How ESPN's Preseason Power Ratings Stand

I'm not sure if there's anyone who takes ESPN's Football Power Index, or FPI, seriously outside of Bristol, Connecticut. But the people inside Bristol, Connecticut refer to it enough that it's probably worth taking a look at.

FPI is the worldwide leader's attempt to have its own advanced metric like S&P+, FEI, or F/+ that it can control. It's largely redundant, but I understand that impulse. If I ran ESPN, I wouldn't want to rely on, say, Bill Connelly's S&P+ for offering a modern, non-opinion poll based rating. What if Bill had some kind of catastrophic data loss and couldn't deliver the ratings some week? ESPN would be caught without being able to offer anything. Besides, Disney has probably made more money selling Elsa and Anna dresses than SB Nation has made in its history. It'd look pretty bad if the company was relying on something from the employee of a much smaller competitor.

Anyway, ESPN describes FPI as follows:

The Football Power Index (FPI) is a measure of team strength that is meant to be the best predictor of a team's performance going forward for the rest of the season. FPI represents how many points above or below average a team is.

A perfectly average team rates a 0. Any non-zero number shows how many points better or worse than a perfectly average squad that the team in question is. For example, Alabama's preseason FPI figure right now is 23.3. If Bama played a perfectly middle-of-the-road team 10,000 times, the average final score of all of those games would be a 23.3-point Crimson Tide win according to this formula.

To come up with a preseason rating, you necessarily have to do some projecting. ESPN describes its method for that like this:

The model takes into account the last four seasons of performance using ESPN’s efficiency ratings, with the most recent season counting most; information on offensive and defensive returning starters, with special consideration given to a returning starting quarterback; a four-year average recruiting ranking of four systems (ESPN, Scouts, Rivals and Phil Steele); and head coaching tenure.

Unfortunately, neither of these descriptions tells us exactly how the sausage is made. These are hints, but we don't know how the weighting of the factors work. Not every advanced stat is an open book, but FPI being a black box hasn't helped its adoption rate.

Anyway, we're going to hear a lot about FPI from ESPN, so we might as well get familiar with it. Here, then is how the conferences stack up in FPI:

The SEC and Pac-12 were the best two conferences last year and should be again this year. I was a bit surprised to see such a big gap between them here, so let's go over why it's there.

The lead the SEC has over the Pac-12 has to do with some rather optimistic projections for some SEC teams. Texas A&M rates 8th overall, for instance. Vanderbilt finished 96th last year, but presumably thanks to three strong years before that, the Commodores come in at No. 51. The Pac-12 has two teams in the 50s with the Washington schools and two others below that with Colorado at 67 and Oregon State at 80. Kentucky at 47 and Vandy at 51 are the lowest the SEC goes.

It didn't surprise me to see the SEC rate the best in defense, though again, the magnitude is bigger than I would have thought. What did surprise me was to see the conference somehow come out on top for offense. The league should have some good ones, but I would have thought the uncertainty at quarterback in so many places—plus only one truly great offensive year for the league in the past four, that being 2013—would have landed the SEC behind at least the Pac-12, if not the Big 12 as well.

So this is how ESPN's attempt at an advanced stat rating sees the country heading into this season. It won't dictate the network's narrative, but it'll play into it for sure.