Writers of Pro Football Prospectus 2008

11 Oct 2013

ESPN: Have Defenses Caught Up to the Read Option?

The read option was a popular trend in 2012 and teams are using it more often (3.7 percent of plays as opposed to 1.4 percent last year), but it has not been as effective with the average gain dropping by 1.8 yards.

It's not a new idea in the NFL. What made it popular last season was that the Redskins, 49ers and Seahawks all made the playoffs with exciting new quarterbacks. None of those teams and quarterbacks are performing as well offensively as we expected to see and the read option has not been as successful for all three.

Now don't get me wrong. We should not credit the read option for their success last year or blame it for this season's shortcomings. One of the major themes in covering it should be that most teams, save for maybe the 2013 Eagles through five games, rarely ever use it in their offense.

Remember Ryan Tannehill's 26-yard run in New Orleans this season? That was a zone-read run and it was a very nice play. What wasn't nice was when Tannehill later ran a simple play-action pass that Tom Brady uses and ESPN's Jon Gruden credited that play to the read option, which is something Miami's ran exactly FOUR times in 2013.

For the read option to really be a significant part of an offense, the quarterback has to keep the ball, which is not happening as often this season. However, it's nearly impossible to keep these quarterbacks healthy when you expose them to hits as ball carriers. Even when executing the fake, the defenders are allowed to target the QB at the mesh point. He loses his usual protection.

Since we know passing offense is the most critical factor to success, and that a team needs a healthy starting quarterback to make the passing game work, then it's hard to see the read option as anything more than a slightly more effective Wildcat-type fad.

Posted by: Scott Kacsmar on 11 Oct 2013

22 comments, Last at 14 Oct 2013, 4:16am by Borkowskowitz


by IB (not verified) :: Fri, 10/11/2013 - 3:55pm

Your evidence either undercuts your claims or isn't sufficient to lead to the conclusion you think it does. We learn that Miami doesn't run read option much, supporting your argument that not many teams do, but we also learn that one of the four plays was a 26 yard TD run. We learn that yards per carry have fallen 1.8, but we get no idea of where that started from or how it compares to other rushing plays.

Most importantly, *on your own argument*, the *only* team in the NFL using the read option to any appreciable degree this season is the Eagles. The Eagles have gone from 25th to 5th in offense according to Football Outsider's own shiny efficiency metric. The other three teams you cite (who aren't running it much anyway, you state) are all in the top half of the NFL in DVOA. Buffalo, who runs it the next most, has also improved (with a rookie QB, no less), albeit less than Philadelphia.

FO as a collective whole has had a bug up its butt about read option since its made its way to the NFL. As with so many other FO foibles (from the universally mocked QBR to the over-engineered Lewin Career Forecast, to the similarly over-engineered and inaccurate SackSEER), you guys just double down on your mistakes. Congrats, guys. Football Outsiders - the Phil Simms of football sabermetrics.

by Perfundle :: Fri, 10/11/2013 - 6:07pm

"We learn that yards per carry have fallen 1.8, but we get no idea of where that started from or how it compares to other rushing plays."

Where what started from? And why does it matter how it compared to other rushing plays? All that matters is how it compares to how they did last year. The article isn't titled "Have Defenses Made the Read Option Less Effective than Other Running Plays."

by IB (not verified) :: Fri, 10/11/2013 - 7:01pm

No, that's not "all that matters." For example, if read option plays gained an average of 20 yards last year (obviously an extreme example), then the fact that they are gaining 1.8 yards less means little. You need some kind of baseline to figure out whether defenses have caught up in any meaningful way.

by Coaldale Joe :: Fri, 10/11/2013 - 7:16pm

Exactly. What about sample sizes ? How many read-option plays were there last year, not the %, but the actual number, plus how many this year, and keep in mind this season is only 1/3 complete, so maybe this year's decline in yards gained just reflects random variation. We don't have any idea, because we don't know the scale of the decline, and we don't know the sample size.

by Coaldale Joe :: Fri, 10/11/2013 - 7:19pm

Nevermind. Just followed the link and saw that the actual ESPN article has the data I thought was missing. That'll teach me for quickly firing off comments.

by Perfundle :: Sat, 10/12/2013 - 6:42am

The baseline that you want is in the article (a drop from 6.4 YPC to 4.6), so maybe you should read it before you make your judgment from the blurb presented here. The current non-read option average is 4.1, so defenses are containing it almost as well as ordinary running plays now.

by IB (not verified) :: Sat, 10/12/2013 - 10:02am

So read option plays are still outperforming regular rushing plays by half a yard despite being run 2.5 times more often, and primarily by offenses that below-average last season and weren't running the read option. This is not a compelling argument for "defenses have caught up to the read option."

by IB (not verified) :: Sat, 10/12/2013 - 10:09am

2.5 more often than last season, that is.

by Karl Cuba :: Fri, 10/11/2013 - 8:43pm

Could you provide some evidence of FO having a 'bug up its butt' about the read option? I read pretty much all of their stuff and I haven't noticed it.

by ClemsonMatt (not verified) :: Fri, 10/11/2013 - 4:47pm

I thought QBR was an ESPN thing? It includes some sort of clutch aspect that has been mocked here as well.

The metrics here are continually refined and questioned to get better. That's the exact opposite of "doubling down". I appreciate that instead of giving up and tossing concepts they're improved.

by IB (not verified) :: Fri, 10/11/2013 - 5:05pm

(1) Go to quarterback stats and QBR is highlighted in green. Whoever is responsible for the origins, FO has trumpeted it as useful and includes it among their own measures.

(2) The metrics are getting better at describing the past, because they are backfit across more data and use more variables. That's the easy part, though. Anyone can run a regression analysis. The metrics are supposed to be predictive tools, and they are far too over-engineered to do that effectively.

by Perfundle :: Fri, 10/11/2013 - 6:11pm

"Whoever is responsible for the origins, FO has trumpeted it as useful and includes it among their own measures."

Uh, no, they haven't trumpeted it as useful. They list it as a comparison because anyone who watches/uses ESPN has surely heard of it and will be asking what the difference is. On that same page the differences are laid out; QBR is highlighted precisely because is isn't their stat.

"The metrics are supposed to be predictive tools, and they are far too over-engineered to do that effectively."

All predictive tools have to be backfit. What do you have in mind that's more effective and not over-engineered?

by IB (not verified) :: Fri, 10/11/2013 - 7:09pm

"Uh, no, they haven't trumpeted it as useful. They list it as a comparison because anyone who watches/uses ESPN has surely heard of it and will be asking what the difference is."

Even more people have heard of regular QB ratings, and those aren't included in the charts. In any event, this article sums up Aaron Schatz's stance towards QBR:


Key paragraph:

"Those are, obviously and inevitably, a lot of "other things." But that's quarterbacking, and taking the position's many facets all into account helps make QBR such a useful metric. "The goal of QBR is to come up with the best possible measure of how well a quarterback played," Schatz says, and he emphasized it was pretty good at this. (It's worth mentioning that Schatz, aside from his longstanding role at Football Outsiders, is a regular on ESPN programming. He noted several times to me that he understands how his praise might be perceived, but stressed he truly does value QBR as a good metric.)"

If Aaron wants to dispute any of those characterizations, then fine. If not, I'll take that as representative of FO's stance towards QBR.

As for a good predictive tool that isn't over-engineered, I've always thought that speed score was a good example. Two variables, strong correlation with success over time, not overly dependent upon outside observation variables like where a player is drafted.

by Karl Cuba :: Fri, 10/11/2013 - 8:39pm

ESPN derived QBR, FO just list it for comparison purposes. That you insist on pinning it to FO just makes it look like you have some bizzare axe to grind, probably because of you have.

by IB (not verified) :: Fri, 10/11/2013 - 9:35pm

I do have an axe to grind; my axe is that I think it's embarrassing FO can consider themselves an advanced-statistics website, then cite QBR. They're doing it either because ESPN cuts them a check, or because they believe QBR is a useful metric. I'll let you decide which is more damning.

by Perfundle :: Sat, 10/12/2013 - 6:28am

Or because they want to compare it with their own stat?

"Even more people have heard of regular QB ratings, and those aren't included in the charts"

Because QB DVOA and QBR are both proprietary stats, and work similarly, taking account of factors that regular passer rating doesn't. How do they help the cause of QBR when they have a completing stat that they consider as better? I know that I definitely don't think highly of QBR in large part due to what I perceive to be the advantages of DVOA.

by Karl Cuba :: Sat, 10/12/2013 - 12:08pm

Or you're a crank and the only person who gets bothered by it, I'll let you decide which is more damning,

by mehllageman56 (not verified) :: Fri, 10/11/2013 - 4:55pm

His argument that the read option is not being used as much because teams would like their new talented quarterbacks to stay healthy is not belied by the experiences of Buffalo or Philadelphia this year, since both of their starting quarterbacks are injured. Of course, Vick always gets injured, and Manuel is hurt because he didn't run out of bounds soon enough, but teams do not want their shiny new quarterbacks to take too many hits.

Since the NFL decided that hitting the quarterback on read option plays is legal even when the qb doesn't have the ball, it follows that teams will cut down on those plays. The Seahawks seem to have cut down on them, instead choosing to have Wilson scramble after trying to pass first.

by IB (not verified) :: Fri, 10/11/2013 - 5:14pm

That can't be his argument, because teams are using the read option 250% more than last season on his own numbers.

by mehllageman56 (not verified) :: Fri, 10/11/2013 - 7:13pm

Ok, you caught me, I only read the paragraphs above because I don't have Insider. I still don't think FO is the Phil Simms of sabremetrics.

by Topas :: Sat, 10/12/2013 - 6:19pm

Stats are useless without Raiderjoe commenting on them!

by Borkowskowitz (not verified) :: Mon, 10/14/2013 - 4:16am

If you're talking about the standard read option, then it wouldn't surprise me that teams are starting to catch on and lock up the teams that are running it like a gimmick play.

As with any system, the teams that treat it like a gimmick will fall behind, but the teams that adapt and treat it as a concept to expand upon will do well. If you look at the Eagles, they lost their mobile QB for a few weeks, but Nick Foles still made reads on packaged plays and operated a screen game that compensated for the diminished run threat. The reads are still there, the bluffing is still there, even if it isn't fishduck.com 101.

If you treat it like a gimmick, it will be a gimmick.