I’m not a statistics guru, but I’ve always appreciated what you can glean from stats. One reason I latched on to baseball early on was because there was so much fun to me in following all the numbers players generated. That being said, until Bill James came along to analyze things, I didn’t appreciate there were numbers and then there were numbers.
Compared to football, baseball is a relatively easy enterprise to analyze statistically, because the bulk of the action is individualized. Baseball is much farther along than is football in internalizing the information; analytics rule the day in modern baseball.
That isn’t to say we aren’t seeing a similar effort made for football. But devising a statistical framework that both enlightens and affects strategies and tactics is a more difficult enterprise than is the case for baseball. My first inkling of what you could do with stats came with Matt Hinton’s sadly missed Dr. Saturday blog, where he sought to discover the level of correlation between different metrics and wins and losses. Although it was a somewhat crude approach, it wasn’t without its revelations. For example, Matt was the first person who made me realize that penalties had very little effect on wins and losses.
There have been plenty of others in the years since then, like Bill Connelly and Brian Fremeau, who have added much to the framework. I find myself convinced by some metrics, not so much by others, but I do my best to delve into the expanded body of work.
Stats at their most relevant, I think, accomplish two different things: they provide insight into the relative quality of play and they also can be a useful tool, analytically speaking, for devising a game plan. The former is more useful to me as a fan; the latter should be more useful for a coach. Either way, if the data isn’t presented in a meaningful way, it’s worthless for either purpose. As yesterday’s PFF post indicated, there is a lot of noise in the system and that encourages people to tune out the story good stats tell.
That’s why I posted the bit from the guys at Dawg Sports Live the other day. Sure, college football is enjoyable to follow on its own terms and if you find statistics to be nothing more than a distraction to that, fine. But, if you dismiss statistical analysis because you find it unconvincing, that’s a mistake.
And apparently some of you do feel that way, because Josh felt a need to tweet something in response to some of your comments to my post.
He provides two charts.
This first one is near and dear to my heart. If there’s only one college football stat you choose to follow, make it yards per play. (I’ve even made it easy to do so at the blog, as I’ve started tracking net ypp in the SEC on a weekly basis.) As you can see from that graphic, there is a very clear correlation between net ypp and wins.
Granted, this chart is more in the weeds-y (and it would help if you listened to the linked clip to understand Josh’s data better), but it tracks two of the more important analytic buzzwords of the day, explosiveness and efficiency. (Kirby may not geek out on the numbers, but he’s hammered steadily about Georgia’s offense needing to be more explosive for some time now.) Where this sort of data can help with game planning and play calling is that you can use it to drill down to what works for given down and distance sets — and if you don’t believe there aren’t coaches already out there doing that to gain an edge, buddy, you’re kidding yourself.
I’m not trying to make your head spin with this stuff. I’m just telling you to approach it with an open mind, because I can assure you I will keep referencing it here at the blog. You’ve been warned!
[By the way, I’ve asked Josh, and he’s graciously consented, to answer any questions you might have about his work. So feel free to query away in the comments.]