PFF: Tannehill's Sacks vs. Dolphins' Blocking Sacks

[Sarnics13]

Take a breather you two...

 

[DolFan in NE]

...hence the confusion.

Not sure why you're confused. The sacks that the line are responsible for are over 1.5 times more that the league percentage. The QB sacks are right on the league average of percentage. That says to me that even though Tannehill should improve on throwing the ball away in certain circumstances, he's not abysmal at just taking sacks. The fact that the line is .4% worse than any other O-Line in the league is pretty telling.

 

[roy_miami]

...hence the confusion.

OK, I got ya. The problem is the definition of pressure (and I never really looked it up to be honest), when I see other QBs get "pressured" quite often they have somewhere to go, like that play with Brady ducking under a swipe from Wake and taking a step into a nice clean pocket. I believe most of our "pressures" are total meltdowns with multiple defenders coming from all angles.

 

[hoops]

captain obvious strikes again

 

[FinfanInBuffalo]

...hence the confusion.

For me it was never confusing. I stated in a post earlier in the season that the way PFF aggregates the OL grades is wrongheaded. It makes no sense from the standpoint of evaluating the impact of their play. If one guy blows a block, it doesn't matter what the other players do. The OL should be graded as a unit for plays, just as they do for sacks.

 

[wonderl33t]

The stats are biased against crappy o-lines.

 

[j-off-her-doll]

The aspect that jumped out to me was the athletic quarterbacks are responsible for a higher percentage than typical of sacks attributed to the quarterback, while the classic pocket types have very low number of sacks attributed to them. In that regard I don't see how this is good news for Tannehill, if we consider him a pocket type. Virtually all the classic drop back peers are below 2% including young players like Luck and Dalton. Only Flacco among the pocket types had a lousy number at nearly 4%. The elite guys like Brady and Peyton are below 1%.

With poor numbers for all the athletic guys like Newton, Kaepernick, Wilson and Griffin, plus the Buffalo quarterbacks, I would be interested to sit in the office when this type of stat is evaluated. Seems likely that it is skewed against that type of quarterback, even if unintentionally. For example, if an athletic guy makes a quick decision to bolt the pocket and is sacked, they tend to blame him, while with a classic quarterback who hangs in there and doesn't move the sack is credited to the offensive line.

PFF generally favors players that make the play within the play, and it generally fails to account for special plays. Guys like Big Ben and Aaron Rodgers have been underrated by the site for as long as I can remember. After all, the difference between a QB making a solid throw for a first down and Rodgers making something out of nothing and hitting a perfect pass for a TD only carries a 1-point difference (1 point for the first down throw and 2 max). I understand that it's difficult to quantify special plays, but they're the ones attempting to quantify play. At any rate, mobile QB's tend to get sacked more - based on the eye test - because they try to buy time with their feet when the play breaks down. You're probably right that they attribute more sacks to the QB than they should, but I have less of an issue here than I do with their lack of compensation for the positive plays they make with their feet.

 

[Shouright]

OK, I got ya. The problem is the definition of pressure (and I never really looked it up to be honest), when I see other QBs get "pressured" quite often they have somewhere to go, like that play with Brady ducking under a swipe from Wake and taking a step into a nice clean pocket. I believe most of our "pressures" are total meltdowns with multiple defenders coming from all angles.

For me it was never confusing. I stated in a post earlier in the season that the way PFF aggregates the OL grades is wrongheaded. It makes no sense from the standpoint of evaluating the impact of their play. If one guy blows a block, it doesn't matter what the other players do. The OL should be graded as a unit for plays, just as they do for sacks.

So what one of you is saying to explain this discrepancy between the number of sacks and the objective measures of team blocking is that nearly the entire line melts down, which differentiates it from other lines in which a lesser number of linemen melt down, thereby leaving a smaller pocket to move in or escape from for Tannehill than for other QBs, whereas the other of you is saying that the discrepancy between the number of sacks and the objective measures of team blocking is explained by the notion that very few linemen melt down, but melt down quickly, thereby getting Tannehill sacked very quickly, while allowing the other linemen to achieve good blocking grades.

I'm still confused. :)

 

[Awsi Dooger]

PFF generally favors players that make the play within the play, and it generally fails to account for special plays. Guys like Big Ben and Aaron Rodgers have been underrated by the site for as long as I can remember. After all, the difference between a QB making a solid throw for a first down and Rodgers making something out of nothing and hitting a perfect pass for a TD only carries a 1-point difference (1 point for the first down throw and 2 max). I understand that it's difficult to quantify special plays, but they're the ones attempting to quantify play. At any rate, mobile QB's tend to get sacked more - based on the eye test - because they try to buy time with their feet when the play breaks down. You're probably right that they attribute more sacks to the QB than they should, but I have less of an issue here than I do with their lack of compensation for the positive plays they make with their feet.

I worked in a sports staff office in Las Vegas earlier this century. The owner brought me in to help with betting related stats, since his clientele was increasingly tilted in that direction and he needed more specialization. I suggested categories, ones that I knew the bettors were interested in, but at the time were not well known within the stats community. Believe me, the process is hardly point blank and calm. There are disputes virtually every day, particularly when two guys who are in charge of the same category have different interpretations. Slammed doors and shouts. Owner shuffling between stations, making split second decisions trying to remain on schedule. Some guys didn't mind when you stood behind their shoulder and watched how they operated, while others refused to continue.

The numbers appear and you're led to believe they are rigid, the indisputable version. But if you witness the process from behind closed doors you realize it's plenty of crap shooting. It reminds me of when I was a sportswriter in college and sometimes my columns would be edited due to lack of space. I wrote it one way but it appeared another way. Yet the way it appeared in print became the absolute, as far as outsiders were concerned.

I know the Las Vegas betting scene. Virtually everything related that I read on sports message boards is absolute bunk, the conventional wisdom version with little threat of accuracy. I'm not fully versed on stats offices, especially ones requiring so much subjectivity like PFF, and dealing with players and not more clear-cut team stats, but my part time experience during those few years doesn't allow full confidence in the numbers. PFF needs a competitor, as I've posted previously.

 

[NCFINFAN13]

I've said it before ... it will rain Cornfed boys this off season in Miami.

 

[Shouright]

This just in:

James Walker ‏@JamesWalkerNFL 2m
#Dolphins LT Bryant McKinnie said he suggested to the offensive coaching staff this week that the team change its rhythm and cadence.

James Walker ‏@JamesWalkerNFL 3m
#Dolphins OT Bryant McKinnie said he believe the #Bills were on to Miami's snap count and getting a good jump much of the game

Cute.

 

[Bigjbizzle2]

So Tannehill is right at the league average for qb's, but the oline gives up almost double the league average. Just saying

 

[beanh8er]

I mean every single ****ing play out of the shotgun Jerry taps Pouncey and then he snaps the ball the exact same way every time. You have to be terrible to not jump that. I COULD JUMP THAT without ever gameplanning for it.

 

[ckparrothead]

One way or another, once the stat pool was populated with enough of a sample, you knew what it was going to show...which is OL play that is far worse than the QB's pocket management. I thought that was evident far, far earlier in the discussion, even statistically, within the first month. I was surprised there was such a strong narrative until about midway through the season suggesting it was all, mostly, or even half Tannehill's fault.

Not that Tannehill's pocket management has been stellar. I don't think any of us ever claimed that. I think both statistically and visually it always struck me as very average pocket management. But the ability for the line to sustain blocks has always been statistically and visually observable to be far, far below average.

Cheers Awsi. Still like reading your stuff.

 

[Shouright]

One way or another, once the stat pool was populated with enough of a sample, you knew what it was going to show...which is OL play that is far worse than the QB's pocket management. I thought that was evident far, far earlier in the discussion, even statistically, within the first month. I was surprised there was such a strong narrative until about midway through the season suggesting it was all, mostly, or even half Tannehill's fault.

Not that Tannehill's pocket management has been stellar. I don't think any of us ever claimed that. I think both statistically and visually it always struck me as very average pocket management. But the ability for the line to sustain blocks has always been statistically and visually observable to be far, far below average.

Cheers Awsi. Still like reading your stuff.

Perhaps then you can help us sort out why, despite the percentage of sacks attributed to the Dolphins' blocking, the objective measures associated with blocking are no worse than average?

On the one hand, PFF attributes 41 of the sacks to the blocking and 17 of the sacks to Tannehill, which does statistically what I outlined in the original post, and on the other hand, the same source (PFF) has the Dolphins' offensive line measured as no worse than average in terms of the only three objective measures of play it compiles: 1) pass blocking efficiency, 2) overall pass blocking grade, and 3) the percentage of passing dropbacks in which the QB is pressured.

In terms of PFF's attribution of sacks to the blocking versus Tannehill, we also have this:

This is precisely what's happening. The correlation between QB sack percentage and QB yards per carry on rushes (a hypothetical measure of QB athleticism) is 0.66.

When the variance associated with PFF's pass blocking efficiency stat (which by the way correlates with its pass blocking grades at 0.96) is controlled for (i.e., in a partial correlation), the correlation between QB sack percentage and QB yards per carry on rushes is maintained, at 0.67. When the variance associated with the percentage of dropbacks in which there is QB pressure is controlled for, the correlation is 0.58.

In other words, more athletic QBs are having a greater percentage of sacks attributed to them, regardless of the quality of their pass blocking, and regardless of the frequency of the pressure they experience when dropping back to pass.

The above (bolded) also holds true for total numbers of rushes and for rushing yards for QBs, which are additional measures of QB athleticism.