Alright, folks. We’re back. College football started this weekend, and that means you NEED someone to tell you what you watched for 12 hours yesterday. Well, you’re in luck, because I’m going to be that someone for College and Mag. If you remember the dregs of 2015 and 2016, I got started on the site by compiling and writing about my own rankings system, the Sterritt Score. It was born out of summer boredom in 2015 (what else was I going to do, my OChem homework?), and inspired by a series of articles by our old friend WarRoomEagle pondering the definition of a “knockout” in college football. You can go back and read some of his writing here and here.
I preferred Method 2 from WRE, which defined a knockout as “the last point in the game where the losing team possessed the ball while still having the ability to tie or take the lead (aka within 8 points)”. Obviously there are issues with just using these KO times to evaluate teams, but I thought it was an interesting way to show how in-control a team was in each game. I combined this metric with average margin of victory (MOV) to create a sort of “power ranking”, since it showed, very basically, how long a team controlled the games, and by how much they dominated. Teams with a positive MOV and KO times would have positive ratings, while teams with a negative MOV and KO would have negative ratings, with 0 being a truly average team. Over time, I also included a rudimentary strength of schedule (SOS) modifier to skew the rankings towards the better teams.
But, that’s enough of a history lesson.
It’s 2018, and after a two-year absence, the Sterritt Score is back and better than ever. A few tweaks have been made to the formula, and now not only will overall rankings be available (after the first few weeks once some legitimate data is available #TeamNoSeptemberPolls), but game scores will also be generated from the basic formula to help show how each team’s performance in their games stacks up. Also, it might be a bad time to admit this, but I’ve upgraded my SOS calculation to using RPI instead of just opponent’s record. Like I said, VERY basic.
I’ll be the first to admit that these rankings aren’t going to be the most robust and mathematically sound things you’ve ever seen. It might put Navy in your top 5 in November. It might not reward a team that’s ranked #1 in the AP Poll but that’s been racking up BS 3-point last second wins over a relatively weak schedule. But what it will do is bring to your attention teams that aren’t performing at a level everyone else things they are. If Wyoming, despite a being in a G5 conference, beats everyone they play by 40 points early in the game, they’re still going to be ranked highly. And at the time of writing this, they’re up 29-0 in the 4th quarter on New Mexico. If Auburn is 7-1 going into November but hasn’t beaten anybody good and was blown out by Washington, they probably won’t be ranked highly. But you know what? This is a dumb, stupid, random sport, and even the best predictors can’t beat the spread much more than 53% of the time. The fun in it isn’t in absolutes, it’s in the stories.
Something else to note is that this won’t be Newtonian physics. Not every action will have an equal an opposite reaction. If Alabama (with a strong RPI) goes out and stomps Mississippi State (also with a good RPI), that will be a great win for Alabama! But losing to Alabama, regardless of how badly, also won’t be the end of the world for Mississippi State. I also won’t be including games against FCS opponents, simply because I don’t have a good way of measuring RPI for them right now. Theoretically I could just assign all FCS teams the same RPI value (equal to the worst FBS team possibly?), but that’s not really a fair assessment and I’d rather just not use the bad date that would generate.
At this point, if you’re still reading, I’ll give you a little taste of what to expect. These game scores don’t mean much since everyone has an RPI of 0.5 so far, but it’s fun to see numbers either way.
|New Mexico State||IND||-27.89|
By the way, I think I need a new name. I’ve tweaked the formula enough times since the first rendition, and after a two-year absence, I need to spice it up a little bit. How about… “The Sterritt Score+”?
I think I’ll drop the “The”.