This year has been something of an eye-opener for me. I came into it with the usual assumptions about the nature of a Rex Ryan defense that most people have, that it's a high-pressure, high-risk, live-by-the-sword-die-by-the-sword defense that only had been slightly toned down by the lack of success that 2011 started to bring. I had read about how Pettine was instrumental in calling off the dogs to some degree, and employing more zone in late 2011 and in 2012, but I simply didn't believe it. My eyes kept telling me how aggressive Rex Ryan was, and the reason is - I suspect - that each and every time the Jets blitzed the last few years it was colored by the aura of aggression that 2009 had created, the year that Rex Ryan swept the NFL and most every Jet fan off their feet.
data normalized to 550 pass attempts: showing number of pressures from each blitzing position - a 57% decline since the initial peak of 2009 - one can see the dramatic decline in blitzing from 2010 on.
In several posts I began investigating and detailing how much the Aggression Myth had mislead us. Yes, 2009 and even 2010 were aggressive years - I was initially finding - but since then the Jets blitz rates were in significant decline against NFL averages. Preliminarily it was looking like the Jets for the past 3 years have not been aggressive at all, yet in examining the numbers a big part of the picture was missing. If Rex Ryan indeed invented a super-aggressive blitz concept in 2009, and then gradually has been forced to draw it down to a state of relative passivity - due to NFL adaptation and roster limitations - what was the rest of the NFL doing all this time? What league trends where underway? Did much of the league also have to become less aggressive as high-octane offenses took over? Rex Ryan was not the only innovator in the NFL. There are a lot of good Defenses out there. Where in the spectrum of aggression and success did the Jets fall? And which direction are they heading in.
Building the PRP-MPR Model
My most recent post How Often Did the Jets Blitz: Year End Numbers I presented a new graphic that showed two dimensions of a pass rush in an aim to quantify or at least position pass rush philosophies by team. On the x-axis was Men Per Rush (MPR) which was the average number of men sent on a pass rush down for the year. This number does have some statistical weakness in analogy to the Yard Per Carry rushing stat in that 4th quarter or big lead contexts could distort it to some degree, but it does seem to have a certain kind of statistical ballast, especially when the data sample gets larger. The y-axis showed the team's Pass Rush Productivity. This is a very good Profootballfocus.com individual stat used to evaluate individual performances. I simply calculated it for entire teams to give a perspective on how efficient a team was in rushing a passer, per each player's rush attempt. The formula for PRP is simple: [sacks + [ (hits + hurries) * 0.75] / pass rush attempts. The valuable thing about this stat is that both hits and hurries are incorporated. Also, because it is a rate stat, a team (or player) which rushes the QB less often can still score highly on it. The end result was an aggression by efficiency graph.
The general way to read these Pass Rush Profile graphs are: the more to the right you fall the more aggressive the defensive play call on average tended to be; the higher up the graph, the more efficient the team tended to be. At the extremes: one could be far to the left and not blitz much, but still be high on the graph (like San Fransisco tends to be) because when you blitzed you were efficient; or, you could be very aggressive to the right, but not very successful.
What I really was after was some kind of relationship between these two dimensions. Did aggression produce efficiency? And if so, at what rate? That is where I have to thank tinely24 who first came up with the idea of an average that helped make the distribution clear. The more I worked with each of the years I found this line to be valuable (red, above). In the graphs of each season I include this rate line which works to show, on average, how efficiency was affected by aggression. You want to be above this line whether you are conservative or attacking because it means that your pressures are more successful than average in terms of commitment efficiency.
An additional, very generalized way to read these graphs is to think that if you are successful as a Pass Defense on the left side, you probably are being successful due to coverage. The reason for this is that you are dropping more men into coverage than the average team does. There are exceptions to this like the 48ners who rush few men, but are still successful due to pressure because their interior linemen are good at producing pressure all by themselves, and when they do blitz it comes from a dynamic high-grade pass rusher with high PRP. On the other hand if you are successful and on the right side of the graph you are probably doing it through pressure because you are bringing more men. The exception of course might be the 2009 Jets who not only brought a lot, but because of Revis and others were quite capable of great coverage. These are ballpark areas of success strategy. Unusual talent can allow you to have your cake and eat it too.
Each of the Rex Ryan Seasons
This is where it gets a little exciting. If we graph out each of the Rex Ryan years against the performances of all the other teams we get a very nice snap shot of not only the strategy (in the context of NFL trends) but also of success. And by following the filmstrip of these seasons we can watch how the Rex Ryan defense has morphed since it's spectacular debut in 2009.
legend: bright yellow = top 5 in DVOA pass defense, dark yellow = 6 - 12 in DVOA pass defense, purple = bottom 5 in DVOA pass defense, boxed team name = a winning record. Dotted white line shows trendline without the outlier of JAX.
Immediately one can see how aggressive 2009 was with 5.2 men per rush (MPR). Rex was sending DBs in particular at an incredible rate, producing his organized chaos in overloads. But notice where the Jet fell on the season trendline. They were actually pretty inefficient in producing pressures per man. Philadelphia, for instance, was nearly as aggressive and was far more efficient. In some sense there is an element in the Rex Ryan approach that we want to say defies these kinds of stats, in particular that of efficiency. One of Rex's strengths is play calling, and we want to say that even if Rex isn't efficient on average, tactically when he wants to dial up a blitz on an important down he gets there. But there is also a risk in this. This is the Rex Ryan aura, the idea that he and his scheme is somewhat infallible. What these statistical pictures suggest, rather, is that the Rex Ryan approach has lived, even from the beginning, on a precarious edge. And if you position yourself below average in your ability to generate pressures, even high-risk, timely blitzes may over time backfire. You may be forced to move closer to the mean. And this is exactly what happened in 2010.
Now this is very interesting. The 2010 Jets fell from 1st in pass defense DVOA to 7th, and indeed found themselves positioned closer to their year's average in aggression. But they still remained significantly below league average in efficiency. His tactical blitzes were still coming at a pretty good clip (in fact he may have compensated for the DB blitzes in fewer OLBer rushes) but the pressures were not being generated to even NFL team average. Pulling so many men out of coverage was not paying off, and the overloads were being solved.
Now look what happened in 2011. This is a sea change. Not only were the Jets significantly on the other side of average on the aggression scale, it was the first year they found themselves at least at average on the league trendline (though still below NFL average in terms of PRP efficiency). A far cry from the hyper-aggressive, blitz-happy Rex Ryan media image, the Jets who where 2nd in the league in DVOA (-16%) were now distinctly in the "coverage" camp. Their highly reduced blitz tendencies were paying off at least to average in terms of trendline, and they were getting more men into coverage.
In 2012 the Jets were 10th in pass defense DVOA (-2.6%). They were closer to league average in aggression than in 2011, but still not even an average aggressive team. They fell below the line again as well, with sub-par efficiency in pass rush.
Then this year the Jets were 17th in pass defense DVOA. They we significantly behind and below both the aggression and efficiency averages. Despite a nice, young DL the pass pressures simply were not coming and with coverage issues at nearly every position the defense suffered.
Initial Conclusions and Trends
pass rush efficiency related to league average
The most salient picture to take from this is that the Rex Ryan approach to defense while with the Jets has never been particular good at generated pass pressures. Even in its heyday year of 2009 it was precariously balanced in a statistical deadspot, trading on surprise and tactics to beat the odds. Already in 2010 the draw down was underway, and by 2011 the Jets were firmly in the "coverage" camp in terms of pass defense success, well below average in terms of aggression.
This leads to some very interesting questions in terms of big picture. We often treat the Jets 2009 year as if it is the rule and performances of the last several years as the exception. Instead 2009 was the exception and the trends of the last 4 years are really the rule. If the Rex Ryan blitz onslaught is something of a myth, or at the very least a strategy that worked well for a single year before the league caught on, and coverage skills are really the glue that has held the Rex Ryan pass Defense together, the state of the Jets roster may create something of alarm. The loss of Revis was more than we realized, and only the beginning of what might be called a tremendous talent hole. There are very few - almost no - Jet defenders that are plus coverage guys. Even Allen who is adequate in man coverage is very limited in zone coverage skills. Davis who is known for his speed has yet to show in 2 years that he can cover adequately either in man or zone. Cromartie, the best coverage player by skill, is a strong question mark in many ways. If pressure is not how it gets done the significant weakness in coverage spells major problems in a passing league.
The Last 3 Years - The Rule Rather than the Exception
view an enlargeable version of this graphic
legend: team name black = 2013, team name dark red, 2012, team name pale, 2011 - dotted white line = that year's trendline
In order to pull back the above graphic plots the performance of all the NFL teams from 2011-2013. The trendline is a polynomial expression of the distribution of these the seasons. The 2009 and 2010 Jet seasons have been added for perspective, and all 5 trendlines (2009-2013) have been added so that changes in the data by year can be seen.
Immediately one can see how aberrant 2009 was. The league (and the data) is changing, but statistically it was the worst PRP year of all of Rex's years. One can see the drift over to a conservative play calling approach, and in this 3 year data view two years fell on the trendline, 2011 and 2013.
Importantly, notice the bell curve of the 3-year trendline. We have been representing the aggression - efficiency relationship linearly, but this curve more truly shows that there are valleys of "too conservative" and "too aggressive" that result in inefficiency. And, notably, rare teams push beyond these statistical limits and have success. Houston has been successful on the extreme aggressive end and San Fransisco on the extreme conservative end for instance.
There is a lot in this data and I'd love to hear what other patterns people see in terms of teams and approaches.
A Problem in the Data: PRP
There also seems to be a significant problem in the data. 2009 was a spectacular, if precariously balanced, pass defense year for the Jets, but positioned amongst the data of the last 3 years it shows very poorly. It looks just like an outlier. And also 2013, while a pretty poor Jet year by recent standards, looks surprisingly good. Well, the problem is that PRP rates have changed a great deal between 2009 and 2013. The whole league is getting PRP pressures at a greater rate than ever before. This is the reason I included the trendlines for each of the 5 seasons, as well as average points (red).
What we really want to know is if the NFL itself has become more proficient by scheme improvement or another league trend to account for this, or if the change is in the data collection itself. Looking closely at league totals kept by PPF I noticed two significant jumps in QB hurries, the largest one coming between 2012 and 2013. Below are the total pressures generated by kind, shown as a percentage of all pass rushes in the league.
Notably sacks and hits appear relatively stable, with slight increases in hits. I dug into this by email with Profootball Focus and indeed in 2013 there was a big change in their data collection. That was the first year they had the budget to "double hand" their data. Now they have two, independent examiners who mark up games, and even have a 3rd who arbitrates any discrepancies. As Neil Hornsy pointed out, the 2013 data is for that reason probably more accurate than ever before, but it does make for season to season comparisons a difficulty. This was one reason why each season is graphed independently.
There is a way around this which is to simply treat the anomalous changes as data collection changes - no explanation for the increase in hurries from 2009 and 2010 was found, but perhaps this too seems like a judgement trend due to the stability of the stats otherwise - and to normalize the data averages under the presumption that all teams were judged similarly.
The Correction - All 5 Years
view an enlargeable image of the graphic
All team performances for the years 2009-2013 have been plotted. The teams of the last 3 years (2011-2013) have been labeled, and the two years previous remain unlabeled. A very interesting things stand out. The uniqueness of the year 2009 really becomes apparent. It is positioned with a few outlier type teams: Houston and Arizona. Here it, the year 2011 and 2012 also show themselves to be the only on-average years in terms of pass rushing efficiency by the trendline average, but the 2009 efficiency was generated in a very different way. If one had to characterize the Rex Ryan Jet Defense looking at this 2009 is definitely the exception. 2010 and 2013 also show themselves to be more similar than one would think, both being poor in generated pressures.
In the larger sense the valleys of the bell curve a more shallow (which one one expect in adjusted data) and we benefit from seeing 5 years worth of data in the trendline instead of only 3. The good pass defenses (yellows) tend to appear above the trendline, and the Jet Defenses seem to have found themselves positioned in a minority, at or below the line. As mentioned above, what is concerning is that while there has been a kind of in-Rex-we-trust attitude towards the Defense, if called aggression is not a primary part of what makes the Pass Defense tick the biggest weakness going forward may be roster and skills.
It is also a little odd to see the 2009 season, which in the minds of Jet fans was an unworldly success in terms of generating pressure, actually bettered by the seasons of so many other NFL teams in terms of PRP, at times spectacularly so. It seems almost blasphemous to say that Rex Ryan doesn't know how to rush the passer productively, or at least better than average.
The 3-4 Defense
A perhaps anecdotal but very illustrative example in all this can be seen in the performance of the Jet Interior line. By every report, and by all eye-balls this 2013 version was a young, best-in-the-league type collection of talent. It is just bubbling with athleticism and potential. And for Jet fans the future hope is that upon this strength a great Rex Ryan defense can be built. But just how great was this line? It would be a surprise that the Interior of the line was actually below league average in generating pressures, among 3-4 DLs and DTs. It simply was not spectacular. It was a little below average, and that is generally to be expected from the 3-4 interior. The biggest difference, in perhaps the Jets and let's say the 49ners, is that the Jets have Calvin Pace and the 49ners have Aldon Smith, or the Jets have Harris and the 49ners have Willis, or the Jets have Davis and the 49ners have NaVorro Bowman. There is a huge qualitative difference in skill level from the positions that attack in the 3-4. The 49ners can rely on their 3-4 interior to produce moderate pressure while being pretty conservative in blitz packages, respecting coverage, knowing that when they are bringing pressure it is coming from dynamic, high-level players that are very difficult to stop. If indeed the Jets are banking on a Rex Ryan defense that will trade on the San Fransisco side of the conservative - aggression scale the Jets seem to be missing the most important parts of the scheme. They have the line, but they don't have coverage guys, and they certainly don't have dynamic, explosive pass rushers to make it count when they do rush. It is a perhaps little like saying you are going to run the Triangle Offense in basketball, but you are missing your Micheal Jordan or Kobe Bryant.
For those waiting on a team built around a great Rex Ryan defense it would seem that at least two years needs to be granted before the skill level of the roster on the Defensive side of the ball even remotely is comparable to the 49ners. And there are even some questions about current player investments like those of Davis (who is nether a fierce pass rusher or a coverage LBers) and Coples who at best can be said to struggle at the Outside Linebacker position. Not only do the Jets appear to need significant upgrades on Defense to make a Rex Ryan approach work, with some of these being the hardest and most expensive positions to fill (pass rusher, shutdown corner), the biggest roster priority would appear to be on the other side of the ball. In a passing league you need to be able to pass. This is going to take multiple years.
Of course all is not lost. The team will be improved next year, and the next. The real issues are perhaps about seeing where Rex Ryan has been taking the Defense, and where that approach stands in the context of the league. This is a coverage Defense that seeks to use punctuated blitzes that accentuate Ryan's playcalling finesse, but a defense that requires big time athletes to make it all go. It is no surprise that Ryan had lost of success with the talent laden Ravens' roster with Hall of Famers at ever level of the Defense. Once the potency of the overload DB blitz was lost (sometime around 2010) the need for skill players on Defense has become more and more pronounced, and the talent drain has continued with the aging of Rex's favorites, and perhaps with some questionable young replacement choices. If the bet is that Rex Ryan will do great with multiple Hall of Fame players that is a long bet, it makes the draft and the cap places where the Jets simply cannot miss very often.
the data: there a huge amount of data collection done here with lots of crunching, so not only might there be errors there probably are errors. I produced the data pictures for my own pleasure seeking the patterns that expose the dynamics of the NFL pass rush. I hope you enjoy them as much as I did. Of course these are novel data presentations and they could certainly be improved. They are only presented here as food for thought.