Writers of Pro Football Prospectus 2008

Most Recent FO Features

BrownMal15.jpg

» Futures: Texas RB Malcolm Brown

DeMarco Murray is the toast of the NFL, but injury and team issues clouded some observers' view of his talent. Texas RB Malcolm Brown might have the same problem this winter. 

20 Jun 2014

Classic Pocket Passers Versus Mobile Quarterbacks

guest column by Gregor Bozic

For many years, the popular belief among NFL analysts was that the success of an NFL team comes with a quarterback who can stand tall in the pocket and deliver the ball downfield. Members of the elite group of active quarterbacks, Peyton Manning, Tom Brady and Drew Brees, for instance, also earned their reputation by making plays almost exclusively from the pocket.

But there is a growing group of young signal-callers who can make plays with their legs, scrambling to escape the pressure or using their legs on an occasional designed run play. Russell Wilson, Colin Kaepernick and Cam Newton are three of the players considered dual-threat quarterbacks who entered the league in recent years and have already had their share of success. Consequently, we now hear from the analysts that the ability to extend plays and make something out of nothing is the most vital component of the art of the NFL quarterbacking.

In an attempt to graphically present the well-known difference between classic pocket passers and athletic quarterbacks capable of improvising, I've designed visualizations based on collected spatial data for three active NFL quarterbacks. In this article, the group of pocket passers will be represented by Peyton Manning. Wilson and Kaepernick will represent the group of mobile quarterbacks.

I divided the backfield into a grid of so-called "throwing cells." I reviewed the games and charted the locations from where the quarterbacks threw passes or took a sack during 2013 playoffs and mapped them out in the pre-defined grid. In addition to "coordinates," I've collected "attribute" data to determine pocket/out-of-pocket throws and sacks, behind the line of scrimmage passes, designed rollouts, and passes thrown under pressure. All of which helped me further analyze spatial-based data for Manning, Wilson and Kaepernick.

The first visualizations are known as heat charts. The charts will answer two questions; from where and how often did the trio of quarterbacks attempt a pass in a certain area of the backfield? I grouped pass attempts into classes based on number of attempts which fall into regular hexagons with a side length of one yard.

Russell Wilson attempted just 68 passes in Seattle's three playoff games. Wilson completed 43 of them for 524 yards with three touchdowns. Of the 68 attempts, Wilson threw 13 passes behind the line of scrimmage, which is 19.1 percent and presents the highest rate of all three players considered in this article.

We can see from the chart above that Wilson most frequently (seven times) attempted a pass 7-to-8 yards deep in the backfield, just inside of the right hashmark. The spots where Wilson threw a pass in the playoffs are scattered all around the backfield. Still, the vast majority of passes were attempted from the right side of the field. On average, Wilson tossed a pass when located seven-and-a-half yards behind the line of scrimmage.

Next is the heat chart of Kaepernick's pass attempts. Kaepernick attempted 82 passes in three games during the 49ers' 2013 playoff run. He completed only 45 passes for 576 yards with three touchdowns and three interceptions. One of the differences compared to Wilson is the number of passes which didn't travel past the line of scrimmage and are thus not shown on the chart. Kaepernick attempted just three passes, or 3.7 percent of the total.

Just like Wilson, Kaepernick most frequently (nine times) attempted a pass 7-to-8 yards deep in the backfield. Here too we can see passes attempted from all over the backfield, but Kaepernick was more willing to pull the trigger while rolling to his left. On average Kaepernick attempted a pass eight yards deep in the backfield.

A completely different picture shows up in Manning's heat chart. Manning threw 128 passes during three playoff games, which is almost as much as Wilson and Kaepernick combined (150). The ultimate pocket passer completed 71.1 percent of his passes and gained 910 passing yards with five touchdowns and three interceptions. Eighteen of his passes, or 14.1 percent of the total, didn't make it past the line of scrimmage. Unlike the two mobile quarterbacks, Manning was able to make subtle movements and climb up the pocket to avoid the pressure. He often released the ball quickly from the spot where he waited for the ball to be snapped, without further dropping back behind the line of scrimmage.

As a result, spots from where Manning threw passes are much more condensed and closer to the line of scrimmage, with an average "throwing" line at six-and-a-half yards behind the line of scrimmage. The quarterback of the Denver Broncos most frequently (20 times) threw a pass just inside the right hashmark, 5-to-7 yards deep in the backfield.

The heat charts reveal the main difference in throwing tendencies, in terms of throwing locations, of Manning, Wilson and Kaepernick. The spot where the center snaps the ball is dependent on the result of the previous play. So besides the obvious out-of-pocket attempts, the ones closer to the sidelines, we can't see how many of the passes were actually thrown outside the boundaries of the pocket. To answer that I've designed another chart with only pass attempts thrown outside the tackle box. All three quarterbacks are in one chart here.

There's no doubt Manning prefers playing from the pocket and it showed during 2013 playoffs in which he threw only five out-of-pocket passes in 128 attempts. Three of those five were on designed rollouts, including a one-yard touchdown pass thrown against the New England Patriots in the divisional round. Overall, Manning went 4-of-5 for just 21 yards and one touchdown on those attempts. Wilson and Kaepernick combined for 37 out-of-pocket passes, almost one-quarter of their combined 150 attempts. Kaepernick was extremely successful on the move. He completed 12-of-20 attempts for 141 yards and two touchdowns. Only seven of those 20 attempts came on a designed rollout, all seven to the right side. Wilson did not throw a touchdown pass when working outside the pocket, but he did gain 165 yards with a 58.8 completion percentage. 6-of-17 attempts were on designed rollout plays, with only one to the left side.

The charts shown so far only include pass attempts and ignore scrambles and sacks. But we know those are pass plays by design. The ability to escape the pressure and avoid the sack is a skill attributed by many to the group of mobile quarterbacks. But the fact is that good pocket passers take far less negative plays and have lower sack rates. Manning never scrambled and took only one sack during the 2013 postseason. His pocket presence and ability to process information pre- and post-snap allowed him to release the ball quickly and avoid those kind of negative plays. Manning's total pass plays count is at 129, if we take into consideration pass attempts, scrambles and sacks, which gives him an 0.8 percent sack rate.

Wilson and Kaepernick, considered as two of the best mobile quarterbacks, absorbed far more negative plays. Kaepernick is generally quick to leave the pocket and sometimes it’s hard to identify whether he actually scrambled or ran a designed draw. Based on my game charting he scrambled seven times, took six sacks, and finished the postseason with a sack rate of 6.3 percent. Of the three quarterbacks, Wilson was the one that was pressured the most and he too tends to leave the pocket early when he feels the pressure and turns himself into runner. He finished with the highest sack rate, with 68 pass attempts, six scrambles and seven sacks. Wilson recorded a 8.6 percent sack rate. In the next image we will see the locations of the sacks the quarterbacks took during 2013 playoffs, further broken down with relation to the pocket.

The lone sack Manning took was a strip sack late in the Super Bowl which happened within the pocket. Four of Wilson's seven sacks occurred close to the line of scrimmage. On one occasion he scrambled and ran out-of-bounds just shy of crossing the line of scrimmage, and three times he scrambled and dived for a minimum loss of yards. Three of six total Kaepernick sacks took place when he was working outside the pocket. Similar to Wilson he took one sack when he scrambled to the left and ran out-of-bounds just short of reaching the line of scrimmage.

The visualizations presented in this article are based on small-sample sizes of spatial-based data. Despite that we are still able to see that each of the players display certain tendencies when playing the most important position in the game of american football. The differences between the two types of quarterbacks can clearly be seen on the charts. Peyton Manning, who has great pocket presence, rarely leaves the pocket. Even under heavy pressure when facing the No. 1 defense in last season's Super Bowl, Manning took only one sack and left the pocket only twice on 49 pass attempts, both times on designed rollouts.

Wilson and Kaepernick leave the pocket while under pressure, trying to make big plays on the move. With those big play chances also come an increased rate of negative plays. As right-handed quarterbacks they are inclined to roll out to the right, as was the case during 2013 playoffs. But as the postseason sack chart revealed, those outside pocket sacks all happened on the left side of the field. When rolling to his left it is much more difficult for a right-handed quarterback to be accurate with his throws and get rid of the ball on time to avoid negative plays.

It is important to remember that a quarterback's performance isn't only a function of his own skills. The trio of quarterbacks operated in different kind of offenses behind different offensive lines. Helped a great deal by their defenses and power running game on offense, both dynamic quarterbacks weren't asked to do a lot in the passing game. Kaepernick, for example, attempted only five passes in the first half of the 2013 NFC Championship Game. The 49ers game planned to use him more as a running threat against Seattle's top-rated pass defense. Wilson, who played behind a porous offensive line, handed the ball off to Marshawn Lynch 50 times and only attempted 43 passes in the Divisional and Championship Game combined. That's less than Manning's attempts total in Super Bowl XLVIII alone.

With a larger sample size of the data we could perform spatial analysis and look for quarterbacks' tendencies when playing against certain defenses. Kaepernick's heat chart when facing teams from the NFC West would certainly look different than a chart which would include data for the rest of the 49ers' 2013 regular season opponents. With visualizations designed based on data splits for opponents or field position, we would be able to further explain and evaluate space-dependent performance of the quarterbacks.

Gregor Bozic is a geodetic engineer living and working in Slovenia, a beautiful country on the sunny side of the Alps, and a fan of the San Francisco 49ers since 1988.

Posted by: Guest on 20 Jun 2014

56 comments, Last at 27 Jun 2014, 10:21pm by theslothook

Comments

1
by PaddyPat :: Fri, 06/20/2014 - 1:12pm

Perhaps a quibble, but it seems to me that Drew Brees is a poor example of a pocket passer. His reputation (whether accurate or not) is as a nimble-footed guy whose pockets role, and who uses gutsiness and mobility to compensate for the fact that he is too short to be a classic pocket passer. Pocket passer? I think Drew Bledsoe.

9
by Kopalec :: Sat, 06/21/2014 - 1:32am

I also wonder at the examples used based on their respective experience. If the purpose is to show the differences in a pocket passer vs. a mobile one, then the gap in experience may skew the result. Until such time as Wilson and Kaepernik improve their own ability to read defenses more quickly while maintaining their mobility.

Further you seem to only refer to "negative plays" in these examples in terms of sacks. While Manning was only sacked once in the SB, anyone who watched the game knows he was under constant duress and made poor throws leading to interceptions, or his receivers being blown up.

10
by G-49er :: Sat, 06/21/2014 - 6:11am

The way I wanted to present the data was to look only at what went on behind the line of scrimmage. Manning was heavily pressured during SB, but he rarely left the pocket and as you noted only took one sack. I used the phrase "negative play" more as a synonym for a sack.
And I have to agree with what you've said in the beginning, we'll have to wait and see how Wilson & Kaepernick develop with regard to pocket passing.

14
by Kopalec :: Sun, 06/22/2014 - 9:10am

I still enjoyed the article, these type of write ups help get me through the off season withdrawl.

If you venture to do this again, perhaps a comparison of QB's like Jay Cutler vs Aaron Rogers, or if you want to compare different era's, Steve Young and Randall Cunningham vs Dan Marino and Jim Kelly? (Fran Tarkentan vs. Johnny Unitas?)

16
by G-49er :: Sun, 06/22/2014 - 2:10pm

Would love to do those kind of different-era comparisons. As a 49ers fan Montana-Young-Kaepernick comparison first comes to mind. For that kind of analysis All-22 and Endzone views really help. The problem is season archives are available only from 2009 season.

2
by PaddyPat :: Fri, 06/20/2014 - 1:23pm

Nice piece. It would be interesting to see larger sample sizes with these sorts of charts. Once thing that stands out to me as a New England fan is that pocket presence and pocket mobility are important for a quarterback irrespective of whether he scrambles or steps up. Tom Brady and Peyton Manning have the ability to extend plays in a different way from Ben Roethlisberger or Russell Wilson. Being able to distinguish between the mobility of Byron Leftwich and Peyton Manning is perhaps more significant than trying to evaluate whether Manning or Wilson represents an "optimally" athletic quarterback.

3
by theslothook :: Fri, 06/20/2014 - 2:06pm

One of the big implications about pocket vs mobile was the view that playing from the pocket is more efficient and overall much preferred. I use to be more dogmatic about this being the case, but others correctly pointed out that there have been a few historical counter examples of qbs who were effective doing both.

I guess we'll have to wait and see, but the chart above does once again confirm that both of these qbs are on teams that do not require their quarterbacks to consistently be excellent to win games. In time, we should expect a level of regression from the rest of the rosters so we'll see how they adapt.

5
by Perfundle :: Fri, 06/20/2014 - 3:22pm

Which counterexamples? Did they have a choice of staying in the pocket? You talk about the roster regressing, but that's hardly likely with respect to Seattle's offensive line, and most particularly the guards, who are arguably the worst in the league. Interior pressure is the absolute worst thing for a pocket passer, and I'm not particularly optimistic about that improving a lot this year, as the FO is sticking to their belief that players will improve. It's sort of fascinating watching them build the team back-to-front. The widely-held belief is that you build a team starting from the trenches, but Seattle seems to be putting everything into the receiving corps and the secondary.

7
by theslothook :: Fri, 06/20/2014 - 4:57pm

Historical examples include Steve young and Aaron Rodgers.

As to Seattle, im referring to the future losses of import depth players like mebane, Wright, Wagner, avril,.Maxwell, etc. obv some will be retained but many will be let go and the elite players will have to carry more of the load. A potential injury to one or more will further burden the rest of the roster. In short, I can't see this style of the Seahawks persisting long term. The regression will need to be offset by Wilson's progression.

Finally, I use to believe Seattle's D was being carried by its elite players, but then pff told a different story. Nearly everyone in their defense was graded well, including many of their role players. I suppose you could argue they.are all a product of Sherman, Thomas, or Carrol's scheme, but I'm inclined to believe its just one of those times when all the picks pan out.

11
by Perfundle :: Sat, 06/21/2014 - 6:29pm

Rodgers' OL was below average for most of his time at Green Bay, wasn't it? Also, and I don't know how much of a factor this is, Rodgers and Young are both short as QBs go. Young was before my time, but Rodgers seems to get a lot of passes tipped. Maybe there's a relationship between shorter QB height and mobility?

As for PFF's grades, have past historically great defenses (such as Chicago last year) had players that were graded below-average? If not, then the grades are not very useful.

By the way, who was the worst-graded defender and what was he poor at? I'd like to see how well their grades match up to my perception of the players.

12
by theslothook :: Sat, 06/21/2014 - 10:36pm

It's pretty hard to tell grade wise how to compare the different defenses, but I attempted anyways using Pff, comparing SEA, 2011 Bears, and 09 Jets who actually had a better pass d than both of these other two defenses.

SEA - They had 30 players who took a snap for the defense with 17 logging an above average number of snaps. Out of that 17, 16 had positive grades.

CHI - They had 26 players log a snap, with 13 logging an above average number and 8 having positive grades.

NY - They had 28 players who took snaps for the defense, with 12 logging above average snaps and all 12 grading positively.

The grades are sort of all over the place. Both NY and SEA shared high grading corners and safeties(rhodes and thomas and chancellor), but sea had higher grades across the board for both their d line in terms of pass rush and their linebackers in terms of coverage. The jets seem to have really benefited from an absurd year from both revis and rhodes(both outgraded sherman and thomas).

The bears are the most hit and miss.

Again, not really sure what to make of all of it.

15
by Karl Cuba :: Sun, 06/22/2014 - 9:45am

"Again, not really sure what to make of all of it."

Perhaps that PFF couldn't tell their arse from their elbow?

19
by Perfundle :: Sun, 06/22/2014 - 5:24pm

I have a pretty big problem with assigning individual grades for a unit as interconnected and reactive as the defense. Do the secondary players get partial credit for coverage sacks? Do linebackers get credit for blocking a running back's path and making him hesitate enough so that a defensive lineman can make the actual, unassisted tackle? What if that linebacker is forced into coverage because a CB blew his assignment and the LB has to cover for him? What if the CB only blew his coverage because he got picked illegally? I just don't see how blame or credit can be assigned given all that minutiae.

20
by theslothook :: Sun, 06/22/2014 - 5:38pm

To assess that, you would need to go through their methodology. You are also trying to gauge their quality control. I know exactly what Karl thinks about it, so I won't bother trying to change his mind.

I would offer these two defenses. First, they do have an admirable mission and they vet their methodology by asking teams and scouts what they use. Their founder and people like Sam Monson are invited to training camps as consultants and much of their data does get requested by teams. Their data also gets used by both agents and general managers for contract purposes. Were all the data they used utterly meaningless, I doubt these things would happen.

To that point, I don't use them as a be all and end all of judging teams. It's a stat. At the moment, it's the only stat that actually attempts a micro view of football. And though it is likely flawed, it's better than being completely ignorant.

Going back to Seattle. I've heard this narrative that earl thomas' presence is what allows the defense to do x y and z. I use to believe that, but then I learned that Seattle not only led the league in pressure rate(it wasn't all coming from coverage sacks), but they also had green coverage grades for everyone involved. And considering that neither the bears last year nor the 09 jets had such similar grades for all their role players, it leads me to believe that the thomas narrative is probably not true. I also think sherman and chancellor are both excellent players and would be great even if they were dropped onto the raiders roster.

21
by The Ancient Mariner :: Mon, 06/23/2014 - 12:09pm

You should still believe it, at least to some degree. Having ET covering centerfield allows Sherman to play fearlessly, because he knows *if* he makes a mistake, Earl will be able to clean it up (most of the time); and because Sherman's game is built on length, strength, brains, and balls rather than on being quick and shifty (which he isn't), he needs that assurance in order to be effective. Having ET + Sherman has also had a lot to do with the success of the CBs on the other side of the field, because the 'Hawks are able to roll ET away from Sherman a lot of the time, which makes it easier for the rest of our corners to succeed. ET's presence on the back end also gives our LBs the freedom to be aggressive, and lets the DC do things like dropping our DTs into coverage once in a while.

Yes, the revamping of the DL last year also had a lot to do with our defensive success, making the LOB's job easier in several ways and reducing the stress on ET in particular. The scheme is still built from the free safety forward, though, and ET is still the most irreplaceable player on the team, aside from (probably) Russell Wilson.

22
by Perfundle :: Mon, 06/23/2014 - 2:40pm

Arguing against your first paragraph is the play of the Seahawks in 2012. Thomas and Sherman were just as good back then, but the coverage of the linebackers was horrible at times, especially on 3rd-and-long.

26
by theslothook :: Mon, 06/23/2014 - 4:31pm

I did say it's possible the whole scheme is dependent on Thomas, but I still am skeptical. Again, going with pff numbers, it's actually been chancellor who has been the better coverage grade safety the last few years, with this year being the exception when thomas was dominant.

I would also add that Sherman is paid more than Thomas and that corners in general are paid more than safeties. That could be because great corners are harder to find, but it also could mean that corners are simply more valuable than safeties.

31
by Perfundle :: Tue, 06/24/2014 - 2:17am

I'm not quite sure what to make of your reply to me, because I'm basically agreeing with your point that Thomas isn't as important to the scheme as people say he is. Clearly, his presence (or even the presence of Sherman and him together) doesn't automatically guarantee improved coverage ability on the part of the linebackers and slot corner.

As far as coverage grade goes, I believe PFF's grades are cumulative, so if offenses are avoiding Thomas, then he's not going to have a chance to accumulate a high grade.

32
by theslothook :: Tue, 06/24/2014 - 3:43am

All that is fair and I agree with what you said.

34
by Karl Cuba :: Tue, 06/24/2014 - 11:28am

Thomas's presence for not guarantee improved coverage ability but it does mean that their assignments are easier.

Which is precisely the type of contextual information that PFF ignore.

To put this another way, Hooky seems to be arguing that every player in Seattle's D must be really, really good because they have good grades. I think that they are getting good grades because they are playing in a well coached system that asks them to play to their abilities in rules made easier by Thomas's ability to take away the deep middle on his own on most plays.

Put a league average safety back there and I think he'd get burnt quite a bit, which would force a change in scheme. Chancellor would be playing deep coverage, where he'd be covering receivers more often and he might give up some plays at his size. They couldn't play that 6-2 front so the run defense would suffer which would result in the linemen having to be less aggressive in shooting gaps and the pass rush would need more time to get pressure.

They'd still be a good defense as they do have good, disciplined players with sound technique but it would be qualitatively different to the monster it is right now.

However, there is no control group for this hypothesis. Unless Thomas misses half the year, which would give teams three or four games to study the Thomas-less defense and then another three or four games for teams to try to attack them, we can't know for sure.

36
by theslothook :: Tue, 06/24/2014 - 8:37pm

We seem to be getting stuck on two separate issues, so I'll address them one at a time.

First, this thread started because I made the claim that eventually, both the 49ers and Seahawks would need to depend on their qbs. This is because talent naturally regresses. The role players above, whatever you think about them, will eventually walk and it's not obvious to me that you can just plug in a replacement 3rd rounder and expect the same results. Seattle isn't going to field a top 10 all time defense every year, especially when defense is so much more volatile year to year.

As to Thomas. I think the issue I have comes with the narrative, "Thomas sets the defensive scheme and everyone else plays off that." Are we so sure it is this way? It's not like Thomas was the first great cover safety. Ed Reed has had a long career of being a rangy speed cover guy, but that hasn't stopped the ravens from fluctuating on defense. Furthermore, since defense to me is much more about depth and overall quality, I'm skeptical one super talented player can make everyone else play so well that the whole defense is top 10 all time.

When I look at the seahawks, I think they have three elite players on defense, but I also marvel at just the sheer number of quality plays they got from a whole host of players. Wagner and Wright played well. Avril and Bennett along with a deep rotation were the best pass rush in football. And of course, Sherman was superb, as was Chancellor. That's what makes a top 10 all time defense.

37
by Karl Cuba :: Tue, 06/24/2014 - 9:41pm

Point one: totally agree.

Point two: I think it's interesting that you raised Ed Reed and I think he's a great comparison for Thomas. The Ravens were able to field high level defense for years because they had Reed operating in a very similar scheme to the one Seattle use today. They used to leverage everything towards Reed in the same way, if you have a HOF safety it can change everything (in my opinion).

Query, who I'd the third elite Seahawks on defense in your opinion? Thomas and Sherman are obvious but I don't see the clear cut candidate out of Chancellor, Bennett and Wagner, their all good but none are elite in my eyes.

38
by theslothook :: Tue, 06/24/2014 - 9:51pm

I think Chancellor plays a vital role in that defense too. I think linebacker coverage(or middle zone coverage) can be very different than deep range freelancing coverage, just like slot coverage is different wide coverage. Although I admit chancellor is the most controversial of my choices.

Part of my appeal for chancellor comes from his versatility. They've asked him to guard tight ends in man, cover shallow zones where the running backs and slot receivers are, and he's even a useful help defender too. I do admit to being slightly enamored with him ever since I saw the redskins hawks playoff game.

Looking over the Ravens defense, I am actually a bit amazed at their consistency so maybe having the all pro safety does make sense. Still, they've also had great players on their defense too so it's hard to disentangle.

As an aside, I've had this debate with my friend so I'm curious to see which side you come down on. Are Ahmad Brooks and Ray McDonald average players, slightly above average, or solidly good? How might you compare them to say Wright and Wagner?

39
by Karl Cuba :: Tue, 06/24/2014 - 11:31pm

I think Chancellor is outstanding at what he does, fantastic but I do wonder if he'd be as good if he had to play half the field in a cover two. I'd have said Bennett was the third elite guy.

Ray McDonald was really good a couple of years ago but hasn't been the same guy and after he played most of last year with a torn muscle (pectoral/tricep? I can't remember) he struggled. I'd be surprised if he was on the roster next year. He looks to be very average right now, I think the niners hope that Carradine will grab that job.

Brooks is a very good player. How many 3-4 linebackers are outstanding at setting the edge, good in their coverage assignments and get eight sacks per year? I think it's a small number and his versatility really helps the niner defense. Having said that, I think he might be replaced by Corey Lemonier after this season too. The cap crunch is coming and Brooks and McDonald will be into their thirties.

40
by theslothook :: Tue, 06/24/2014 - 11:54pm

Yes. I think both of them are valuable but expendable when you have too many outstanding players to pay. It's how I feel about wagner and kj wright.

I will also agree overall with your chancellor argument. AS for bennett, he's versatile, but as a pass rusher, i think there's a noticeable difference between him and say Aldon Smith, who is absolutely elite.

43
by Perfundle :: Wed, 06/25/2014 - 7:21pm

Why does it matter if a player would play worse in a different system? If you put Peyton in a zone-read or Air Coryell scheme, he wouldn't be as good as he is now. I don't see it as an indictment on a player if he is put in the best position to exhibit his talents, as long as his responsibilities aren't reduced.

46
by Karl Cuba :: Thu, 06/26/2014 - 7:48pm

To put it simply, Chancellor's value is being boosted by the awesomeness of the guy stood behind him. And I think his responsibilities are being reduced, he doesn't have to play deep half etc.

48
by tuluse :: Thu, 06/26/2014 - 10:19pm

He is only able to play the role he does because other better players play harder roles around him.

Like a slot receiver with a really high per play value. Someone like Patrick Crayton.

50
by theslothook :: Thu, 06/26/2014 - 10:41pm

See...are we sure it's structured that way? We don't act like ed reed's deep safety skills were what made ray lewis so good. We don't act like justin smith enables patrick willis to play well. And I don't think urlacher made briggs overrated either. Sometimes you can have two defenders doing different types of things. Chancellor is labeled as a safety, but his job is outlined differently and it's not obvious to me that plugging in thomas means you can find a close facsimile to Chancellor.

53
by Perfundle :: Fri, 06/27/2014 - 4:43am

The fact that this same argument has been used for Sherman makes me doubt its veracity. Fahey showed that when Sherman does play man coverage he's one of the best in the game; that he doesn't play it the way that Peterson plays it is irrelevant to Sherman's ability. Why does someone who is essentially a LB/SS hybrid need to have the same skills as a SS/FS hybrid? Are you arguing that it's impossible for him to become the best SS in the game in your eyes as long as Thomas keeps playing deep center?

54
by Karl Cuba :: Fri, 06/27/2014 - 7:54pm

Sherman might be a press cover 3 corner and Chancellor might only be a box safety but they are clearly outstanding at what they do. Some people, like me, think that it's Thomas that allows them to excel in such roles.

55
by Perfundle :: Fri, 06/27/2014 - 8:23pm

Well, Sherman's job is certainly easier than Peterson's because of Thomas, but he makes up for it by completely dominating his position as far as passer rating allowed and interceptions are concerned. Chancellor's not quite there yet, but he would be if he could put together an entire season at the level of last season's playoffs. When players don't have the same level of responsibility, then it's not enough to merely excel. So if Chancellor can lap his competition (adjusting for opponent strength, etc.) for the whole of 2014, then he should be considered a truly elite safety.

56
by theslothook :: Fri, 06/27/2014 - 10:21pm

Thing is, pff says he's actually been great in year's past. It was just this year he turned in a slightly above average regular season performance with a monstrous post season. Yes yes, pff, I know, but I tend to think Chancellor is really really good I think what he does is extremely valuable.

23
by Perfundle :: Mon, 06/23/2014 - 2:42pm

I have to think that anything will get used for contract purposes if it suits their side. It also depends on what actually gets used, because they offer objective as well as subjective data on each player.

24
by Karl Cuba :: Mon, 06/23/2014 - 3:59pm

This.

Plus lots of other stuff.

25
by theslothook :: Mon, 06/23/2014 - 4:27pm

True, but I don't understand why it's obvious that their subjective data is garbage or as Karl put it, "asses to elbows." FO makes similar subjective decisions when they calculate DVOA and we accept that it's just an estimate.

27
by Jerry :: Mon, 06/23/2014 - 6:43pm

DVOA is based on what the official play-by-play says. There are cases where FO cleans some stuff up, or doesn't count kneeldowns as rushes, but that's a lot different than looking at video and deciding who's at fault on a given play. (When FO does it, it's labeled as charting stats.)

Every now and then, Ben Muth will break down a play and point out that you'd need to be in the meeting room to know exactly who had what responsibility. I doubt that PFF has more insight into that than Ben. It doesn't mean everything they do is garbage, but I take their ratings with a grain or two of salt.

28
by theslothook :: Mon, 06/23/2014 - 7:21pm

Yes, it's based off play by play data, but calculating dvoa and dyar come from decisions about how to weigh certain factors. Those are decisions they make based on whatever model they are using. Since we don't know what went into the sausage making process, we're going on faith that they're doing a good job.

29
by Karl Cuba :: Mon, 06/23/2014 - 10:56pm

I just wrote a lengthy response to this but the internet swallowed it, so more succinctly:

- Thomas allows them to play cover 3 most the time without exposing the deep middle, meaning they can play their 6-2 front all day, which in turn means that their linemen are able to one gap and get to the qb more quickly with less risk.

- Seattle's defenders ranked near the top of the DVOA rankings across all different types of receivers. It is far more likely that this is due to coaching and the influence of the game's best safety than each and every other player being so good.

30
by theslothook :: Mon, 06/23/2014 - 11:35pm

I might have agreed with this, but then there was a difference in their pass rushing numbers and their linebacker coverage numbers between last year and this year.

33
by Jerry :: Tue, 06/24/2014 - 3:54am

Aaron's spreadsheet is a black box, but there's a huge difference between that and Aaron deciding that since he likes a particular play, he'll add a couple points for it.

44
by tuluse :: Thu, 06/26/2014 - 5:51pm

FO doesn't pretend that you can use their stats to grade individuals with any certainty. There are disclaimers everywhere.

45
by theslothook :: Thu, 06/26/2014 - 6:12pm

When FO calculates dvoa, I imagine they are using a predictive model that requires different weighting assumptions, like how much more value a 3rd and long on the 2 yard line produces versus a 1st and 10 at the 25 yard line.

47
by tuluse :: Thu, 06/26/2014 - 10:17pm

They don't pretend the value created is one player's doing. On every article they disclaim, so-and-so produced this value while in team X's offense playing with team X's players and coached by team X's coaches.

49
by theslothook :: Thu, 06/26/2014 - 10:34pm

I understand that. FO is doing macro analysis, pff is doing micro. We can debate about which site uses much more subjectivity in their analysis, but all I was saying is that FO isn't just taking raw data logs and reorganizing the data. They are taking data logs and modeling them to spit out one number that is supposed to encompass team quality. There is subjectivity involved in that.

51
by tuluse :: Thu, 06/26/2014 - 11:22pm

They use the same model on every player and team.

Can we be sure PFF is doing the same?

52
by theslothook :: Fri, 06/27/2014 - 12:25am

that's fair to say.

13
by LionInAZ :: Sat, 06/21/2014 - 11:36pm

Steve Young and Rodgers are listed at 6'2, same as the not very mobile Matt Stafford and The Former Packer QB Who Can Not Be Named Here. Jay Cutler is 6'3, Kaepernick and Cam Newton are 6'5, same as the statue Peyton the Manning. I don't think any of them are considered especially short as far as QBs are concerned, at least compared to Wilson, Brees, and Vick.

18
by Perfundle :: Sun, 06/22/2014 - 5:05pm

Of course there are tall QBs who are mobile and short QBs who aren't. I'm just wondering if shorter QBs (let's say <= 6'2") are more mobile than average.

41
by Scott Kacsmar :: Wed, 06/25/2014 - 2:53am

On average, I lean towards yes, though it's probably not a significant difference. I wish I had career rushing data that was wiped clean for kneel downs (Some day I suppose...)

4
by jtr :: Fri, 06/20/2014 - 3:15pm

Interesting analysis. In the future, I would rather see positions relative to the center (and thus whole offensive line) than relative to hash marks. Wilson's tendency to throw from the right hash might mean he likes to scramble a little that way, or it might just mean the 'Hawks tend to snap it from the right hash. Without knowing where the snap came from, we can't tell where the qb was relative to the pocket close to the middle of the field.

6
by panthersnbraves :: Fri, 06/20/2014 - 4:05pm

That might also allow for some minor weighting - a throw from the sideline would "count more" than a throw from a single step outside the Tackle...

17
by G-49er :: Sun, 06/22/2014 - 2:21pm

Good point, I'm well aware of that, just as I wrote in the article, the spot from where the offense operates is depended on the previous play. If the 'Hawks like to run the ball to the right then chances are they will snap the ball more from the right hashmarks.
The attempts could be split, for example, centered, pocket left side, out-of pocket left side,... But for that I'd have to come up with different chart design.

35
by ChrisS :: Tue, 06/24/2014 - 2:27pm

Interesting article with some useful data. Assuming you know where the ball was snapped can't you just adjust the spot of the throw to compensate? e.g. ball was snapped 3 yards inside of the right hash, and the ball was thrown 3 yards outside the right hash, so the pass was made 6 yards right of the center and then locate that point 6 yards right of midfield.

42
by G-49er :: Wed, 06/25/2014 - 4:25am

That could be done, that way all the throws would be mapped out relative to the "axis" defined by center and quarterback, in your example, 6 yds to the right of that reference line.

8
by Pen :: Fri, 06/20/2014 - 10:03pm

Will finish reading after this post, but up front there is a huge flaw in your analysis: You just look at playoff performance of a pocket passer who only faced one elite defense and two mobile QB's who faced nothing but elite defenses. Well, except for Denvers defense in the SB.

You couldn't have a less controlled subject matter.

Gee, guess which QB's were being chased all over and which one wasn't?