Traditional thought has always said that the lynchpin of a successful 3-4 defense like the Packers run is a giant nose tackle who can contain the pocket against a double team and a versatile outside linebacker who can rush the passer. Ted Thompson definitely followed that philosophy to a T the very first year that the Packers implemented the 3-4 defense by drafting a nose tackle #9 overall with BJ Raji and then following that by trading up to select Clay Matthews III. However since then Thompson has been rather apathetic when it comes to nose tackles; since Raji the Packers haven’t really drafted a player who fits the mold of a 3-4 NT (Boyd and Thornton might be the closest but both have played more DE than NT so far). At this point, with the prevalence of the passing game and the subsequent shift to more nickel and dime looks that feature only 2 down linemen, nose tackle has become a rotational player more than a foundation of a defense.
With the Packers on the clock at 30 in draft, plenty of people were clamoring for an inside linebacker (myself included) but a sizable minority of fans also wanted the Packers to take a nose tackle like Jordan Phillips or Malcolm Brown but the Packers passed on both positions in favor of cornerback Damarious Randall. Much like inside linebacker, Thompson doesn’t appear to value nose tackle that much. Was Thompson right to ignore the nose tackle in favor of a cornerback, where the team definitely has more depth? Or to put it another way, how much effect does a nose tackle have on the outcome of a game? Is he really the key to the defense as has been said in the pass?
To test this question, pun/pass/total grades were pulled of nose tackles playing 3-4 defenses for every regular season game from the 2014-2015 season from pro Football focus and tabulated with game data from Pro Football reference. New England (who plays a 3-4 on paper) was excluded from the analysis since they run so much 4-3 defense and really are a hybrid defense at this point.
The easiest and most straight forward way to see just how important nose tackles are is to compare how well they play versus whether or not the team won the game. Unfortunately, there is no statistical difference between PFF grades and whether or not their team won the game. In other words, a nose tackle was just as likely to play well and win the game as he was to play well and lose the game. Obviously wins and losses are way more complicated than just how the nose tackle plays; a nose tackle could play spectacularly only to have the offense sink the game or perhaps the defensive secondary couldn’t stop any receivers.
While it doesn’t appear as if nose tackles directly control the outcome of a game, it is possible that good nose tackle play could limit the number of points that an opposing offense can score, which would factor out how well his defense plays. This graph looks at the correlation between the PFF scores that each nose tackle received and the number of points the defense gave up in that game. There does appear to be a statistically significant negative correlation between nose tackle play and score, meaning the better a nose tackle played, the less points his team gave up. However, only 14% of the variation in score can be accounted by nose tackle play, so while it is a significant result, the expected outcome is rather small.
Nose tackle is supposed to be the anchor in run defense, but ironically there isn’t any significance between how well a nose tackle defends the run versus how many points his team give up. It is possible that the prevalence of throwing the ball skews this correlation, but I decided not to split points into pass and run since I didn’t want to overemphasize drives where the quarterback throws the ball down the field and the running back punches it in from 2 yards out.
Interestingly, good play from nose tackle defending the pass does seem to affect the ability of opposing teams to score points. Again while it’s significant, only 15% of the variation in points can be linked to nose tackle passing defense, which again would seem to be quite minor but what’s also interesting is that you don’t see much variation in PFF pass defense scores, especially compared to run defense scores. While this might be an artifact of PFF’s grading system, it would also point to the fact that nose tackles don’t contribute much to pass defense (which makes sense intuitively) one way or another, which of course would agree with the fact that importance of nose tackles on pass defense is quite small. You can interpret the data as you see fit, but personally I would say that even if nose tackle pass defense actually influenced how many points were scored, the effect would be very minor.
In conclusion, nose tackle play seems to have at best a very minor role and at very worst is not correlated with winning or losing games. I will say that context is important and having r values for quarterbacks and punters would be useful in really framing how important 15% correlation is (if anyone else wants to take a stab at this you have my gratitude). Furthermore, nose tackle play is notoriously difficult to grade, BJ Raji has been quoted that it’s pretty rare for a nose tackle to run a “jet” (i.e. rush the passer, which would be easier to grade) play and a nose tackle could appear to be blocked by an offensive linemen when in reality he’s actually walling off the offensive linemen and allowing other players to fill in the gaps. Overall, I would say that this probably corroborates with my previous analysis that shows that nose tackles aren’t all that important. I’d say on the importance scale of quarterback to long snapper, nose tackle is probably right next to inside linebacker as one of the less important positions on the team.——————
Thomas Hobbes is a staff writer for Jersey Al’s AllGreenBayPackers.com.
49 thoughts on “Big Man, Small Effect: How Important is a Nose Tackle?”
I am a firm believer that a defense has to be strong up the middle know matter what the formation.
I do not know exactly where I heard that, could be from the great one himself (Lumbardi).
I think with Raji/Gruin, Matthews/Barrington and HaHa we look pretty good up the middle.
Do not under estimate the second coming of Raji.
in Ted we trust.
My rebuttal the article;
“Nose tackle is important and I think we are going to be O.K. with what we have”.
I’m still not convinced that nose tackle is all that important; keep in mind another option to your rebuttal could be “Nose tackle is NOT important, but I think we are going to be OK with what we have”.
I don’t have numbers or stats to back this up but if you do not have a good nose tackle you better have a damn good offense.
If raji is our NT then hopefully it is not an important position, and in fact, like ILB is going to disappear completely as it almost vanished the last half of last year
I would have been very interested in seeing what Raji could have done last season; from what I recall, he was a lot better at NT than he was at DE and in 2013 he played mostly DE. With the Packers a lot more stocked at DE than before Raji will probably play exclusively NT and that could mean he performs better.
Football is the ultimate team game. Everybody needs to beat the man in front of him in order to win.
Data can be analyzed and studies can be done, but if I had to guess, QB is the only statistically significant position you’ll find.
This is quite possible, I did mention that it would be nice to see the r values for all positions, but frankly I don’t have the time to do all the analysis. However, even if the data is noisy, I still think that some information can be gleaned.
Not only is it quite possible. Its very likely. QB is really the one position where performance directly relates to winning.
NT have lost some importance, even since Raji was drafted, but the premise that a strong and productive NT still makes all the sense in the world. It helps free the ILB to make tackles on run downs.
A better study might be how NT play correlated to wins when the opponent runs the ball 50% of downs. Against the pass a NT isn’t nearly as likely to make a big impact, but against the run I bet it does.
what actually be more interesting would be seeing the correlation on running/passing plays when the NT is on the field versus when he is off. Common sense would say that having an NT on the field against the run would be a positive, I wonder if that would be offset by having a NT on a passing play since a lot of times defenses don’t know whether it will be a pass or a run.
As much info that stat sheets and those that produce them can offer,I’m always left thinking and agreeing with what Rodney Dangerfield offered to the Business professor in the movie Back to School…..”You left out a whole bunch of stuff.” 🙂
My feeling is stats can answer every question; whether or not the author has done a correct and proper job is another matter. I’m quite positive I haven’t done enough of a through job and there are variables that I am aware that I didn’t cover. However, rather than endlessly chase that by myself, I’m showing everyone to see if anyone has anything to add.
Obviously nose tackle and ILB are not important… we haven’t had either in 5 years. When was the last superbowl? oh, I get it…
Desmond Bishop and AJ Hawk were the two inside linebackers for SB45; was Bishop good enough to carry Hawk? Was Hawk good enough in 2010 than 2014? (-6.0 vs. -14.4 PFF). I don’t know. However, if Thompson’s drafting history is any indication of how he views importance on the team, then neither NT nor ILB is all that important to him.
Desmond Bishop was a monster during that run we made to get to the superbowl. He was a ruthless tackle machine and Raji held the point of attack. We also had Cullen Jenkins and Ryan Pickett/Howard Green. They were all monstrous defensive lineman. We had a top 10 defense with all of those players in place and that’s without having a reliable suitor to pair with Clay Matthews. There was a time where our run defense was #1 in the league and even A.P. couldn’t run on us. And those were the times with Nick Barnett and company in the middle. Don’t underestimate having a great front 3.
It’s interesting that the Packers have actively moved away from the “big dudes up front” to a smaller and quicker defensive linemen who can penetrate into the backfield (say Pickett vs. Daniels).
I believe the drafting history stuff is largely nonsense since it rarely gets put into any context, but if something gets repeated often enough, folks start to think it (the suggestion that TT devalues the position) is true. Watch the draft, the $, the FA signings, and the personnel on the team.
Pickett replaced Grady Jackson, who was pretty good. TT paid good money for Pickett, who led the NFL in tackles in 2005 and was a coveted FA in 2006. See link below. Then TT put the Franchise tag on Pickett in 2009, and resigned him for 4 yrs @ $24.75 million. One can watch the draft history, sure, but also watch the $. TT then spent the #9 overall pick on another NT, Raji. Sounds like TT spent a lot of capital on NT during his tenure!
LB is trickier. Despite the -6.0 PFF grade for 2010, TT signed Hawk for 5 yrs/$33.75 million. Seems clear to me that TT graded or at least saw Hawk differently than PFF. Bishop was playing well, and TT had CM3. Before that, TT inherited Nail Diggs and Barnett, a 1st rounder in 2003, and promptly took Hawk using the #5 overall pick and Hodge at #67. Lots of capital. GB only had 3 starting LBs and two were 1st rd picks. TT took Poppinga #125, then Bishop #193, CM3 and Brad Jones 7th rd., then DJ Smith 6th, then Perry and Manning in 2012 and paid elite money for Peppers in 2014. Still, a lot of those LBs were OLBs. The context might be that TT was focusing on fixing the hole at OLB in lieu of spending capital on ILB.
TT has never used a 1st rd. draft pick on a CB. Oh wait, he just did, because for the first time he needed a CB! I was always sure that TT hadn’t devalued the CB position since GB has had good CB play for most of TT’s tenure. TT inherited Al Harris and Ahmad Carroll, who was bad. 2006, TT buys Charles Woodson (& Manuel). Tramon started to play well by the end of 2009 and Shields came out of no where in 2010. TT did not NEED any CBs, just some for depth.
An interesting thought is which position has the worst/best bust/disappointment % for 1st round draft picks? I’d guess worst would be QB, DL, CB, and LT, but I’ve done no research.
“Favre is watching the Packers’ free-agent acquisitions to see what kind
of team they will have. If he likes what he sees, he is expected to
postpone thoughts of retiring for another year.” (March 2006 – link below). I’d heard this kind of stuff from Farve before then, who was getting old, and had been beaten up. GB also knew about his drinking, parties, pills, maybe girls, and his Diva behavior. I just read that Rodgers was the only player on GB’s board with a 1st rd. grade when they drafted at #24. I thought at the time that drafting Aaron Rodgers was a no-brainer, and still do.
BTW, I think Malcom Brown is a 3-4 DE. He isn’t just a NT. We’ll see what NE does with him & how he does.
Finally, I’m not doing hours of research regarding other positions so I can post it at the end of a thread again. Been there, done that, but thanks for your obvious effort in doing the research for this article.
The caveat with the Pickett signing is that I’m guessing Thompson didn’t envision him being a 3-4 NT since they were playing the 4-3 at that point; is it possible that Thompson was considering switching that far ahead? Possible, but that’s a pretty big stretch. DTs in a 4-3 defense are paid considerably more, so it’s debatable if Thompson would have done the same thing if it were a 3-4 defense.
Same goes for Hawk, obviously everyone outside the Packers front office is still confused as to why Hawk was resigned for decent money, but when he was drafted, Hawk was definitely a 4-3 OLB. Evaluating Barnett is tricker since he was a hold over from Mike Sherman, would Thompson had valued ILB more if he didn’t have Barnett? Furthermore, you can’t say either that Thompson would have spent a 1st on Barnett either.
I agree with you when it comes to CB, but a lot of that probably has to do with the fact that he won the lottery twice with Williams and Shields, how many teams can say both of their starting cornerbacks were UFAs? I will say that he’s been putting more stock into the defensive backfield as of late; Hayward was a 2nd, Hyde was a 4th, House was a 4th, Burnett was a 2nd and Ha Ha was a 1st (alongside Randall and Rollins)
As far as Hawk’s value to the defense, it will always boil down to this- people who don’t understand the role of the jack backer in the 3-4 defense will always say Hawk was horrible and got stuck on blocks and didn’t make tackles; people who understand the role of the jack backer in a 3-4 will see that AJ Hawk had a role of sacrifice that many players would complain about, but he did it and never complained about it, and never once said “I did my job, but I can’t speak for the guy who is supposed to make the tackle.”
I really don’t care to be ridiculed as I have in the past for stating what the Jack ILB’s role is, so any interested can do their own research and come to their own conclusions. I will say this, once somebody understands what AJ Hawk’s job duties in the 3-4 was, they can go back and watch him play in an entirely different light.
While I don’t think many fans will notice him eating up a blocker or rushing up the center so that someone else can slip by, fans will notice the missed tackles, the poor coverage and the times where he looked a step slow. Is that fair? I think fans keep in mind he was the 5th overall pick and adjust their expectations accordingly; I’ve mentioned before if he was a 3rd rounder I think everyone would think a lot differently.
All too often, it was “Hawk is always making tackles from behind, after the RB already gained 5-7 yards”, which many times was because Hawk was doing his job, the other ILB missed his tackle, and hawk had to disengage and run after the loose ball carrier. I fully agree that Hawk never lived up to the kind of game changer fans expect from a #5 overall pick (and they would have been much happier with him had he been a 3rd rounder); however, many top ten picks are busts and wash out, at least Hawk was a solid contributor for a decade.
The ultimate reality that most fans don’t see is that it really doesn’t matter what round a player is drafted. Of course you you want your highest pick to pan out, but what most fans miss is that you want ALL of your picks to pan out. With rookie contracts being restructured the way they have been, fans should have an even easier time getting over draft slotting expectations, since the money has been reigned in substantially.
Money is the only thing that matters, not draft slotting. If the Packers spend roughly the same amount of cap on each draft- it doesn’t matter which picks pan out and which don’t- so long as we get a haul of quality players and our financials are in order (in the big picture). Poor AJ Hawk. It’s not his fault he was drafted #5 overall.
Burnett was a high 3rd rounder (#71). All your points have validity. I knew LB was dicey. Seems to me that things are usually more complex than they seem. Dom Capers was hired 1/19/09, so I’d have to admit you’re right and Pickett was signed as a DT and not as a NT. There is a good chance that we will get an answer or at least another clue to the issue of how much TT values NTs next off-season.
TT doesn’t value NT and ILBs. No wonder we can’t stop the run.
Do you value Edith?
You can’t have it all in and at least personally I would rather see the team do well against the pass than against the run.
It shouldn’t be an either/or situation. There are plenty of teams that defend both adequately.
Sure, but then other aspects of the team get minimized, there’s simply no way to field a completely balanced team unless you get incredibly lucky in the draft repeatedly.
Team speed is a great equalizer.
Only if used properly, I think Al Davis pretty much dispelled the “speed is king” theory during his last years. Speed is good, but you can’t sacrifice everything else to have speed.
Dom’s scheme is intended to force the opponent into passing by denying them the run. That’s the situation they intent to create, it’s the situation where they feel the defense has the biggest advantage. Why? Because even the best QB’s in the league fail to get a single yard on a passing attempt 35% of the time. Because the biggest opportunity for a turn over is when the ball is in the air. Because the best opportunity for a defense to back up an offense with negative yardage is when the QB drops back to pass.
In short, the idea is you want to defend the run well because you want the opponent to choose to pass.
I wonder if that’s the case anymore, considering how many defensive backs Capers likes to used, followed by the general trend of getting smaller and faster defensive linemen, I almost think his current strategy is to force the opponent to run by denying them the pass. Overall, I think the pass is definitely more dangerous than the run.
Let me be clear on this, the game is changing.
I am talking expressly about Dom’s 3-4 scheme(we are talking about the NT position, after all), which admittedly, sees less time on the field than Dom would probably like. Some of that is attributable to our available personnel, some to changing trends in offense, and some probably due to the HC making the decision to move in a different direction.
We play in nickle the majority of the time, with either 2 or 4 down linemen (Last season, we saw more 4 down looks with 3 linemen and a OLB/DE tweener as the 4th down ‘elephant’). What all this means is, we’re not seeing much true 3-4 anymore, and that means we’re seeing less true “0 tech” NT play, or even shading to 1 or 2 from the NT. In that regards, you could say that in GB, NT isn’t that important, because we play less and less 3-4. However, I believe fundamentally that the NT is the base building block for a sound 3-4 defense, and I believe most defensive coaches would agree with that statement- you build inside out, the scheme relies on it. It’s a big part of the reason the D started performing better towards the second half of last season- Guion coming on, and improved play at ILB with the move of CMIII. Inside out.
Your argument is becoming a double edged sword; do the Packers run a fundamentally sound 3-4 defense with a NT or do they play the opponent and field whoever gives them the best chance to win? I would argue if you have a truly dominant defense, you can do whatever the hell you like; if the Packers had super stars on the defensive line and linebackers, then you could probably have a NT on passing downs and still be successful. It’s kind of how the Packers offense is built, they are too talented to really have to drastically adjust to the opponent. Of course the Packers defense isn’t dominant so it makes more sense to go with what you have and make life as difficult for the opposing defense and let your offense win the game (I have no problem with this philosophy). In that case NT isn’t as important since sub packages are more common.
Respectfully, Thomas, what is quickly becoming the double edged sword is your focus. Your article clearly states in the first sentence that you are talking about the value of the NT to a 3-4 defense, which therefore means suggesting the “NT isn’t as important since sub packages (alignments and sets which are not our base 3-4 defense) are more common” is an argument outside of the stated scope of your article- although one I would agree with as it pertains the Packers’ defense as a whole.
As it were, however, when we are talking 3-4…the NT is, IMO, still as important as ever.
FWIW, I mean none of this in a combative tone. We simply disagree. 🙂
You’re a great writer and I always look forward to reading your articles.
I’m not specifically referring to the Packers 3-4 base defense when saying 3-4, I’m mostly using the term to differentiate it from 4-3 defense in terms of personel. While I don’t have proof I’m almost certain that a sizable proportion of the data I used comes from 3-4 teams that weren’t running a base 3-4-5 alignment. There will situations where a linebacker is subbed out instead of a linemen and I seem to recall one time where Capers played the psycho package with Raji playing 0 tech.
Mark Twain once said, “Facts are stubborn things, but statistics are pliable.” As you wrote, “NT play is notoriously difficult to grade.” A good NT doesn’t get the credit he deserves. We only note the absence of a good one when we see the fact of the opposing team continually running it up the gut. Good article, though, Mr. Hobbes.
The better Mark Twain quote is there are “lies, damn lies and statistics” Unfortunately, unless you actually have the play assignments, which only the Packers have, you will never know how any player performs. Assuming that no one will ever get that specific information, you have to go with incomplete data. Does this kill the analysis? I would think that major wins and losses would still be gradable (like collapsing the pocket or falling down) so there is still some validity.
Thomas: Thanks for the reply. FYI, your quote of Mark Twain about lies and statistics comes from Twain’s autobiography in which he credits PM Benjamin Disraeli of Great Britain for saying it when he wrote, “Figures often beguile me, particularly when I have the arranging of them myself; in which case the remark attributed to Disraeli would often apply with justice and force: “There are three kinds of lies: lies, damned lies and statistics.”
The problem with your assessment is that the value of the NT in a 3-4 isn’t measurable in tackles, or passes deflected, or sacks.
The NT is supposed to be such an immovable force at the point of attack that he takes the full attention of two offensive linemen (creating a potential opening in either the 1 or 3 gap for an ILB to fill or giving a pure one-on-one for the OLB on the edge), holds his ground and clogs up the middle of the line so that the ball carrier must divert east-west (a win for the defense since time and yards are being used laterally instead on vertically, and theoretically funneling the ball carrier into the arms of the OLBs or DEs). Yes, the NT needs to be a great two gapper as well, meaning, he has to be able to get off the double and make the tackle if the ball carrier actually decides to attempt the burst up the middle. However, a sound NT means the offense quits trying that because there isn’t anywhere to go.
I admire the research you’ve put into your ongoing assessment of the NT in the 3-4; however, I think it is flawed because you seem to be looking for statistical proof in the form of tackles and such, where as the position is lynchpin because the NT is truly supposed to divert the direction of plays away from his position.
Give that man a thumbs up. You get “it”. Not so sure the author w/ all his stats does!
The analysis is premised on the notion that PFF knows what it is doing (which I take leave to doubt) when it assigns grades, and takes into account the difficulties in assessing NTs. PFF is supposed to note whether the NT did his job whether the play comes near him or not. However, PFF is one of the few sources available, since we don’t have access to the grades given by the Packers’ coaches, which probably are more accurate, so if one is going to try to write an article that is not purely subjective, such data is all that there is.
It looks to me like Oppy really is suggesting that PFF can’t properly grade NTs, so using its data in a statistical analysis is doomed to failure. Fair enough, so I upvoted Oppy. But the author did note such difficulties: “Furthermore, nose tackle play is notoriously difficult to grade…and a nose tackle could appear to be blocked by an offensive linemen
when in reality he’s actually walling off the offensive linemen and
allowing other players to fill in the gaps.” I think the author does “get it” and properly acknowledged the limitations in the data and the need to analyze more positions to see as a check on his methodology (eg., to see if using the same analysis whether any defensive positions show significance).
I think your argument is valid, but it depends on how much you trust PFF. Admittedly, they don’t actually post their grading rubric, but it is supposed be based on technique, so it should account for holding up double teams, being able to 2-gap and making sure LBs can get off clean. You would think that over every NT play over an entire season there would be some trend one way or another.
1st, a question. There seems to be a lot of dots. Did you plot the grades for every NT? Only about half of the teams use the 3-4, and there is only one main starting NT. For instance, Wilfork played 802 snaps. I’ve no interest in the results of his back up, who presumably only played a handful of snaps and whose dot merely skews the data. I wonder if your analysis would be different if only starting NTs (say, minimum of 350 snaps) were considered.
I read your 1/28/13 article. It suggested that safety and DE were the most important positions. I’m not clear though: was that using data for both DEs and both Safeties? I mean, Daniels by PFF is about a top five 3-4 DE. Boyd, er, wasn’t.
Only teams that play 3-4 were included and every dot is a NT for one game. Also I didn’t use NE since they are a hybrid defense at this point and I didn’t want to skew the data. Finally, I tried to use only the starting NT where possible, but sometimes this wasn’t clear in the data so I made an educated guess.
I think the fans clamoring for Malcom Brown or Phillips or Goldman at #30 was a function of two things:
1. perceived player value based on lists generated by draft evaluators.
2. low-impact play by the DL over recent years.
In the end, the line must play better. NT is an easy position to point toward, but I would argue that DE play out of the Packer DL–especially in odd-man fronts–has been more troublesome over recent years.
Your first reason is not exactly complimentary to those of us who clamored for Brown. The second reason seems reasonable. As to the paragraph, I view Brown as a 3-4 DE/NT. By PFF scores, in 2014 Daniels was an elite (top 5 or so) 3-4 DE overall, and had impressive grades in both pass rush and run defense. Jones ranked 17th as a pass rusher (his grades were pretty good – that is, closer to the 10 best guy than to the 27th ranked player), and Boyd was not listed in the top 34 3-4 DEs for pass rushing. The article I read only showed the top 19 run defending 3-4 DEs, and neither Jones nor Boyd were listed.
“Your first reason is not exactly complimentary to those of us who clamored for Brown”
This was me, too. I was absolutely flabbergasted (and more than a little flabby) when the pick wasn’t one of those DTs at that point, and many of us were pulling our hair out as the draft progressed and some of these guys (Carl Davis?) kept sliding and sliding.
I’d like to see what Daniels’s snap count was in base over the last two years…and maybe what the total snap count by the different DE candidates in base has been. I don’t know how PFF generates those grades and I only put a little stock in their results (any algorithm has inherent bias), but my guess is that Daniels plays a lot more snaps in sub-packages that have what amounts to even fronts. Whatever the case, Daniels is a beast…I’d take 4 more of him. Boyd and Jones? Meh.
As far as I know PFF actually hand grades every player, and then only normalizes that grade to the “average player”. I also assume they normalize to even out the graders themselves. So there’s isn’t really an algorithm like DVOA.
There has to be some kind of algorithm. You can’t have a grade without a scale and you can’t have a scale without a norm and you can’t have a norm without an expectation of performance.
How do you establish the expectation?
There has to be an algorithm.
If you want to go straight by the definition of algorithm then yes PFF uses one, but in the end they are normalizing a human observation as opposed to using game data to generate their statistics. Yes, every algorithm will have bias, but so will every other observation you can make. On top of that, as long as the bias is all in the same direction it doesn’t matter because this analysis is comparing players based on the same algorithm, so any bias will cancel out.
Comments are closed.