Articles

Blogging about college football by an Oklahoma Sooners fan.

Lincoln Riley and the epistemology of good defense

Since this deal about Lincoln Riley and Georgia’s defense remains a thing, I figured I’d add a few more logs to the fire. The full exchange between Riley and Danny Kanell from a spot on ESPNU Radio on SiriusXM is below.

LR.jpg

Riley is correct: Georgia would not have finished in the top five nationally in total defense in the regular season if it had played the Big 12 in 2017.

This is just a matter of arithmetic. The Big 12 teams generally played faster than the teams that populated Georgia’s schedule last year, which results in more offensive plays per game.

As an illustration, six teams in the Big 12 finished in the top 50 in the country in plays per game last season. There were three teams from the SEC in the top 50, and UGA only played two of them. Conversely, there were five teams from the SEC in the bottom 20 nationally – Georgia played four of them. The bottom 20 included one Big 12 team, Kansas State.

If you want to get a bit more esoteric, five Big 12 teams ranked in the top 50 of Bill Connelly’s Adjusted Pace metric last season, versus four from the SEC. Georgia played three of the four. On the other hand, the SEC placed six teams in the bottom 20 of Adjusted Pace, and the Bulldogs played four of them. Kansas State was the lone Big 12 team in that group.

Say Georgia’s defense had to face just five more plays per game on average in the regular season a year ago. Using its average yards allowed of 4.5, the Bulldogs would have given up about 294 yards per game during the regular season, instead of 271.

Voila. “More plays yield more yards” makes for an easy rule of thumb.

I don’t have the energy to go back and calculate what all the other teams’ numbers would have looked like prior to the postseason, but it’s a near certainty that would put UGA outside the top five in total defense.

The real takeaway here – and the point that Riley was making – is that statistics like total offense and total defense really aren’t that informative. Therefore, teams such as TCU and Iowa State can field good defenses that appear less so when looking at raw numbers.

If you want to take what Riley said as a shot at anyone, look to the people who espouse the idea that “Big 12 offenses only put up gaudy statistics because they’re facing pathetic defenses every week.”

Should the way you determine good defense be different in the Big 12 because of the offenses you’re facing?

That was the question Riley was answering. The question is moot if we’re really just talking about offenses that look better than they are because they’re beating on tomato can Ds.

However, as Riley noted, OU’s offense produced against two of the best defenses in the country last year in UGA and Ohio State. He also pointed out that some of the other Big 12 teams put up yards and points in games against teams from outside the conference.

Check out the numbers from:

That might be as close to proof of concept as you can get in a college football universe defined by inherently limited sample sizes and low connectivity.

If you’re trying to argue that Big 12 teams’ outsized offensive stats are the product of weak defenses – or the corollary that the big numbers put up by offenses of Big 12 teams prove that the defenses in the conference all suck – you have to reconcile that with how the teams perform when they’re not playing each other.

Since raw stats are bunk, what do more rigorous measurement methods have to say about this?

Riley’s comments offer an implicit endorsement of some of the Moneyball-type evaluation systems that have popped up in recent years.

I looked at the only published offensive efficiency metrics that I’m aware of – S&P+, FEI, ESPN, CFB Analytics – as well as raw yards per play to get a better picture of the lay of the land in the power conferences. These systems all aim to level out the impact of tempo and competition to give a more standardized measure of offensive efficiency.

How many offenses from each league ranked in the top 50 in offensive efficiency?

P5Os.jpg

Not surprisingly, you can spin those numbers in different ways. For example, Brian Ching of forbes.com used yards per play to argue that what Riley said about Georgia “was based upon the false premise that Big 12 offenses are decidedly more explosive than their SEC counterparts.” (Riley didn’t use the term explosive anywhere, and yards per play is a fairly limited metric... but whatever.)

Ching points to the fact that both the SEC and Big 12 had six offenses ranked in the nation’s top 50 in yards per play. He also noted that both leagues had two offenses that ranked 100th or worse. “Pretty even in my book,” he concludes.

The other systems suggest the SEC offenses are even stronger than that. ESPN, for instance, says 10 of the 50 most efficient offenses in the country reside in the SEC.

However, when it comes to the weekly consistency that Riley referenced, unbalanced scheduling in other leagues can have a significantly dilutive effect. The Big 12’s round-robin conference slate means that its members will necessarily play between four and seven games against top 50 teams, depending on the metric you pick. It’s more of a crap shoot in the other conferences.

Take a gander at the SEC numbers, for instance, using Ching’s preferred measure of yards per play. Some of the SEC teams, including Georgia, only had to play two teams in the regular season that ranked in the top 50. Others had to play five, but a majority played no more than three.

Meanwhile, OU had to play five in the Big 12.

None of what Riley said negates the reality that a large portion of Big 12 defenses are bad. Case in point, the Sooners.

In a sardonic sense, I have to admit the funniest part of this entire kerfuffle to me is the number of times I read or heard people making detours into extrapolating what the Bulldogs’ offensive mauling of OU meant. UGA lit up the Sooners’ defense… just like six or seven teams did in conference play.

OU’s defense was objectively poor last season. Riley himself has coachspeaked some semblance of that sentiment on multiple occasions since the season ended. (He even did it in response to a question during the same interview in which he made his comments about Georgia’s defense.)

In fact, it’s strange to me that Riley’s comments are being read in some corners as an argument that OU’s defense was better than it looked. There is not one word about that in his response. Notably, he brought up other Big 12 teams’ defenses – TCU and Iowa St. – as examples of quality units, but there was no mention of his own.

There are probably more bad defenses in the Big 12 than there are good ones. That shouldn’t come as a shock to anyone. Again, Riley was talking about how to identify the good ones.

-Allen Kenney