Category Archives: Stats Geek!

It’s gonna be a long season.

I always hope when I see articles like “Reverse Engineering a Champion”, there will be some really cool insight that leads to an unexpected conclusion.

I really shouldn’t get my hopes up like that.


Filed under College Football, Stats Geek!

Strength of schedule, 2015 edition

Talk about your picture being worth a thousand words… take a look at this graph, particularly the overlap between the top two teams:

I guess that’s what everyone who’s talking about the SEC’s decline must be referring to.  Jeez.


Filed under Stats Geek!

All those spikes… they’re not a good thing, are they?

As I joked with Brian Fremeau on Twitter after seeing this, it turns out that advanced stats can quantify brain farts.  Who knew?


Filed under Georgia Football, Stats Geek!

The (wins) Matrix

I’ve started pondering what to say in my annual SEC predictions post and I find myself growing more receptive to advanced stats in my analysis.  Something else that’s registering with me is Dave Bartoo’s wins matrix, which you can see here.  The gist of his calculations is very simple:  79.8 percent of all teams have been plus or minus two games of their previous season win total.

Yes, there are always going to be outliers.  Sometimes a team like 2013 Georgia gets crushed by injuries.  And sometimes a team like 2010 Auburn emerges unexpectedly because a unique talent like Cam Newton.  (There’s also the Malzahn rabbit’s foot that comes and goes.)  But in any given year, essentially four out of every five teams’ win totals are going to putter along fairly consistently from the season before.

If you want to see how Dave’s math projects for all P5 teams this season, here you go.


Filed under Stats Geek!

As Nick Chubb goes…

The only interest I have in the Heisman Trophy is that it’s an indication of which teams enjoyed a successful season. Nobody on a 5-7 team will be tripping to New York.

So with that in mind, you may find this Dave Bartoo piece on Heisman winner trends of interest.

Group I
This is your top 10 talent ranked teams minus Florida, with a new head coach, and Notre Dame, due to inconsistent play.  With 10 of the last 13 winners coming from top 10 talent ranked teams, this is your best ‘odds’ group. Nick Chubb of Georgia, Kyle Allen of Texas A&M, Ezekiel Elliott and J.T. Barrett of Ohio State, and Cody Kessler of USC are my favorites from this group.

Like I said, if that’s the group that makes it to the next awards ceremony, we’re likely to be a bunch of happy campers about the place Georgia is in by then.


Filed under College Football, Georgia Football, Stats Geek!

I’ll hang up and listen to your answer.

Because we haven’t had enough to argue about lately…


Filed under Georgia Football, PAWWWLLL!!!, Stats Geek!

Another reason for the quarterback muddle?

There’s an interview over at Football Study Hall with the Arkansas high school coach who’s going to try bringing some rugby-style tactics to his offensive gameplan that’s a good read.  But there’s something in particular in it I wanted to focus on.  It’s about his motivation for this innovation in his tactics.

The article details how Kelley analyzed a database of college football stats and discovered a strong relationship between explosive plays (specifically 20+ yards) and winning. And further, he found that the more players that touched the football on any given play (3+), the greater the chance of a 20+ yard play. So, while watching rugby on TV one night, he made the connection: why not add laterals as an additional wrinkle to his already innovative offense? Would that increase the likelihood of more explosive plays?

… The idea for Kelley’s newest offensive innovation began with a conversation with Brad Edwards at last year’s Sloan Sports Analytics Conference. Using ESPN’s Stats and Information database for a little while at the conference, Kelley found that winning the explosive play ratio was more important than even winning the turnover margin. [Emphasis added.] This echoed earlier findings…

…and Bill’s own Five Factors, here and here.

Last season, Georgia was 32nd nationally in plays of 20+ yards.  But when you look closer, it’s more revealing:  10th in rushing plays of 20+ yards; 78th in passing plays of 20+ yards.  (Before you ask, Mason was tied for 23rd nationally in interceptions, with four.)

Now certainly some of that is a factor of the orientation of Georgia’s offense in 2014.  Georgia ran the ball 555 times, compared to 322 passing attempts.  But there’s a little chicken and egg aspect to all of that, too.  Georgia didn’t throw the ball as much because it played to its strengths and those strengths didn’t include a serious downfield passing game.

Richt’s pooh-poohed analytics before, so maybe this is something that doesn’t matter to him.  But what if behind the scenes, this kind of thinking has factored into the equation?  It would certainly reinforce Richt’s general instinct favoring a downfield passing attack as a major part of his offensive philosophy.  If so, the question becomes how much risk are he and Schottenheimer willing to tolerate in an attempt to juice Georgia’s offense beyond last year’s record-setting pace.

And just to add one more wrinkle to that equation, Bill Connelly has looked at the data and finds there’s almost no correlation between completion rate and yards per completion, and it’s close to the same story with yards per completion and INT rate.  That goes against what I would have expected to see.  The conclusion he draws from that:

Well, among other things, quality matters. That’s the ultimate “duh,” but this suggests that quality and skill matter even more than one would think. You can’t really generalize about a QB or a passing game based on merely his per-completion yardage or completion rate.

“Quality matters”.   Duh, indeed.  But when you’re looking at three guys working with a new offensive coordinator, how quickly can you make that determination?


Filed under Georgia Football, Stats Geek!, Strategery And Mechanics