Art Kafuman's Defense: Arizona

This post is going to build toward talking about the UA game, but it's also going to be my first general introduction to some important concepts in defensive football. We're going to look at some defensive structures before getting into the problems that UA caused for them.
I wrote the bulk of this post before the CU game, so some new issues obviously came up. I won't comment on much of that here, so this post is really just an isolated commentary on our UA gameplan. I'll also say that I couldn't bring myself to break down last year's defense, and I've since deleted all digital memory of that horrible season, so I can't immediately compare it to Buh's defense except on a general level. Theoretically it'd be awesome and really informative, but for me the ratio of frustration to value wouldn't be worth it when we have the current season to talk about.

Defensive Concepts

I've already talked a little bit about coverage shells and the most general characteristics of each coverage, so this post will talk more about fronts and techniques in an effort to fingerprint our defense. To get started talking about defensive fronts, let's recall the way that gaps and DL assignments are called.

 Run defense operates on the understanding that, because a ball-carrier can't run through another human being, as long as the defense can cover all of the spaces, or “gaps,” between offensive blockers, they should be able to shut down the run. Therefore, the defense won't be worried about individual blockers per se, but rather about open spaces on either side of those blockers. These gaps are labelled with a letter starting from the inside out. The gaps outside the center are the A-gaps, the gaps outside the OG's are the B-gaps, the gaps off-tackle are the C-gaps, and anything outside a TE (or far outside an OT when there is no TE) are the D-gaps. So, when you teach your defense what their assignments are, you'll tell the NT that he's responsible for the A-gap, the DT that he's responsible for the B-gap, etc., and you'll make sure that you've assigned each gap to someone.  So how many gaps are there?  If X = the number of blockers, then the number of gaps = X + 1.  A spread formation with just the five OL as blockers will be best defended with six front defenders, an I-formation with 5 OL + 1 TE + 1 FB will be best defended with eight front defenders, etc.

This is different from the way that you'll call alignments. Alignments are labelled by number, and tell a defensive player where to line up relative to an offensive blocker:

Even numbers indicate that the defender should line up heads-up over the offensive blocker. Odd numbers indicate that he should line up shaded outside of the blocker, and even numbers with an “i” added mean that he should line up shaded inside the blocker. These numbers increase as you work from the inside out. So, a 0-technique lines up heads-up over the center. A 2-technique lines up heads-up over the OG. A 3-technique lines up in an outside shade on an OG, and a 2i-technique lines up shaded inside an OG. The TE numbering is weird and I'm not sure anyone knows why, but for whatever reason an inside shade on a TE is called a 7-technique, while an outside shade on the TE is a 9-tech.

The combination of a player's alignment (technique) and assignment (gap) tells you how the play is going to work. So, a typical alignment and assignment might tell your DT to line up in a 3-technique and to play the B-gap: 

Whatever he does, he can't let the RG get outside of him (or the RT get inside of him). It's OK if he gets blocked a little so long as he can stay in between the RG and RT, because as long as he's doing that he's got the B-gap covered.

You can also use this system to indicate line movement. I could have that DT line up in a 3-technique, but then angle inside to the A-gap:

No matter what the defensive call is, every player in the front will be given their alignment and assignment according to this system.

The assignments for individual players will look something like that. From that nomenclature we can start building up into different defensive structures. The two most common defensive fronts in the 4-3 are the “shifted” fronts referred to as “over” and “under.” These fronts are said to be “shifted” because they put a DL over the C, and only one of the offensive guards. Whichever side has the OG covered, we can say that the defensive line is “shifted” in that direction. This “shifted” side of the formation is most commonly called the “reduction.” The characteristic feature of an “Over” front, then, is that the reduction is set to the TE:

This front is shifted toward the diagram's right, where we're covering the C and the RG, and away from the diagram's left, where the LG is uncovered. In terms of our nomenclature, in a typical over front, on the TE side of the diagram the DT is a 3-tech and has the B-gap. The DE is a 7-tech and has the C-gap. Away from the TE, the NT plays a 1-tech and has the A-gap while the DE plays a 5-tech and has the C-gap. To simplify things for talking about the LB's, we can say that the SLB has the D-gap outside the TE, the MLB has the A-gap to the TE side, and the WLB has the B-gap away from the TE, although as we'll see LB assignments are more complicated than this.

This is contrasted with the “under” front, which shifts the DL away from the TE:

The under front puts the reduction away from the TE. So, from left to right, the DE is a 5-tech C-gap player, the DT is a 3-tech B-gap player, the NT is a 1-tech A-gap player, and the DE is a 5-tech C-gap player. Because this front is shifted away from the TE, it'll also put the SLB on the LOS as a 9-tech D-gap player to get an extra guy on the LOS. The MLB has the B-gap to the TE, and the WLB has the A-gap away from the TE.

Why does it matter where you put the reduction? Don't you have all the gaps covered regardless? Basically, it all comes down to who's controlling which gap. It's hard to run through a gap that a DL is covering, and for some blocking schemes it's impossible. The DL are 275-300 lb. guys doing their best not to move, and for the most part they're responsible for one gap and only one gap. On run plays they're stuffing that one gap, and on pass plays they're pass-rushing through that one gap. There aren't a lot of blocking schemes that can get them out of there. Accordingly offenses want to attack the “soft” gaps that are covered by LB's. LB's are starting out off of the LOS, they have to worry about reading run or pass, and they have to flow with the ball on lateral plays. Basically, there's a lot going on for the LB's that keeps them from racing straight downhill into their gap, making it easier to block them out of their assignment. When a DC is choosing where to place the reduction he'll have this in mind. Say that an offense's best running play is to the TE-side B-gap. Well, the DC might call an over front, which will put a DT in that gap and make it hard to run that play. This is just one out of many considerations, as we'll see below.

Art Kaufman's Defense: Front Play

It looks to me like our base defensive call is Under Cover-4 (although we do mix in a fair amount of Over, and in coverage we play Cover-0 and Cover-1 with varying degrees of frequency according to situation and gameplan). If any of you read my BI post on Buh's defense long, long ago, this might terrify you, because it's the same base defense. The devil is always in the details, though. There are only a limited number of defenses that most college teams base out of (generally either the over, under, or odd (classic 3-4) front combined with Cover-1, Cover-2 (or 2-man), Cover-3, or Cover-4 behind it), but there are tons of ways to play each of these fronts and coverages. So, calling our defense a 4-3 gives you almost no information. Calling our defense an Under Cover-4 is a little better, but still won't distinguish it from many of the other defenses out there. For the rest of this post, we'll look at some little details that make our defense what it is.

In the previous section I talked about placing the reduction either to or away from the TE side of the formation. We've played against a lot of spread formations that use no TE, so our front obiously can't be set based on where a TE lines up. We usually run an under front, meaning that we'll put the reduction away from the “strong side” of the formation. For now, we'll define the strong side as the side with the most WR's:

 

In this diagram the left side of the diagram has 3 WR's while the right side only has 1, so we'll designate that side as the strong side. Because we're running an under front, this means that we'll place the reduction away from the strong side, or to the single receiver side. In terms of gap assignments, even though the offense is in a spread trips formation, the run assignments are no different from what we saw above when there was a TE. The space between the LT and Y is still the C-gap, and the DE is still responsible for it. The space outside of Y is still the D-gap, and the SLB is still responsible for it.  The main other difference is that the SLB lines up off the LOS to put him in better position to cover something like a slant by H, and so that he can see his keys on the OL and in the backfield better.

When the offensive formation doesn't have a clear strong side, we'll set the reduction to the RB's alignment:

 

In this diagram there are two WR's on each side of the formation, meaning that neither side has more WR's. The RB is aligned to the right of the diagram, so we set the reduction to that side.

Those are some of the ways that we'll set our base front. Above I mentioned that, although many teams run Under Cover-4, the devil is in the details. I also mentioned that LB assignments are less straightforward than I made it sound. Well, those LB assignments are the kinds of details that I'm talking about, and right off the bat we can see some differences from what we were running with Buh.

First, how is run defense supposed to work, in more detail? In any defense, LB's can't just blindly rush into their gap. Let's see why:

 

This is a spread version of “power” against our base defense. From the diagram it looks like the RG is pulling through the hole to block the MLB, but in reality thinking of him as blocking the MLB doesn't fully describe the problem. What this play is really doing is removing a gap from the right side of the formation and inserting it into the left side of the formation. On the right side of the formation before the snap there is an A-gap outside the center, a B-gap outside the RG, and a C-gap outside the RT. When the RG leaves, however, there is just one big gap between the center and RT where there used to be two gaps (one between the center and the RG, and another between the RG and the RT). Meanwhile, pre-snap on the left side of the formation there was an A-gap outside the center, a B-gap outside the LG, and a C-gap outside the LT. When the pulling RG inserts himself between the LT and LG, he splits the B-gap into two separate gaps (one outside of him, one inside of him). If the WLB just blindly rushes downhill at the snap, the defense won't have the numbers that it needs to stop the play:

 

Whichever side of the pulling RG the MLB attacks (I've drawn him attacking the outside shoulder of the RG here), the RB can cut to the other side and gain yards. In order to stop this play, the WLB has to recognize that the RG is pulling and follow him across the formation:

  

Put into the terms we've been using, the pulling RG takes a gap away from the right side of the formation and re-inserts it on the left side of the formation. If the defense is going to stay gap sound, they also have to take a defender away from the right side of the formation and insert them to the left. The MLB can then attack the outside shoulder of the pulling RG while the WLB attacks his inside shoulder. The defense stays gap sound.

So the LB's (and to some extent the DL) can't just rush their gaps, they have to read the offense's blocks and react in the correct way to counter them. The details of Kaufman's specific brand of Under Cover-4 are all about how they're taught to react to different kinds of plays. Let's see how the run fits worked against just one of UA's favorite plays and formational tendencies.

Here's a classic UA/UO-style play. Early in the game UA wanted to run inside zone toward the reduction and away from trips:

  

The actual playcall here is to hand the ball off to the RB. All of the other bells and whistles attached to this play are just ways to protect that running play by punishing the defense for overselling to stop the RB run.

For the inside zone part of the play, the OL is zone stepping to the right. The RT will block the DE, the RG will block the DT, and the C will release to LB level. The LG will take the other DT, and the LT will release to the MLB.

The key block here is the RG vs. his DT. This block is the RB's first read, and he cuts off of it. He has a number of choices depending on what happens. He's seeing whether the DT goes inside or outside of the RG. Say that the DT goes outside the RG and plays hard into the B-gap:

 

In this case, the right A-gap should be open. If the WLB tries to fill that A-gap, the C will have a great angle to block him (as shown in the diagram above). The RB will cut to the left of the RG and get downhill. What if the MLB gets over the top of the LT and into that A-gap?

If the MLB comes over the top, as I've drawn, or if the A-gap is otherwise filled, then the RB will cut it all the way back and get yards inside the DE. What if the DE crashes down to take the RB on that cutback?

This is where the offense wants to start punishing the defense for selling out to stop the RB run. In this case the QB should keep the ball, read the SLB, and decide either to keep the ball himself or to throw the bubble screen to H. As a side note, we can see from this set of rules that the old-school option defense of hitting the QB on every play doesn't work against the zone read. If the DE plays the QB, either by slow-playing him or straight up hitting him, then the QB will just hand the ball off and there won't be anyone in position to defend the cutback by the RB:

 

What's a defense to do? The offense has an answer no matter what happens! Well, not quite. The defense CAN take it all away, and here's how we did it against UA:

This all starts with the defense recognizing the RB's alignment and reading the blocks of the OL. This does not appear to be a stunt that is called in from the sidelines. Rather, the defense is ready for the tendency based on the offensive formation and reading the blocks of the OL. If the OL zone steps to the right, that sets the chain of events in the diagram above in motion. On the right of the diagram, the DT and WLB see the OL zone stepping towards them. The 3-tech DT falls inside the block of the RG into the A-gap, where he bangs into the center and prevents him from getting to LB level. The RG is still trying to block that DT, so the DT is able to take out two blockers (C and RG) all on his own. Meanwhile the WLB takes the B-gap hard. This automatic stunt takes away the front side zone play, forcing the RB (if he gets the ball in the first place) to cut back. On the backside, when the DE sees the OL stepping away from him, he crashes down hard and plays the cutback. Because the DE crashes down on the RB, however, the QB should keep the ball. That's fine, because we're scraping the MLB to the outside as the QB player. This is a read that the MLB is making. When he sees the OL zone stepping away from him, he comes around outside of the crashing DE. This leaves the bubble screen as the only option left, but we're assigning the safeties to any bubble screens. This gives us a 3-on-2 advantage (X and Y blocking vs. the CB, FS, and SLB):

We wouldn't game-plan the reads like this against every team, but because UA often runs this particular play, we do. Most importantly, although there are a lot of moving parts, the reads are simple: For everyone in the defensive front, they just ask if the OL is zone stepping toward them or away from them, and react accordingly. For the safety, he hardly has to worry about the run at all. He keeps an eye on the receiver that he's already covering, and the second he sees him bubbling for the screen he flies downhill. Easy.

This solution shows a key way in which our defense this year differs from Buh's, at least as far as I can tell with my limited breakdown of last year's defense. Buh wanted to involve the safeties against the run in order to outnumber the offense.  Last year Michael Lowe was our leading tackler, Cameron Walker was third, and Avery was all over the place before he got hurt.  That's not all because defenses completed a lot of passes against us.  This increased responsibility unfortunately also put the safeties in run-pass conflict. On pass plays they had to drop back and be deep defenders, while on run plays they had to fly up to the line of scrimmage. This isn't as crazy as it sounds and is a reasonably common way to play Cover-4, but we clearly couldn't make it work. As a result, our safeties easily got caught looking into the backfield on playaction passes, and we gave up a ton of big passing plays as a result. Our CB's probably looked worse than they were last year because of these safety assignments.

Kaufman, even though he's running the same front and coverage, involves the safeties in the run game much less, at least against these spread sets. He only asks the safety to line up over his receiver, and to fly up on him if that receiver runs a screen. In general, Buh wanted the LB's and DL to spill runs out to the sidelines, where the safeties could clean up.  We often failed to do this, resulting in those massive outside runs that we saw so much of last year. Kaufman, on the other hand, puts much more of the onus on the LB's. He wants to keep all runs inside the OLB's so that the rest of the defensive front can make the tackle in pursuit, and as a result Jefferson and Barton are our leading tacklers this year. This does ultimately make us more vulnerable to cut-backs (this is something that I've seen going back to his Cincinnati defenses), but that's where the safeties can come up and at least stop the play from going for big yards. If we start to see big runs up the middle, we might have to rethink this strategy, but for now it's working.

This element of our UA game plan let us do several things: (1) it let us defend the run with a 6-man box. If you check out my post on ourown QB run game, you'll see that the defense needs a 7-man box to have good numbers against 5 OL + 1 RB + 1 QB. Our auto-checks to the stunts described above messed up UA's run blocking schemes and let us defend Solomon and Wilson with only 6 men, giving us an extra defender against the pass and screens. Also, by making our OLB's our primary run force on the ouside, we've freed up the safeties and limited some of the plays that killed us last year: Certain types of screens, QB runs, perimeter runs, and easy deep passes down the middle of the field. As bad as the defense has looked at times this year, we aren't giving up more against runs between the tackles, and we're greatly improved against several of our weak spots from last year.

 

Art Kaufman Defense: Coverage

So where do things go wrong? Against UA (and CU), things go wrong against the pass, and mostly against the quick passing game. Before getting into coverage, let's talk for a minute about Anu Solomon: A lot of people are saying that they aren't impressed with him as a QB, and to an extent I have to agree because of a lack of versatility. He has three throws that are his bread and butter: The hitch, the slant, and the fade. He doesn't seem comfortable reading floating safeties and LB's in the middle of the field, and I don't see him working many multiple receiver concepts downfield. This doesn't mean he won't be a good QB, since Goff had a lot of the same problems last year and has since corrected them, and Solomon and his WR's are way ahead of where Goff and our WR's were last year on the fade. Right now, though, he's limited. Fortunately for UA, these throws complement each other quite well.

Getting back to coverage. We've already looked at some basic coverage shells in another post, so let's talk a little more about technique, and the different ways that a CB can line up on a WR. The tightest coverage will come by lining up 1-yard off the WR and pressing him. If you don't want to do that, you can play off of him (with the depth depending on the coverage) with either inside or outside leverage. When you line up inside of the WR, you're partially conceding routes with an outside release but putting yourself in position to stop inside releases. Playing with outside leverage is the opposite. When you're playing with leverage, you'll want a second defender who can help on the routes that you're partially conceding. So in Cover-2, the CB will line up with outside leverage on the WR and will prevent him from taking an outside release. Meanwhile the safety is inside and over the top. The CB uses his leverage to force the WR inside to the safety, and the safety's inside alignment lets the CB play hard against outside stuff. The CB protects the safety, and the safety protects the CB. Their respective leverages complement each other:

 

The problem with UA, in part, was that the fade has an outside release, while the slant has an inside release:

 

When you press a WR and he beats you on each of these routes a few times, you can get caught guessing, and your play against both routes can deteriorate as the game goes on. If you give up on the press and play off of this WR with inside or outside leverage, then that impacts the rest of your coverage.  A defense is like a spider web; if you pull on one part of it, every other part of it will change to some degree.  If you play with inside leverage to take away the slant, you need another player deep and outside to help on the fade:

 

This, in turn, affects the way that the rest of your deep coverage has to play, and so on.  If you play with outside leverage to keep the fade from getting to the sidelines, you want an inside short player to take away the slant:

 

Those pictures look like pretty good options, right? Wrong, because the offense has two or three other receving threats in the game:

 

This is Cover-2. Here the CB's are playing outside leverage, and would hypothetically force the fade inside to the safety near the hash marks. This puts them in a bad spot against the slants that're being run, though. The next defenders inside of them are the OLB's, but they're in a bind. If they cut the slants by the outside WR's (the dashed lines), they leave a lot of space for the slants by the inside WR's. If they stay inside the inside WR's (the solid lines), they leave a big hole for the outside WR's. The MLB can only take one of the inside WR's slants and really needs to be on the lookout for a route by the RB anyway, so the defense is in a bad spot here. What about other kinds of leverage?

 

Here the defense has rolled coverage to the diagram's left and is playing Cover-3. The problem here is that you can only roll coverage to one side of the formation. This might take away double slants to the left of the diagram, but on the right of the diagram you can't take both slants, at least if we're assuming that we need to have at least one LB (MLB) in the box to defend the run.

The simplest solution to these problems, by far, is to have the CB's press:

 

The CB's play tight coverage on the slants by the outside WR's, and the OLB's can sit with inside leverage on the slants by the inside WR's. It's can also be good against the fade if your CB's are physical, because they can get a good jam to take speed off the route and can then force the WR to the sidelines so that there's no space for the QB to fit the ball into. This is what we would've liked to do, and if we could've upped our percentage of success against either the slant or the fade from the press it would've worked.

Unfortunately, our CB's are overmatched against good WR's with size. It's not just Cayleb Jones. Now against UA and CU, whenever we have our CB's covering the fade, opposing WR's are almost always more open than our own WR's are on similar routes, and our WR's are pretty good. If we had Marc Antony, 2008 Darian Hagan, 2008 Chris Conte before he moved to safety, or even SQT (just to show it's not all about size), we beat UA and the CU game is much more convincing. All of those guys, regardless of measureables, defended the fade much better than anyone we have right now. More length would also help us break up some of those slants from a press alignment. More length means longer arms, making it easier to get a good jam on a tall WR. As it is, our CB's don't have the combination of size, strength, or technique that we need. Our best hope of a quick fix is for Darius White to get things figured out, but the potential of that is all based on physical stats right now, and those are kind of meaningless when compared to all of the other factors that make a good football player. If you can't press, then you're back to playing the leverage game, with all of the accompanying problems that we saw above.

There are a number of things we could do to try and disrupt this, but I've given too much info too fast as it is, so let's try to get a group discussion going. Obviously we aren't college-level coordinators and won't solve anything, but discussions like the one I'm about to propose will help make sense of different coverages much better than any of my posts can, so let's try this: These are three plays that caused us repeated problems in the 3^rd and 4^th quarters:

Hitch:

 

 

All slants:

 

 

Verticals:

 

You have 11 guys, and you need a call that can defend these three plays. Preferably, you'll also want your call to limit huge runs even if you're willing to concede some things in the run game. Assume Cayleb Jones is the left outside WR (X). Give me your personnel group of choice, tell me where you want them to line up, and give me the coverage rules that will let you cover all three concepts. I'll draw up any suggestions and post them in the discussion thread, and we'll hopefully get some good conversation going.