BGA Weekly – Should Rex Substitute Less?

Bent ,

Now that the season is over, I’ll be writing weekly analysis articles from now until the beginning of the league year and also during the period between the draft and training camp. I’ll be breaking down some of the data from the 2012 season and revisiting some of the things I wrote about over the last two offseasons to see if any patterns identified at the time have continued or if any new patterns have developed.

In this week’s BGA, I’m going to be looking at how the Jets rotate players and consider some of the implications of how often they substitute. As ever, thanks to PFF for giving us access to data and information not available to the general public.

After the jump, an analysis of Jets’ substitution patterns together with a look around the league at some other teams and how what they do compares with the Jets.


I want to start by recapping perhaps the craziest sequence I’ve seen since I’ve been recapping Jets games. This happened in a game late in the 2011 season.

The Bills had first and ten down at the Jets’ 15 yard line as the Jets lined up in a 3-4 front. Ryan Fitzpatrick handed the ball off to CJ Spiller who ran for seven yards to the outside, making it second and three. Would the Jets consider this a short yardage down, or did they need more versatile personnel on the field because the Bills had the option to throw and still end up with third and short if the pass fell incomplete? A lot would depend on the Bills’ offensive personnel and how the Jets would decide to match up with it.

As the play ended, Eric Smith sprinted onto the field. A few seconds later, Kyle Wilson started to run on, turned to get some instructions from the sideline and then continued on towards the defensive huddle. Garrett McIntyre, who was on the floor having been blocked to the ground saw Wilson running on and ran off to the sideline. A slot cornerback replacing an outside linebacker is the obvious move if you’re going into a nickel package, which made sense as the Bills ran an extra wide receiver onto the field. However, after a few more seconds, the Jets’ sideline changed their mind and decided to put their specialist pass defense personnel onto the field. Jamaal Westerman, Josh Mauga and Brodney Pool all ran onto the field and into the defensive huddle, which now had far too many people in it. Smith, Bart Scott and Sione Po’uha saw those three entering the game and then left the huddle and started running towards the sideline. Halfway there, the coaches motioned to them to go back and Scott and Po’uha stopped and ran back to the huddle although Smith did not. When Scott and Po’uha made it back to the huddle, Westerman and Mauga went back to the sideline.

Finally, the defense was set and, to their credit, they managed to stop the next play short of a first down despite all the pre-snap chaos. Buffalo would end up scoring a touchdown on the drive, but the Jets ultimately won the game 28-24.

If you watched the condensed game version of the re-run – or the coaches film – you don’t get all that stuff between plays, so you’d have seen the Jets transition from a 3-4 front into a 3-3-5 nickel package. Essentially, the only personnel change they made was that Wilson came in for McIntyre so they could more readily match up with the Bills in their three-wide set. All that other stuff was unnecessary. There was one other change, which serves to emphasize one of the main drawbacks of having such a complex substitution policy: On first down, there was no deep safety so the Jets were playing with 10 men. Although Smith sprinted on after the play, suggesting it was his fault, it was actually Pool who ended up as the 11th man on that second down play. This might have been because the Jets benched Smith for missing his cue.

In the Jets’ defense, this was kind of a unique set of circumstances for this game. First of all, the Bills are a team that has abnormal versatility in terms of personnel (ie Brad Smith can enter the game and you don’t know whether it’s to be a third receiver or to take a snap and Spiller regularly splits out wide or in the slot). Perhaps more importantly, the Jets, who would usually have two safeties playing all the snaps and rotate a third in every now and then, had announced that week that they were going to reduce Eric Smith’s and increase Pool’s playing time, something they did by rotating them in and out on a series-by-series basis. This game was also on a long week, so maybe they felt they would have enough preparation time to put a plan in place to match up specifically with the Bills’ personnel.

Although they won the game, the defensive gameplan didn’t exactly pay off, as the Bills racked up 336 yards and almost scored a game winning touchdown on the last drive, highlighted by two near-catastrophic blown coverages by Pool. This begs the question: Are they over-complicating things and would they be better served by removing some of the complexity and confusion, even if it means that from time to time they might end up with someone in a role they aren’t ideally suited to (ie a linebacker covering in the slot)? Or did they have so little faith in the versatility of their personnel that they felt that would be too easily exploited?

Obviously, this is an extreme example, because I’ve already said it was the craziest sequence I’ve ever noticed. However, an extreme example like this does underline the issues with it. Let’s consider some of the pros and cons of such a policy in more detail and look at what happens with other teams around the league.

The Basics

The reasoning behind this is something Rex Ryan has talked about in the past. If you can play, this defense is versatile enough that he can find a role for you. However, he has also said that he wants the best 11 players on the field. Based on their substitution patterns, who the best 11 players are might change depending on what he anticipates the offense is going to do. If the other team is expected to pass, he wants personnel out there that can rush the passer or handle coverage assignments. On the other hand, if they’re expecting a run, you don’t want your pass defense in the game, otherwise teams will overpower you in the running game. If you have a player that excels at one area, but is weak in another, then you can try to maximize their production by restricting the amount of time they are put into situations that expose their weaknesses. It follows that if you have someone who is versatile enough to do either role, but perhaps does not excel at anything, simply playing that guy all the time would mean you could expect a lower standard of play than if you used a guy who excels against the pass for passing downs and a guy who excels against the run for running downs.

However, therein lies the problem. Since you don’t know when a team is running and when a team is passing, you need that versatility on the field, otherwise teams are going to recognize and exploit where your weaknesses are. The Colts ran the ball down the Jets’ throats on third and long in the 2010 wild card game because they were able to blow the Jets’ six or seven DB sets off the ball, picking up a couple of key first downs. The Patriots have always looked to isolate Jets’ personnel, sometimes switching to the no-huddle so they can dictate which players remain on the field defensively and tweak their offensive approach accordingly. That would also lead to players being on the field for longer and in situations they perhaps aren’t as prepared for as they might be. With the improving versatility of many offenses, that’s something you need to be able to counter these days.

The other problem, of course, is the confusion that is sometimes caused. Whether it leads to blown coverages, the wrong number of players on the field or whatever, if this is something that could be eradicated by having a more straightforward rotation then the team should consider if it’s the best way forward or if these downsides outweigh any benefits.

The advantages are not just limited to putting your personnel into a position to succeed. There is also something to be said for the unpredictable nature of the defense Rex puts out there. Often mixing up formations, packages and coverages can cause a team to rethink its offensive gameplan. The Jets are trying to cause confusion too. Again they need to weigh up how easily teams can overcome this and whether it does more harm than good to the defense.

Does it matter?

It’s an interesting question to ask 1,500 words into a study into this topic, but could I be overstating the negative points of this system? Yes, the Jets looked confused on that play, but they got themselves lined up on time and managed to prevent a first down. Sure, they only had 10 men on the field on the previous play, but it was the only time they made that mistake all day and it only cost them seven yards, which Spiller probably would have still gained even if Smith was back in centerfield. A few plays later, the Jets lined up with no drama and the Bills scored a touchdown anyway. While there’s certainly plenty of examples of things like blown coverages I could give you that were perhaps caused by confusion or a communications breakdown in the defense, who says that was because of the complexity of the defensive rotations? Maybe someone just had a brain cramp for some other reason each time.

I know a lot of people have questioned the complexity of the offense in recent years and the expectation was that things would be simplified this year and the offensive personnel would benefit. Obviously, it doesn’t work that way. In the end, I’d say it came down to personnel on the offensive side of the ball and perhaps that’s the case defensively too.

Needless to say, even with things the way they are, the Jets defense has been highly ranked over the past four years. Maybe if things were simplified they would have been even better. Then again, perhaps this was a product of the fact Rex didn’t have the requisite personnel on his defense and this was his best chance of having success. Now that he’s into his 5th season, you’d hope that maybe he would be building a core of players more prepared for a full-time role (or close to it) and this would allow him to simplify things.

Full time role

Ryan would likely feel more comfortable if he had more players that he knew could excel in both disciplines. Muhammad Wilkerson is certainly developing into someone who meets this criteria. In 2012, he played over 75% of the snaps in every single game and even played every snap in two late-season games. Maybe if Quinton Coples could stay on the same development curve, he could become one of those core pieces Ryan’s defense needs. Within two years, perhaps guys like Demario Davis and Antonio Allen could be on the same curve.

As an indication of how much substituting teams do, we can look at how many times a player played every snap in a game. For the Jets, that happened 42 times in the 2012 regular season. If you compare that with Pittsburgh, that was a team where it happened a lot more. They had 71 players with a 100% snap count.

The 49ers are an interesting case. They had ten players with over 1,000 snaps this year (compared with just five for the Jets). Essentially their entire starting lineup were full time players, although they’d rotate out a lineman for a slot cornerback on passing downs. Having said that, they only had 40 total games with a 100% snap count. Their number was skewed somewhat by the fact they had some blowout wins where they rested their starters in garbage time (and they also rested their starters in the last two games with their playoff place secured). Even so, they had 75 instances of a player with a 97% or higher snap count – 14 more than the Jets.

Houston is another top team that has a relatively straightforward substitution pattern on defense. However, they had four fewer 100% snap count games (and 10 fewer at 97% or more) than the Jets. This is because although their substitution patterns are straightforward in terms of positions and formations, they do rotate their personnel more than many other teams.

New England is another team that, like the Jets, has some complexity to its substitution patterns. Their number for 100% snap count games is just 30, although this is again skewed by the fact that a lot of their games were blowouts. (Note: Ryan tends not to rest starters in blowout losses). The Patriots are, of course, notoriously well-drilled.

Jets in 2012

This would all be pretty meaningless if the Jets had already simplified their substitution patterns, but that doesn’t seem to be the case (and who can predict which direction Mike Pettine’s departure will send things in?)

The thing about the substitution pattern, is that it isn’t a pattern. Sometimes the Jets will match up with the offensive personnel, sometimes they’ll line up based on the down and distance. On third and long, they might substitute in a completely different personnel package than they did in the same situation on the previous series. Some other teams substitute a lot less, but there are also those that – although they might make as many changes – substitute in a uniform manner with maybe just three or four different personnel groups that enter the game based on the situation – and always in that same situation.

Picking out a Jets game I charted from last year, the Seahawks lined up 73 times against the Jets defense and the Jets made a total of 103 substitutions. There were 15 plays where the Jets didn’t make a substitution, although five of these were as Seattle went no-huddle for a two minute drill at the end of the first half and one was while they took a knee at the end, so there was actually less than ten conventional plays where the same 11 stayed out there at the end of the play.

How does this compare with some other teams? Let’s consider Pittsburgh. They have a very uniform substitution pattern. Essentially, they line up in a 3-4 (and within that, the linebackers typically stay in the same positions). Then on passing downs, the nose tackle (usually Casey Hampton) comes out for a slot corner and they go into a 2-4-5 nickel. They do rotate the defensive linemen based on tiredness and obviously some plays (goal line, prevent defense) deviate from this pattern in exceptional circumstances. In a game a charted for them against San Diego, they substituted just 59 players on 82 plays. There were also 28 plays where the same 11 stayed out there.

The Steelers are pretty exceptional though. Let’s consider Houston, who – as noted – has a similarly straightforward package rotation, but does rotate personnel more. In their game with Tennessee, they substituted 76 players in 76 plays and had 31 plays where the same 11 stayed out there.

Tennessee themselves are a pretty interesting case. Here’s a team that rotates a lot but has a couple of basic packages that they tend to stick to. In the game mentioned above, against Houston, they substituted 110 players in 74 plays. That’s a higher rate of substitution than the Jets had in the above example. However, this is usually simple player rotation rather than changing up personnel groupings, formations and packages. The Titans have a deep defensive line with four backups that all get reps, sometimes all at once. So the Titans could have four substitutions on their defensive line (perhaps one on one play, two on the next and one on the play after that) without ever shifting out of their basic 4-3 alignment. This leads to a high number of individual changes, but you don’t get the complexity of roles changing from play to play.

Perhaps the most telling example is your new Super Bowl champion. Obviously their defense evolved out of Ryan’s system and seems to be a great example of the potential that scheme has once you get good enough personnel. They are capable of changing things up and hitting you with a bunch of unexpected packages, but they don’t need to do it all of the time. They can gameplan specifically for this and get everyone prepared and on the same page. Against the Jets in 2010 and 2011 they went crazy with some of the complexity and player rotations, also throwing some complex blitz packages at the Jets. That made sense in the 2010 season opener because you have more time to prepare (and Rob Ryan’s defense did the same thing in 2011’s season opener). Also, they may have identified the Jets as a team they could have success against by doing this, especially in 2011 with Nick Mangold out. They also made some interesting adjustments in the Super Bowl to try and slow down Colin Kaepernick.

While that’s something they are capable of doing, it’s not something they have to do all the time and some of their games this year did follow a pretty standard pattern. In the week 11 game against the Steelers, there were 27 of 73 plays where they did not make a substitution and they did have some distinct defensive packages and personnel groupings that they favored throughout. Is this a product of the current Ravens coaches cutting out some of the complexity from the Ryan system that their current scheme evolved from, or could this be what Ryan would ideally want to work towards if he had the perfect personnel? I can’t tell you that, but at least their success is a sign of what’s possible.

Next man up

One further aspect of this could be player development. Let’s use Aaron Maybin as an example of someone who was a pass rusher and the Jets used him almost exclusively in that role. Did Maybin develop faster because the Jets put him in a situation where he could do what he did best? Or, alternatively, did the fact that he wasn’t getting reps in other situations stunt his growth and lead to a situation where it was a lot easier for teams to figure out what to expect when he was in the game? It’s an interesting question.

On the other side of the coin, you have teams like the Steelers who have targeted versatile personnel. If it’s second and ten and the offense goes to a three wide set, they don’t need to substitute because they have faith in their outside linebackers to line up in the slot, even if it’s only to jam a receiver and slow down their route into the secondary where someone faster can pick them up. The problem with this is that they have so many full time players that when one goes down, you’re not just giving a backup a few extra reps, you’re literally turning a guy like Jason Worilds or Ryan Mundy who doesn’t usually play into a full time player. However, at the same time, everyone knows the system and is ready to step into that role, so the Steelers often manage to cope when someone has to fill in.

The Steelers have had their own cap issues over the last few years and there’s some talk about whether they can keep the team together and how long they can keep relying on new guys to come through (as Brett Keisel did for Aaron Smith and Lawrence Timmons did for James Farrior). If the Jets lost a defensive starter, they’re just as likely to be forcing a player into extended action in a role in which they don’t have much experience though.

2011 Jets-Denver game

One final example to bring us full circle is the game from last season where the Jets lost to Tim Tebow’s Broncos on a short week.

This is an interesting case because the Jets put Bob Sutton – the linebackers coach who, like Pettine, left the Jets last month – in charge of the gameplan. Sutton had coached at Army, so he was used to equipping a defense to cope with an option style attack. What was interesting about this game was that straight away you noticed that they weren’t mixing up formations and personnel groupings as much as usual. It was a standard 3-4 defense on first and second down and then they went into a nickel package on passing downs. There were still a lot of substitutions, because the Jets made wholesale changes on every third down, bringing Marcus Dixon, Jamaal Westerman, Josh Mauga and Kyle Wilson into the game and also rotated a lot of their front seven personnel during the game in an effort to keep them fresh. There were 27 of 67 plays where the same 11 stayed in the game at the end of the play, much higher than usual for this team.

This would prove to be one of the more disciplined defensive performances of the year and the Jets completely shut down the Broncos offense only to allow them to drive 95 yards in the last five minutes to get the win. On that drive, the Jets abandoned their previous rotation and were in a nickel defense on all but one play. Maybe the success of the defensive performance earlier in the game is a sign that these players would make fewer mistakes if the defensive rotations and substitution patterns were more straightforward.

Interestingly, this was the game before the Bills game referenced in the introduction and the cause of the safety rotation that led to so much additional confusion. Smith’s blunder on the last drive, coupled with lingering injuries that had plagued him most of the year, caused him to lose playing time to Brodney Pool and the short term benefits of this didn’t last long enough to get them into the postseason.


If there’s one recurring weakness in the Rex Ryan regime, it’s that he seems to sometimes have a tendency to expect too much from his players. Could it be that his efforts to be creative have caused him to overthink things to the point where the defensive players are on their heels as much as the offense? There are obvious advantages to an aggressive and unpredictable defense, but some of these seem to be cancelled out by the disadvantages, of which there are certainly a few.

Maybe Ryan has deviated too far from the usual pattern of having a base defense with an ability to do lots of different things on top of that. If you don’t really have a base defense, is it any wonder you find yourself searching for an identity?

I’ll be back next week with some more food for thought as we head towards free agency. I’ve got several articles in the works, but if there’s anything you’d like me to investigate or revisit, let me know in the comments.