BGA: The Double Team Project – Part 1

Traditional statistics and most modern analytics measure how effective a player is without necessarily taking into account assignments or degree of difficulty. While I’ve made passing reference in my game analysis to how often certain players have been double teamed, nobody tracks this, so I’ve been keen to figure out whether the reality matches up to our expectations and my recollection from watching the film.

Over the last few weeks, I’ve been charting every defensive snap from the 2013 season to try and get a handle on how often each of the Jets defensive linemen were doubled in order to investigate tendencies based on situation, different teams’ schemes and certain individuals.

I’m still in the process of compiling the numbers, which – as you might expect – are fascinating. We’ll get to the first set of numbers in part two in a few days, but before we get to that stage, I wanted to write about what I’ve learned about teams’ approach to who gets double teamed and how they set up their run blocking schemes and protections.

After the jump, I’m going to talk about the methods I’ve employed in compiling this data, how I’ve treated certain situations and what this data tells us. I’ll also be outlining what the data does not tell us.

Introduction

When I began this project, my initial intention was simply to add up the total number of double teams so that we could see which players were doubled the most often and in which situations during the season. However, it soon became apparent that this would be inadequate, because there were players who were double-teamed temporarily and there needed to be a way to differentiate between those and full-on double teams.

Here was the system I devised:

I attributed a double team to a player if he was being actively blocked by two players at the end of the play. Anyone who was initially doubled, but not throughout the play, got credit for a “peel” (on running plays) or a “chip” (on passing plays) which I basically used as catch-all terms for anything that slowed their progress without constituting a full double team.

Run Blocking

In order to attribute a peel, I determined whether a second blocker impeded the progress of the defensive player. This might be anything from a getting one hand on him on the way to the second level to a temporary double-team where one of the blockers peeled off to the second level. Just waving a hand out in his general direction on the way through would not be sufficient, but if solid contact was made to hold him in place or even move him then that would be a peel. A bump on the way through would also count.

Let’s look at an example:

On this play, Harrison and Richardson would get credit for a “peel”. Although they are both initially doubled, one of the blockers peels off to block a linebacker in each case. Everybody else is single-blocked and nobody is double-teamed at the end of the play, so nobody gets credit for a “double team”. Although Coples gets caught up in traffic, the other blocker he makes contact with is not actively seeking to block him. Had the right guard turned around and left his man to double-team Coples, then he would have been credited with a double-team.

Pass Blocking

As with a peel, for a player to be credited with a chip on a pass play, their progress needed to be impeded by deliberate contact from an offensive player. It’s important that the offensive player needed to be actively trying to make a block. If the contact was incidental, because a blocker got driven into someone, or even if the defensive player deliberately initiated the contact that would not constitute a chip or a double, unless the blocker reacted to that and started making an effort to block them.

This is a simple example of a pass play. Everyone on this play is single blocked, but Wilkerson does get a shove from the left tackle, impeding his progress. While this is only a one-handed shove with a minor effect, I would still credit him with a “chip” on this play.

Let’s now look at a more complicated pass play.

On this play (which was actually from 2012 so not included in the study), Wilkerson is initially doubled. However, #74 ends up impeding the path of DeVito who has deliberately rushed at an angle in order to free up Ellis. Ellis is chipped by #55 who ends up chasing him to the quarterback. McIntyre's interior rush is single blocked and any contact with a second blocker is incidental because the player is not actively trying to block him. Ultimately, Wilkerson would get credit for a "chip" because he is not being doubled as the ball is thrown. DeVito would be credited with a double, even though he has initiated that contact himself, because the blocker has made an effort to impede his progress. Had the blocker stayed on Wilkerson then the contact from DeVito would have been incidental. Finally, although Ellis is chipped, he would not get credit for a chip because there is no second blocker making an effort to block him. In order to get credit for a chip or a peel, you must be blocked by two different players.

There were plenty of special situations, which needed to be treated consistently. In the event of a stunt, if the offensive linemen simply swapped men without any contact, nobody would be credited with a chip. However, if a player was contacted prior to the stunt by one blocker and then picked up by the other, that would be a chip because their initial progress was impeded. If a player was chipped but then left unblocked, that does not constitute a chip. However, if they were chipped twice and then left unblocked that would constitute one chip. Finally, on some occasions a would-be pass rusher occupied more than one blocker but never made contact because they deliberately made no effort to rush the passer. On these occasions, I credited the pass rusher with a chip for the extra blocker(s) as long as they did not react by going to block someone else instead before the pass was thrown.

A player can be chipped twice on one play, as long as you're chipped by two different players before (or while) being blocked by a third. You could also even be chipped and doubled on the same play. On the rare occasions where there was a triple team, the player would get credit for two double-teams. However, this only happened a handful of times all season, because if a player was being triple blocked, one or two of the blockers would usually break off from the triple team to try and pick up someone else. In that instance the defensive player would only get credit for a chip and a double-team (or two chips if two blockers broke away). In some situations, teams used moving pockets which would often result in three guys blocking two pass rushers almost in unison. Where there was no obvious double-team, each player would get credit for a chip here. Also, if someone that was double teaming a player stayed on that player, but clearly tried to impede the progress of another player at the same time, the initial double would still count and that could still constitute a chip. Similarly, if an offensive player chipped someone but then ended up blocking them anyway, that would not count as a chip.

After talking to some analysts who’ve been in contact with NFL linemen about how they handle double teams and what they’re trying to achieve, they agreed that my approach aligns closely to what they would expect teams to do if they were compiling similar information. The point was made that I could have differentiated between a one-handed momentary chip like the one on Wilkerson in the second clip above and a more substantial two-handed temporary double-team like the one on Wilkerson in the third clip. Nevertheless, the data produced should still be pretty informative.

What I learned about double teams

There are a lot of different ways an offensive line will set itself up, which vary from play to play and are affected by scheme, situation and personnel. When we get into the data, we’ll see that there are some major differences from team-to-team and even from the first meeting between two teams to the next.

It seems apparent that double teams are not always dictated by the offense as you might think. The way the defense lines up will often force the defense to single-block a dangerous player. For example, you might see the Jets leave a cornerback with no safety coverage, but the safety comes into the box to cover, which enables one of the inside linebackers to blitz, which engages a would-be double-team blocker on one of the interior pass rushers. Also, the Jets defense will shift just prior to the snap, often putting one of their defensive tackles opposite the “A” or “B” gap so that this player can try and shoot that gap and potentially occupy both blockers, at least temporarily.

The most amazing thing is how often a pass rusher makes no effort to take the shortest route to the quarterback and instead appears to deliberately draw blockers out of position to create space for someone else. You’ll often see an outside rusher run extra wide to spread out the pocket or an interior lineman making his initial burst laterally rather than straight ahead, as DeVito did in the third gif above. Also, it was not uncommon for a rusher to appear like they gave up trying to rush, when in reality the fact that they backed off prevented them from driving their blockers into the path of imminent pressure off the edge.

One thing that’s obvious is that the defensive tackles are doubled initially a lot more often than the ends. That makes sense because obviously those are the closest players to the ball, so you would often want to ensure that right up the middle is not the path of least resistance. With a defense like the Jets where they have versatile players lining up all over the front seven, this suggests that a lot of the time the players who are initially doubled are dictated by how the team lines up moreso than who those players are. However, where the players’ talent comes in is more in terms of who ends up getting doubled.

In many cases, a tackle will be doubled initially to slow down that immediate burst and allow the quarterback to drop back unencumbered (whether that be to throw or to hand off) but this will be a token double and then they will either peel off to the second level or drop off to double the rush off the edge. You’ll also see more offensive linemen forced to stay on the double team if the defensive player is starting to get penetration or collapse the pocket. Finally, you’ll see offensive linemen having to leave their initial man to rescue someone else who is about to be beaten one-on-one. In that regard, the importance of someone who gets credit for a double team is generally reflected by the fact they have contributed by being disruptive.

There is not necessarily always a double team (or chip/peel) on every play. Sometimes linemen head straight for the second level or sometimes the Jets would send extra players and everyone would end up single-blocked. In zone blocking, particularly on stretch plays, the blockers are on the move, so while peels are not uncommon, you rarely get a double team unless it’s on the backside.

As noted earlier on, most basic statistics, analytics and analysis do not account for how often a player is double teamed. However, they are taken into account to some extent by Pro Football Focus, insofar as they would not typically grade down a player who was double teamed out of the play on a running down and would generally give a higher grade to anyone who beat a double. Is that enough though?

What the data will tell us

Once I’ve finished compiling the raw data, I’ll be able to provide you with a total figure for double teams and chips on pass plays and double teams and peels on running plays for each of the Jets linemen.

What this will tell us includes the following:

- How often each player was doubled with reference to one another, taking into account position.
– How often a double team on each player constituted a full double team, again with reference to one another.
– How the frequency that players were double-teamed developed over the course of the season.
– How certain teams’ approach differed from one another.

What the data will not tell us

This initial data will not be definitive. However, it will give us the raw data to do some further investigation.

- If a player was double-teamed a lot this doesn’t necessarily prove that they should have received better consideration for Pro Bowl or similar honors. (To determine that would require a comparison to see how often some of the other elite players were doubled, perhaps in terms of common opponents).
– If a player was double-teamed a lot this doesn’t necessarily mean their role was more difficult than it was last year. (To determine that would require a comparison with similar data from last year).
– If a player was double-teamed a lot this doesn’t necessarily mean that was any more or less than a former Jets player. (To determine that would require a comparison with similar data for that player in a prior year).

Naturally those additional studies will be conducted in some of the later installments of this series. I also intend to look at it from the other side of the ball, in terms of which offensive linemen for the Jets were employed in single blocking or in double teams.

In Part 2 of the Double Team Project, in a few days, we’ll be looking at the numbers for double teams on pass plays and what we can glean from these.




42 comments
levi
levi

Damn Bent, thats a lot of work going into this BGA. Im looking forward to part two.

bklyndude
bklyndude

This is the stuff that makes the Jetsblog such a good site,  outstanding !!!  

Not only am I getting psyched for the season watching our DL,  but it's nice to see how these guys operate as a unit.

UncleJoesJetFarm
UncleJoesJetFarm

Its not just happy talk folks!     "Smith, for his part, said he started to "get it" as the season wore on and his film study helped him notice things he could have exploited with his feet as he went through his progressions. It also helped that, later in the season, Smith was able to process what he had to do much faster, something he chalked up to gaining experience." http://www.nj.com/jets/index.ssf/2014/06/why_the_jets_sound_so_optimistic_about_quarterback_geno_smith.html

harold
harold

Very excited to see part two...


Dare I say... Can't Wait!

Brendan
Brendan

Can't wait to read the next part in this. Are you going to keep track of players who start out doubled vs. those who end up doubled as separate counts? 

bradysucks
bradysucks

This BGA on double teams makes me think about double and triple teams which take place on the Jets blog.


Perhaps we should have Bent write one of these articles about Hank/Naples and how he'll make a statement and suddenly be double teamed (blocked) by Brendan and tsjc68....and even triple teamed when Bent himself chimes in.....


Poor Hank....Lets give him an award for going against all of those double (and triple) teams.......lol

Lord Harry Latts
Lord Harry Latts

As always Bent, we all appreciate this type of stuff. Can't wait to see Part 2.

Bent
Bent moderator

Off-topic:


"These things happen in the box.  There was chest to shoulder contact and I got hit in the eye" - Luis Suarez to Uruguayan media in response to allegations that he bit an Italy player in today's game.

Pete 57
Pete 57

Great stuff as always Bent. Some of what you wrote confirms what I thought I had seen. Many of Rex's pass rushing schemes have players that do several other things other than rush the passer. Obviously he knows what he's doing and I do not, but many times it's frustrating seeing D lineman simply standing in front of his man and not making any effort to get to the QB. I think it was 2 or 3 seasons ago with the last game against Buffalo, and Rex stated that he would let the D line just get after the QB. I loved it. I think the talent we have assembled in the front 7 is superior, and selfishly I want to see more of the just letting the players go after the QB. I miss the sack exchange days. It was so exciting to watch that D every time there was a pass play.

williamg1
williamg1

Thank God for Bent's stuff the fill the void during the break before training camp. 

Alec Wilson
Alec Wilson

My god you have away to much time on your hands...lol....STATO !!.....I'm off my work the now but even with mega time on my hands I don't think I could do what you do man..cudos to you

SackDance99
SackDance99

Interesting.  The second gif was a fairly common stunt by the defensive linemen (DeVito and Ellis).  I'm not sure that I'd say that DeVito was "double-teamed" because he was intentionally clearing out the OLs for Ellis.  The OL had no intent to double DeVito and it's likely that Rex and Co. saw that the Texans would be vulnerable to this stunt (looks like the Texans splits were too close making them bowling pins for DeVito).

Monty
Monty

Cool.. thanks

Bent
Bent moderator

@levi FTR it takes about 1.5 hours per game because I have to watch the coaches film 5-6 times, constantly pausing it, to determine who doubled who.

Bent
Bent moderator

@Hanknaples Oh, sorry guys, Fonzie thinks this is a waste of time.  Guess I won't bother with the rest of the series then.

Bent
Bent moderator

@Brendan Basically yes.  You'll get a total (and percentages, and in some cases breakdowns) for double teams and peels/chips.  Double teams would comprise anyone who ends up being doubled.  Initially doubled guys that aren't doubled throughout the play would come within the chip/peel column.


As a general rule, most (not all) doubles on running plays are initial doubles that are held onto throughout the play, whereas most peels would be initial doubles where one guy peels off to the second level (although some of these are just momentary contact).  The main exceptions would be on stretch/zone plays where everyone is on the move so the double sometimes doesn't lock on until the middle or end of the play.


Pass plays are a little different because there would be more plays where a chip happens in the middle of a play and doesn't always happen as part of a double team.  On double teams on pass play there are a more even split of initial doubles sustained throughout the play and doubles at the end of a play.


I've got the numbers done now.  Part two will be on Thursday.

Monty
Monty

@bradysucks please dont put the image in my head of jets fans being "double and triple teamed " :) 

williamg1
williamg1

@bradysucks Hank is never by himself. He's has Fonzie and possibly more personalities. 

BDarc23
BDarc23

@bradysucks

That's too funny. Everyone has to start partnering up.

NYCPEinGermany
NYCPEinGermany

@Bent

he clearly bit him. An italian gets the red card for basically nothing and this guy pulls a mike tyson and gets away with it!

Disgruntled Jets Fan
Disgruntled Jets Fan

@Bent Do they? Can't recall any biting from my days playing soccer.  Shoving, kicking, punching,  yes.  Biting?

a57se
a57se

@Bent 

Then you look at the video and he looks like Hannibal getting started early on Dinner before play acting like he got hurt.

williamg1
williamg1

@Pete 57 I think these guys are going to get after it this year. And I think a guy like Pryor will be coming on some exotic blitzes. 

Bent
Bent moderator

@Alec Wilson Funnily enough, Statto was my nickname in college.  In fact, nobody knew me by my actual name.  Good preparation for TJB, I guess.

Bent
Bent moderator

@SackDance99 That one was admittedly on the edge, but I gave DeVito the credit because the lineman turned back towards him and threw a shoulder at him to impede his progress.  While there was no intent, the reaction was there and although DeVito instigated it, I feel he should get credit for occupying an extra blocker with that move.


Most stunts are more straightforward to make a decision than this because it's more clear-cut who picks up who.  There wouldn't have been too many plays like this, I picked it deliberately to show how difficult it is to decide in rare cases.  With a whole season's worth of data, my hope is that the accuracy is still good even with plays like this which could go either way.

levi
levi

@Bent  As I started reading I was like holy crap thats gotta be time consuming. 

NYCPEinGermany
NYCPEinGermany

@Bent @Hanknaples

or you could just ban fonzie (not necessarily for this, but for the grand accumulation of his trolling)

Pete 57
Pete 57

William, I'm very confident Rex knows what he's doing, it's just my own personal desire to unleash ALL the hounds at once!

Brendan
Brendan

@Bent (Currently ranked the 69th CB in the 2016 draft by nfldraftscout.com) 

Bent
Bent moderator

@bradysucks @Hanknaples But there's also Fonzie and as explained in the article, 3-on-2 just counts as two chips.