Mysteries of the Offensive Line

Our friend Ben Alamar had a project to try to judge offensive linemen this year by looking at their performance on passing plays. I must admit to being reticent about some of the methodology; I'm not fully comfortable with the idea of judging linemen without knowing the line calls. However, Ben did a few things to try to make sure his results were accurate both scientifically and "football-wise." (Some of this information is from him, not the article linked.) He looked at only passing plays. He tossed out plays if the camera angle didn't allow him to see the whole line. He didn't blame any blockers for untouched rushers, and he didn't blame them for not holding a block more than 4.5 seconds. He also established averages for comparison purposes and used Cox Regression to make sure he had statistical significance. His results were big on Baltimore and New England, but not so big on last year's FO Offensive Lineman of the Year, Nick Mangold.

By the way, note what Jake Scott says at the end about the creation of individual stats for linemen becoming even more difficult with the rise of zone blocking.

View Full Article


20 comments, Last at 17 Jan 2011, 10:12am

1 Re: Mysteries of the Offensive Line

There is something called the "smell test", and I think we might suspect a study that says that

- PIT has the 7th best pass blocking line in the NFL
- Jonathan Scott is playing 700 yards per season better than Flozell Adams

I suppose they assumed that Scott couldn't have been asigned to block the "untouched rushers" coming around the left side of the line.

2 Re: Mysteries of the Offensive Line

Actually PIT didn't have the 7th best line in the NFL. They have the 7th best line of the 8 teams that are currently still playing.

Though Bryan Bulaga rating out higher in pass protection than any other lineman that is still playing doesn't seem right to me either. Though I can't disagree too much with the rest of the grades for the Packers. If Bulaga had come in at 250 and not 557 or so I'd feel pretty good about how he graded them.

3 Re: Mysteries of the Offensive Line

Agree, seems like a major flaw in the results. I know you can never know the line calls. But look at the Polamalu-Flacco strip-sack in the Pitt @ Balt game closely. Grubbs comes off ready to block left. Oher picks up Harrison rather than picking up Polamalu. Grubbs, seeing Oher picking up Harrison, blocks right instead, doubling Birk's man. Polamalu comes in untouched.

1) Oher doesn't get dinged for this? I can understand the argument that he was assigned to block inside, but with Grubbs clearly having a call to either "block left" or "block left or right" it seems pretty clear to me Oher's in the wrong.

2) It seems faulty not to be dinging the Ravens OL in some capacity for this. It's a major flaw in the line or blocking scheme in some fashion. And to say the Ravens have a great blocking line when they're pretty consistently allowing untouched rushers in at the passer doesn't seem to me to be a good, objective measure of their talents.

4 Re: Mysteries of the Offensive Line

I wouldn't say it's necessarily a flaw in the methodology as it is just a limitation in O-Line stats. There's just stuff we can never know, and I would say it's worth knowing how well an O-Line does when they get a hat on all the rushers... even if you'd also want to know how often people are coming through untouched.

5 Re: Mysteries of the Offensive Line

Vollmer had a negative score? Seriously? I always thought he could develop into a viable LT or starting RT and that he was rather good than bad...

Any Pats fan wants to confirm this?

15 Re: Mysteries of the Offensive Line

I thought Vollmer had been starting all season right? When Mankins was gone they moved Kaczur to guard and then Neal and Kaczur went down. Anyway, I don't think that these scores are useless but all the really high scores went to guards (with the exception of Bulaga) which makes me wonder.

7 Re: Mysteries of the Offensive Line

I'm no expert on line play, but he always seemed to be a better run blocker than pass blocker, which is not taken into account in this study. He has also been playing next to a revolving door at right guard, since Stephen Neal went down with a shoulder injury in mid-November, which would also affect his numbers.

10 Re: Mysteries of the Offensive Line

maybe he had a bad year? I've heard that marshall yanda is maybe the most talented guard on the ravens roster, but he sure made a crappy right tackle this year. Of course, my sources could be wrong, and he could suck. But still, I figure circumstances can play heavily in such a thing.

17 Re: Mysteries of the Offensive Line

Yanda is a fantastic player, but he's at his best when playing Guard; he is indeed absolutely lost at RT, although this season seems worse than his typical "bad" experience with it.

To pile on, I agree with other comments regarding the Ravens being #2; anyone who has watched the pocket collapse around Flacco over and over again throughout the season has to feel that the line play is somewhat lacking on passing downs.

8 Re: Mysteries of the Offensive Line

As a Jets fan, I'm truly suprised at Matt Slauson's rating as the best pass blocker on the Jets OL. From my observation, he was very good at run blocking, but was the weak link in pass protection. Interesting.

9 Re: Mysteries of the Offensive Line

This sort of confirms something I think every non patriot fan has suspected for a while. I think it is probably the biggest explanation for why the Patriots can get away with throwing predominantly short passes over and over and why the other teams that throw short passes over and over routinely suck at it. The truth is, you can't get away with doing this unless you have the best pass blocking offensive line in the NFL

11 Re: Mysteries of the Offensive Line

Without access to the report itself, I fail to see how Cox regression applies for this type of analysis. What is the baseline hazard ratio? For that matter, what is he defining as the hazard event? I don't see a survival model applying here.

And what were his regressors? I would imagine there would be 5 (one for each lineman) + some number of situational variables (down/distance, length of throw, etc.). But given that the linemen all play together, the correlation matrix would be hugely skewed (i.e., too much correlation to establish linear independence). You would need a huge number of samples to tease out significant values for individual linemen. And based on the fact that his cutoff was at least 60 plays, that's not nearly enough.

12 Re: Mysteries of the Offensive Line

its good to see that someone is trying to analyze line play, however a good analysis seems a long way away

13 Re: Mysteries of the Offensive Line

I have to question this since he has Quinn Ojinnaka as the Patriots Right Guard, when there are three players (Stephen Neal, Dan Connolly, and Ryan Wendell) who all have more offensive snaps. In fact, I can almost guarantee that Ojinnaka didn't take 60+ snaps at right guard, which violates the "To be included in the study, a lineman had to be analyzed on at least 60 plays" rule.

14 Re: Mysteries of the Offensive Line

The results say Jonathan Scott is about as good as Dan Koppen and Matt Light at pass blocking. I don't think any fan of either the Pats or the Steelers would agree with that position.

16 Re: Mysteries of the Offensive Line

Matt Slauson is the best on the Jets??? Whenever I saw the Jets play he seemed to be getting beat like a redheaded step child. Does anyone agree?

20 Re: Mysteries of the Offensive Line

I had not watched the Ravens a lot this year until the last couple of weeks, but these metrics confirmed what I observed; Birk was simply outstanding.

I hate Brad Childress.