While doing research on my last article, I noticed one very interesting fact: Dominant 3-4 defenses tended to have a star 5-technique defense end. The 3 best 3-4 defenses in terms of Advanced NFL Stats’ dEPA (defensive expected points added) in the NFL right now are San Francisco, Arizona and Houston and each team boasts impact 5-technique defensive linemen like Justin Smith, Calais Campbell and JJ Watt, each of which is among the top five 5-technique defensive linemen according to ProFootballFocus. This got me to thinking: everyone knows that the quarterback effects offensive success more than any other position on the field (hence why Aaron Rodgers and Peyton Manning can keep winning games without good offensive lines and running backs), but is there a position on a 3-4 defense that is most important to defensive success?
Traditionally, the hallmarks of a good 3-4 defense has been it’s nose tackle and outside linebackers; indeed in 2009 when Green Bay switched from a 4-3 to a 3-4 defense, general manager Ted Thompson drafted nose tackle BJ Raji with the 9th overall pick and then traded up back into the 1st round for outside linebacker Clay Matthews III. The argument has always been made that a dominant nose tackle that can eat up multiple blockers and outside linebackers who are athletic enough to rush the passer are the keys to a dominant 3-4 defense. You could argue that Green Bay seems have both positions covered, both Clay Matthews III and BJ Raji are both dominant players but while that seemed to have translated to success in 2009 and 2010, it didn’t seem to matter much in 2011 and 2012.
What I’ve done is a correlation analysis using ProFootballFocus’ player grades and comparing them to overall defensive efficiency measured in dEPA. I’ve flipped the signs for dEPA to just to avoid making it an inverse correlation. I’ve included both Pearson’s r and chi2, I’m not really much of a statistics guy so I have no idea what the difference is between them, but if you happen to know more about this, leave a comment and I can adjust my analysis if needed. Overall, the way to read these figures is that a value of 0 means there is no correlation at all while a value of 1 means that there is perfect correlation. So for this case, the higher the number the more “valuable” that position is to defensive efficiency. I’ve also included a positive control by correlating dEPA vs. dDVOA (from Football outsiders) and they are 91% correlated, which basically means this analysis holds for both metrics. Finally, I’ve included a negative control by looking at the correlation between how well the offensive center plays versus how well the defense does; presumably how well the center plays has no relationship to how well the defense plays.
| Parameter | Pearson r | X 2 |
| DE | 0.4905 | 0.2405 |
| NT | 0.05412 | 0.002929 |
| OLB | -0.2552 | 0.06515 |
| ILB | 0.1474 | 0.02172 |
| CB | 0.4116 | 0.1694 |
| S | 0.5194 | 0.2698 |
| FO | 0.9168 | 0.8406 |
| C | 0.1327 | 0.01762 |
The results are pretty striking. First off nose tackle seems to be completely irrelevant in the 3-4 defense to the point that it appears that it’s more important how well your center plays than your nose tackle when it comes to defensive efficiency. A great example of this is that the best 3-4 nose tackle combination in the NFL in 2012 was Jay Ratliff and Josh Prince-Brent but the Cowboys were the 3rd least effective 3-4 defense in 2012. On the other hand San Francisco had the best 3-4 defense this year but featured Issac Sopoaga, who graded out as the worst 3-4 nose tackle.
Moving on to outside linebackers, it appears even worse where the better a team’s outside linebackers, the worst the defense is overall. While I wouldn’t say that the Packers should go out and find some bad outside linebackers (as they have the worst 3-4 outsider linebacker in the NFL in Erik Walden), there isn’t any statistical correlation between how well outside linebackers play and how well the defense plays. Some of the best 3-4 outside linebackers played on terrible defenses this year; Justin Houston (KC), Victor Butler (DAL), Ryan Kerrigan (WAS), DaMarcus Ware (DAL) Lamarr Woodley (PIT) and James Harrison (PIT) all have overall positive grades according to ProFootballFocus but played on some of the worst 3-4 defenses this year. On the other hand, Houston seems to have shown that you can have a dominant 3-4 defense with Brooks Reed (-.9), Connor Barwin (-19.0) and Whitney Mercilus (-13.8).
So what seems to be the most important position in a 3-4 defense? Apparently it’s 5-technique defense end and safety. In other words, the better your defensive ends and safeties are, the more effective your 3-4 defense. If this doesn’t strike you as cathartic nothing will. In 2009-2010 the Green Bay Packers had a dominant defense, which included defense end Cullen Jenkins, and safety Nick Collins. In 2011-2012 they did not. More importantly, Green Bay hasn’t managed to recover from losing both of them.
Below I’ve run another correlation analysis looking at how well the defense ends and safeties have played compared to overall defensive efficiency over the period of 2009-2012 (the years where the Packers played a 3-4 defense). The results are rather startling; defensive end play had a Pearson’s r of .7455 while safety play had a staggering Pearson’s r of .9115 (again a value of 1 is perfect correlation). I think both results shouldn’t be all too startling when you think about them. A perfect example of how important defense end play is when you consider the effect that DE Justin Smith had on the San Francisco defense; with him, OLB Aldon Smith was on pace to break the NFL sack record, but without him Smith hasn’t managed a single sack. In terms of efficiency, Aldon Smith managed a grade of 1.37 per game with Justin Smith in while grading out at -.76 average for every game Justin Smith was out.
| Parameter | Pearson r | X 2 |
| DE | 0.7455 | 0.5558 |
| S | 0.9115 | 0.8309 |
| C | -0.08444 | 0.00713 |
I don’t think any fans were thinking that losing Jenkins and Collins would be a improvement, but I doubt many really knew the true affect that both players had on the defense. It could be argued that if Collins hadn’t been lost with a neck injury in 2011, the Packers very well could have won back to back Super Bowls.
Luckily, it’s not all bad news for the Packers; wisely it might seem as if the Packers already know this (heaven forbid if I’m the first person to notice this trend), which might explain why BJ Raji has been spending less and less time at nose tackle and more and more time at defensive end (where he actually has a +12.1 grade). While safety Morgan Burnett isn’t as good as Nick Collins, he is still a good player and maybe with a season or two more of experience will be able to fill Collin’s shoes; perhaps adding Charles Woodson at safety full-time might make the situation better. Furthermore, adding Mike Neal and Jerel Worthy are active attempts to improve the defensive end position. However, I wouldn’t be all too surprised to see a safety or another defensive end drafted fairly high in this years NFL draft.
——————
Thomas Hobbes is a staff writer for Jersey Al’s AllGreenBayPackers.com.
——————
I’ve spent the last 1.5 years in a fetal position mumbling…
This wouldn’t be happening if Collins didn’t get hurt.
This wouldn’t be happening if Collins didn’t get hurt.
This wouldn’t be happening if Collins didn’t get hurt.
Who knows, you might be right
I’d love for the Packers to have a defensive end like Pittsburgh used to have in Aaron Smith…
There are a couple interesting free agent defensive linemen that the Packers might be interested in; however interested and willing to sign are two completely different things for Ted Thompson, hopefully this years draft class has a strong DL group
Who’s available? If they’re interesting I’m sure Thompson will consider them.
Jason Jones (SEA) was a potential candidate last year from Tennessee and he’s likely going to be testing the market again after he injured himself this year.
Richard Seymour has experience in the 3-4, but is on the older side and is coming off an injury as well.
Finally, Henry Melton is sort of a wildcard, he’s a DT with the Bears, but who knows what scheme they’ll ultimately play now that Lovie Smith is gone so he might be available, and stealing a good player from a divisional rival is always fun.
Will Thompson spend the money to get any of these guys? I doubt it, I’m sure he’s more worried about locking up Matthews and Rodgers right now
They could have had him in last years draft. His name is Devon Still, we took Worthy instead.
I’ve said it over and over again. From the end of the 2010 to the beginning of the 2011 season, Cullen Jenkins left, Charles Woodson got old, Nick Collins’ career suddenly ended, and for whatever reason Tramon Williams and BJ Raji have struggled (or at least, they’re just not the same players they were in 2010).
We always say no excuses here, but at the same time, what is Dom Capers supposed to do exactly? Outside of Clay Matthews, those were the best five players on the team.
Like I’ve always said, you don’t need the best players, you simply need the right players. Philadelphia is living proof of that.
Right now, I just don’t think Capers has the right players to work with at the moment. The only star player on that defense is Clay Matthews. Outside of him, all we have are just a bunch of rookies.
This defense needs to develop, and they need another pass rusher up front. That alone could make all the difference for this team IMO.
Based on the analysis, it’s actually a waste to keep BJ Raji at nose tackle, in fact it’s better to just keep him at DE and find another NT to help Pickett. As for pass rushing, they already have Nick Perry and outside linebackers apparently aren’t all that important in a 3-4 defense.
Given the assumption that teams look for the stud DT to build the team around, the stud DE after that provides a formidable front.
I’d like to see a correlation between DT and DE in the 34. Is there a positive correlation between a positive DT/DE combination and team defense? What happens to the relationship if the DT is weak (negative) and the DE is strong (positive)? Or vice-versa?
My point is, many teams in the 34 first try to have that stud at DT. It isn’t until they get two-thirds of the line built that the defense starts to work. Thoughts?
Based on what I’ve analysed, I would say that NT has really no effect on D-line play, even the best 3-4 NT really don’t contribute all that much. In terms of how much a NT helps a DE, that would be very hard to quantify, as you would have to determine how many times offenses doubled the NT instead of the DE.
Isn’t the NTs job to allow all others to do their job? So if the DE is making plays, or LBs, the NT is properly taking up double teams. But if I know PFF, they build that into their analysis of NTs.
I’m pretty sure PFF’s grading system are secret, but I would agree with you chances are really good that they take that into account for NT. After all, it’s not like your typical NT is making many “splash” plays anyways.
I have always thought Raji to be more the DE type. Kind of thought he was miscast at the nose…but I don’t paid to make those decisions..:)
Raji is a little too big and too short to fit into your typical 3-4 DE mold, but he does seem to do better at DE than he does at NT, perhaps another reason why the Packers are using him less and less at NT
So do you think the Packers should draft John Jenkins if he’s there in round one? Put Jenkins at nose tackle and move Raji to end. They say that Jenkins is very quick and had 15 QB pressures this year, and that’s playing in the SEC.
These are interesting numbers and they tell a regular season story when teams play poor teams and get all kinds of fat data. Houston is a good example of a team that showed great on paper but had trouble against the better offenses. The playoffs have a way of seperating the stat teams from the real men.
To the conclusion that having a dominant DE and a great safety is a big key to the 3-4, I wouldn’t argue. But if your D-line or LB’s are not doing their jobs or missing assignments, then it won’t translate into good numbers for anyone. The 49er’s game showed that you need to do your job so that the other components can do their job. Defense really is about team.
Beware the statistic!
Arguably there are too many variables to simply say that if the Packers somehow signed JJ Watt and Troy Polamalu suddenly the Packers would have a dominant defense (it might happen, but who knows). What it does say is that teams can get around deficiencies in certain positions (Houston and their linebackers is an example) but you can’t hide from poor D-line play, just like you can’t hide from poor QB play on the other side of the ball.
Totally the right framing for our D-line play.
My belief is that they will draft 2 defensive players early. The D-Line has plenty to offer in this draft. I also believe the Safety position is loaded too.
The Defense End position you noticed is important. That’s why there was a sharp drop off when Cullen Jenkins left the team.
I guess the question is, do you think the Packers feel they have enough with Mike Neal and Jerel Worthy, both of which were 2nd round draft picks? You could make the argument that Morgan Burnett was initially drafted to take over for Charlie Peprah and not Nick Collins, so a second safety draft high would be expected.
“I guess the question is, do you think the Packers feel they have enough with Mike Neal and Jerel Worthy, both of which were 2nd round draft picks?”
As Curly of the Three Stooges would say,”I certainly hope not!”
Can you really make an accurate assessment of Worthy after only a handful of games as a rookie? He did grade out as the worst defensive linemen on the Packers, but that’s not really saying much of a rookie, even a high draft pick. As for Neal, he did some good things during the 2nd half of the season.
Regardless of whether Worthy will be any good or not, he will not see the field next season so I hope they draft another DE high in the draft this year. As far as Neal, if he stays healthy this coming year and improves he will be a keeper.
Do you think Worthy’s injury will preclude him from playing all next season? I have my doubt’s that he could make it for opening day, but getting stashed on the PUP and then started around week 6-8 wouldn’t be out of the question.
You can’t use the analysis of Watt, Campbell and Smith to make your point since none of the 3 played at the ‘5-technique’ full time. Hell, there are games Watt played at the 2, 3 and 4-technique more than the 5. Same with Smith. Not to mention that none of the above defenses are static. In other words, they don’t play base full-time where a ‘5-technique’ even exists. The players you’re referencing are all-stars in any scheme.
Also, Cullen Jenkins was a one-dimensional defender in the 3-4. Effective as a nickle rusher, mostly on the inside, and average to awful against the run when asked to play the true 5-tech.
Like any argument, it’s truly nuanced. I believe any defensive scheme works when players execute it and the opposition match-up is favorable (it’s always variable) and personnel flexibility in any front.
P.S. – Good work on the analysis, not easy to do when PFF and FO are among the few offering numbers to work with.
P.S.S. – Both are deeply flawed and offer a baseline for discussion.
Hard work, though. No desire to discount it.
I agree that with most 3-4 defenses the defensive linemen play all over the place, but the same could be said for the Packers. For instance Neal and Worthy played 5 technique in the base and 2, 3 and 4 technique in the nickel. Regardless, Neal and Worthy didn’t have good enough seasons to really effect defensive efficiency, hence the problem. Perhaps I should change my argument to DL are important to defense success outside of the nose tackle?
I would make the argument even simpler. The 3-4 is predicated on stopping the run in base and forcing unfavorable down-and-distance, i.e. 3rd and long, at which point it’s easier to confuse and hide your defensive looks pre-snap out of a 3-4 than it is a 4-3. That’s why I prefer the 3-4 with proper personnel.
The problem for me, especially evident late in the season: The Packers frequently put there opponent in 3rd and medium to long but couldn’t get home on 3rd down. That was the most prominent issue to me. The ‘what to do’ is more difficult and based on match-up and personnel. I believe it’s a combination of:
1) You now have depth, speed and range in your defensive backfield, but they don’t keep plays in front of them by tackling immediately upon a completion or pressing the line on running downs to tackle. Great coverage, inconsistent finishers when tackling.
2) Combination of coaching staff and on player discipline: Force a runner to pass, force a passer out of the pocket so he only has 1/3rd of the field to play with.
3) They need better personnel at the OLB position. Perry and Moses deserve development. Walden and Zombo are maxed out, they need to walk. Must bring in 2 additional guys.
4) Need an interior push against good Packers. Raji needs the green light to ‘JET’ more often, also they need depth.
For the foreseeable future, it’s a passing league. Defenses will earn their keep on 3rd and long in their nickle packages. You need more ILB’s and interior lineman that can get skinny and push the QB outside. You need better discipline and concept against running QB’s. Contain, don’t rush.
I love the versatility of a 3-4, but they need to get better on 3rd down. I wouldn’t count on a 3 down 5-technique to come into play. The Watt’s of the world are extremely,extremely rare.
This defense isn’t far away, but the growth of these rookies as 2nd year players is more important than a rookie class. Those interior guys rarely contribute their first year in the NFL.
Supposed to be ‘passers’ not ‘Packers.’
I agree with most of what you mentioned, hopefully we see improvement from an offseason and more experience in the defensive front 7 which ultimately leads to more pressure on the quarterback.
Agreed..
JJ Watt was doing the most damage from the inside, not the edge, this year.. At least, from what I’ve seen of him.
You lost all credibility with me when you Cullen Jenkins was awful vs the run. Pack defense vs teh run improved greatly the day he replaced KGB and fell apart the day CJ was let go in FA. End of story.
Jenkins did his damage in sub-packages where they moved him inside, especially on passing downs. He only played the ‘5-technigue’ part time in their base 3-4 defense for a reason, he wasn’t a stout run-defender.
You lost all credibility when you wrote the comment.
Ed isn’t exactly wrong about Jenkins.
He most certainly did replace KGB as the Packers starting DE, putting Kabeer into a pass-rushing specialist role. Cullen played the run very effectively as the DE, he was an instant, and massive, improvement to the run D at DE.
The issue is, that was back when DC Bob Sanders was running Bates’ 4-3 match/press-man scheme, not during Dom Capers’ 3-4 scheme.
So, Ed isn’t exactly right, either, at least not in the context of this post..
Let’s just all agree that I have no credibility at all, and call it a day.
I will say that according to PFF grades, which is what this article’s statistics are based on, in 2010 Cullen Jenkins had a +14.3 rating against the pass, good for 3rd among 3-4 DE. Jenkins had a -1.0 against the run, good for 25th among 3-4 DEs. You can debate that for it’s worth, but that’s what PFF says at least.
Lots of work for nothin. Can make stats say whatever you want. Seems to be the case here. Even w a lot of injuries we were bout #10 overall. With a healthy defense I could see it bout #5. Hope you enjoyed the work…
Except that Thomas seems to have looked at the stats and THEN came to a conclusion on their meaning. He didn’t appear to start with an assumption and then try to use the stats to support it.
It’s better than using vague judgements with no concrete foundation of analysis.
“Can make stats say whatever you want” pre-supposes this started with a hypothesis.
Statistics do not lie, despite what the adage says. Liars lie.
Mr. Hobbes has statistics to verify that a defense (surprise surprise!) is won in the trenches.
Wonderful work Mr. Hobbes, as always. Being a mathematician, I appreciate the effort you have put in and the results.
Should I start the war again? Many disregard the stats because of a lack of understanding. Others accept the presented information in faith. Still others look at what is presented and carefully consider its merit.
Ah Stroh, I’ll make a comparison to poker. While you may not play proper pot odds or call when it is statistically worth your money and still even win some money, it doesn’t mean that the statistics don’t exist. Statistics are a part of the world and if you know how to use them, you will have a significant (har har) advantage. The Packers defense was good this year, but could you argue that the defensive line contributed to that?
Since Defense has to be played as a unit, I guess you have to come to the conclusion that the DL contributed don’t you? LB and DB’s can’t play good D w/o the DL can they?
Defenses do have to play as a unit, but some pieces are more important than others. I don’t think anyone would argue that having a terrible quarterback but great linemen, receivers and running backs would lead to much offensive success. Defense is apparently the same, great cornerbacks and linebackers can’t hide poor DL play
I appreciate your work. As a trader, you always are looking for a statistical edge. It may not always turn out the way you like, but it doesn’t mean you stop looking for the edge. It’s only when you are consistently wrong that you question the methodology
Thanks, how bout a stock tip then? 😀
Buy high and sell low.
Er, wait . . . where the hell is that edit button!?
Time of possession tells you plenty about the play of your defensive and offensive lines. The Packers average ToP for 2012 was positive in ToP with 30:26 ToP versus 29:33 for their opponents. This is 50 seconds positive. While this favors GB it is not even close to any kind of dominance by either line. Both the OL and the DL were inconsistent at best. The OL allowed 55 sacks and the running game struggled. The DL rarely created pressure or penetration in the opponents backfield, especially on third downs. OL keeps your team on the field and the DL gets your defense off the field. If not, all the other positions struggle. The SF game exposed these issues for both sides of the ball for GB and the Time of Possesion and final score reflected it. Thanks, Since ’61
I would argue that time of possession isn’t the reason why teams are good or bad, just one result of a team being good or bad. For instance, the Packers often held the time of possession in the wild card game against the Vikings, but I don’t think anyone really thinks that McCarthy’s plan of clock management really was all that effective. As for defensive linemen, I would argue that you need both a good defensive end and a good secondary or else quarterbacks are fast enough now with their releases that even the best linemen can’t get to the quarterback fast enough now.
Very Interesting… I wonder how much drive, focus, discipline, and love of the game come into play. When I watch SF, Ravens, Seahawks, Texans, and the Candinals defenses play, they all play well together as a team. They communicate and tackle as well. The Packers do not seem to have the chemistry or the willingness to put themselve out there 100% of the time.
The things you’ve mentioned are beyond measure. I will say that there are a lot of young players out on the Packers defense that many not have developed drive. focus, discipline or communication. Hopefully with more experience the defense can settle down and start being more consistently productive
In response to Thomas Hobbes, I didn’t say that time of possession made a team good or bad, I said that tine of possession is indicative of the quality of play of both the OL and DL. Since the Packers are 51-21 over the last 4 seasons including the playoffs I would say that the Packers are one of the best teams in the league. Despite that they need to improve both their OL and DL to become a dominant team. Thanks, Since ’61
Very in-depth article.. It does provide some statistical evidence to back up these claims.. I think the Packers really wanted to draft Muhammed Wilkerson in 2011 to replace Jenkins but the Jets got him just before.. There is only like 1 maybe 2 players every year in the draft that could turn out to be a dominant 5 technique.. Sometimes they don’t even make it to the Packers cause they are always picking somewhere in the late 20’s.. Would you fans be ok with them say trading up to the Top 10 to grab someone?? I think it’s worth a shot if you trust your scouting and interviews and grades of the player. The window of opportunity to win with an elite QB is closing, I wouldn’t mind seeing GB be more aggressive in the draft and free agency this year by moving up to select someone they really like.
From an economic standpoint, with the new CBA agreement, trading up into the top 10 of the first round is no longer suicide. But the effect of that is now teams are probably going to be asking for more since high 1st round draft picks are now more valuable since the players drafted are cheaper. Traditionally Thompson has always shied away from trading up, but last year showed that maybe in previous years it was just not worth it but with the new CBA maybe it is. Then really comes down to is there someone who the Packers really think is worth it? My feeling is that if the Packers felt a top 5 player slipped to the 10-12 spot then maybe it would be a option.
Mr. Hobbes… Excellent article!!! Who do YOU think we can draft that will ACTUALLY make a difference? (And ARE they obtainable (draft able by the Packers)?
(e-mail me with your thoughts on this if you can)…
I would ask our dear leader for more information about potential draftee’s. I look more at the statistics rather than the actual player evaluations. I do know that Al has a good rapport with some of the media people; Al was in love with Derek Sherrod and Casey Hayward before they were drafted.