Y'all, I'm sorry, but this analysis is garbage. In a perfect world, I'd like the PAC to survive and thrive, because I like regional conferences. But there's so many biases in this data it's ridiculous. He did one good things (throwing out games against USC/ UCLA/ UT/ OU and 2020 games altogether), which is just enough to make someone who doesn't understand stats happy, but then the rest is dog****.
A short list of immediately problematic biases:
- He's using more Big XII games than PAC games despite the Big XII have fewer games. This automatically means he's including a larger percentage of low-ranking Big XII games than PAC ones.
- He's not including lowest-tier games from both, but for Big XII teams that was typically one game per year; for the PAC it was many more shown on the PACN, accentuating point #1.
- He's not controlling for network and timeslot, which is wildly variable. Games on ESPN and Fox draw dramatically (a million +) better than games on FS1, etc.
- Contributing to the last point, in a point he almost acknowledges, is that the late timeslot gives the PAC an uncontested window that drives up viewers already, but since it's uncontested it automatically gets those games on better networks, which really drives it up.
- You can read the last point and think "well then PAC is obviously better" but not so fast my friend. If this is a situation where theoretically the Big XII might take a few PAC teams to join BYU out west, that advantage immediately goes away.
- Thus what matters is comparing apples to apples by controlling for timeslot and network, seeing which conference does better when all is equal, and then getting the top teams from the worse conference to join the better.
-
All of this and more is why Navigate Research, which actually crunches these numbers as a business so as to inform conferences and media companies of what is realistic, said in March that the new Big XII would barely be behind the then 12-member PAC 12.