario. P5s play each other in conference basically once a year. This data set attempts to match that.

Suppose Utah were Indy, and had a yearly rivalry game with Washington (1 and 7), but could only schedule the rest of the P12 teams once a year for the last decade on the road. How would that look for Utah's status against P5 competition? Now, if you think Washington is an unfair stand-in here, the support gap between Utah and BYU is no different. But if you insist, take Arizona State instead (3 and 6). Let Utah play Arizona State yearly for a decade, the rest once a year on the road, and go run the numbers. How would Utah stack up? Not quite a fair analysis, I'd say.

So the first data is very far from a generalizeable sample: It's pear shaped, skeewompas, katerwompas... it repeats certain teams far far more than others. The first data set measures head to head matchups if it measures anything at all. It can't get to the general point about how we stack up against P5s. Adding road and home additions to the set was also an improvement from the original. But still leaves us with matchups rather than general info about P5 status.

This second data set also confirms my priors: BYU is a mid level P5, despite not having P5 support.

One more point: I'd bet that if you put us in the Big XII we'd scale up very nicely. Start beating out Stanford for culture matching recruiting talent? Forget about it. (But as we've seen as recently as yesterday, Sitake is overperforming in the recruit category as it stands.)

Also, if BYU were to play their 6 scheduled P5s in 2020... Well, I think the data would look pretty good for BYU this year on either approach.