Upon Further Review 2014: Defense vs App State

FORMATION NOTES: They promised it all offseason and they delivered on it:

In your face bro. Note that this was an example of something I started calling "30 slide"*, as the linemen are basically head up on the tackles and center with Frank Clark as a sort of standup end/SAM.

*[The idea being this is a 30 front (three linemen head up on tackles and the center like a 3-4) with three linebackers slid as if they are in an under.]

Another example is even clearer:

Note that in both of these shots, the three-tech is in fact to the bottom of the screen instead of between Clark and the nose as you might expect. I had not seen this before, because Michigan doesn't run it and they weren't spread enough last year for anyone to do it against them but since Appalachian State also ran it…

this was in fact their base D probably

…and they are right in the heart of spread on spread warfare I figure it is the latest fad when you need to account for the QB in the run game. I'll get Adam to ask Mattison about it next week—unfortunately, they moved the coordinator pressers up a day so I was not educated on anything before that time came. I'll try to accelerate my UFRing process, something that is now feasible with fast downloads and the lack of TWIS on my plate.

They also of course ran a lot of standard nickel:

Michigan also debuted a weird 3-3-5-ish package with Frank Clark at "MLB":

This happened twice. On  both plays Clark was running at the frontside guard on the snap, impacted him, blew him back, forced a cutback, and then no one was there. More on that later.

Michigan also played some bonafide dime snaps:

These had three DL, two linebackers, and six DBs. Generally it was Delonte Holowell getting the extra nickel snaps but that's more in the…

PERSONNEL NOTES: Deep breath. On the line it was Beyer-Henry-Glasgow-Clark to start with copious substitution. Your nominal second string based on playing time was Charlton-Wormley-???-Ojemudia, with the NT ??? a combination of Pipkins, Mone, and Hurst. Pipkins looked by far the best of those guys; I expect that NT rotation to quickly settle down into Glasgow and Pip alternating with scattered snaps elsewhere. Godin got some real PT early at 3-tech.

At linebacker, Ryan, Bolden, and Morgan seemed to get about equivalent PT. Ross got a number of snaps as the game went along as an ILB. IIRC, Jenkins-Stone only saw snaps as a nickel DE late. Gedeon and McCray got in for the last drive.

Michigan played nickel on I think literally every snap they weren't playing dime. That was Peppers spotted by Hollowell and then Hollowell after Peppers got dinged. Taylor and Countess got starters' minutes at outside CB with Lewis coming in frequently; Stribling did not see time until heavy substitution began in the third quarter. Richardson got in there too.

Starting safeties were Wilson and Clark; Thomas got quite a lot of PT starting in the second quarter, with walk-on AJ Pearson seeing the field on ASU's interminable second scoring drive.

And hamburgers: I thought I was done calling people CGordon and TGordon and just realized we have two Clarks. I tried to clarify who was who below; I imagine you can figure it out if I missed a couple.

[After the JUMP: a big table! and some other stuff.]

Well that seemed like a good start.

Yes. Michigan has to work out some mental kinks. They messed up a zone, their linebackers didn't understand what the line play meant to them sometimes, and they got dinged on that belly play a bit. Those are all easily attributable to first-week goofs that have a good chance of being resolved.

Aside from that, Michigan felt physically dominant against an overmatched opponent everywhere. Wide receivers couldn't get a micron of space. OL had a hard time not getting flung into the QB. It was what it had to be if Michigan is going to make a leap to kick-ass. Necessary but not sufficient, of course, but an encouraging first step in Michigan's new Teddy KGB defense.

Remember easy yards?

Nope. Not this year.

CHART

You're supposed to interrupt me…

CHART

…about this. Oh. I get it. You're bei

CHART

aggryessive

Chart. But first:

UFR CHART REMINDER/ORIENTATION SECTION

I give pluses and minuses to players who do things that I think are good or bad and then add them up. A +3 is a monster play, +2 very good, +1 quality, and half-points are for things that are not too hard but done well (or hard and done okay) and play a part. Minuses go the same way.

Because of the way football works this is not fair. Mentally adjust yourselves: DL need to be positive to be average. It is a MAKE PLAYS position. +4 is our approximate "okay" benchmark, though with the number of plays and heavy rotation in this one take it with a grain of salt.

Linebackering is hard and we are content if a linebacker is at 0. Truly good players will be consistently positive.

DB is both hard and off the screen unless you screwed up most of the time. We have a coverage metric that should always be your first point of contact when trying to figure out any individual secondary member. Also, members of the secondary can get +4 or –4, even –5 I think I gave out once, for massive huge plays like INTs or 80-yard coverage busts. (Any individual long run has a thousand fathers and the negatives get spread out.)

In general I'm trying to make it so that a 4 or 5 yard run or 6 or 7 yard pass is a push all around and for numbers to make sense as results diverge from there. We adjust for situation, of course, and also based on the quality of the opponent: if they miss a WR open for 30 yards you're still getting negged.

CHART

okay okay

Keep in mind that nobody got a ton of plays so even on the DL anything positive is good.

You're just mad about eating a lemon.

Bro, do you even chart?

Well no. I just like dismissing data based on supposed personal biases of the presenter when I don't actually have a counterargument.

Do you work in cable news?

I do, actually.

All this makes so much more sense.

Anyway, I thought the linebackers were not so good. The first play from scrimmage could have been a TFL as Clark blew up a tackle and forced the RB to a dead stop, but by that time Bolden was on the ground and Ryan had run himself way outside the hashes as the backside LB.

Bolden checks a gap that Ryan has and Ryan just runs his ass off: if we are trying to defy our stereotypes, Bolden hesitating because he doesn't know what he should do and Ryan running as fast as he can somewhere is not a good start. (This still should have been a zero yard gain but Beyer does not go parallel to the LOS, yes.)

To get some more on the LB issues, we have to look at…

What was up with Clark at "MLB"?

It was a weird callback to what the 3-3-5 is supposed to look like. The 3-3-5 is really a four man front that doesn't tell you which guy is headed to which gap before the play. It always sends a guy to the line from the LB corps and everyone takes a gap; the idea is to use the element of surprise and the benefit of momentum to cover up for the fact that all your dudes weigh about 230.

So Michigan put Clark at MLB and had him pick an OL and go donkey him. This worked pretty great. Not many people are going to take a 280-pound train coming at their face and not stumble backwards. Clark impacted a guard, blew him back, and immediately forced cutbacks to unblocked guys…

…who weren't there. Ryan runs at the frontside of a play that isn't going to have one. Then he goes to the bench.

The second one actually sent Ryan and Clark. Both guys got picked up but the play was forced to the backside hole. It was again completely open:

Hell no I ain't happy.

This time both Bolden and Ryan were immediately yanked.

To me this seems like a reoccurrence of an issue Michigan was having a couple years ago where they'd slant the line and the linebackers would play it straight, either unaware that the playcall meant they should expect a cutback or unable to prevent themselves from executing their Pavlovian response to guys stepping to inside zone. The Clark gambit is designed to blow up the frontside of that play, get the guy blocking Bolden blocking air, and have Bolden waiting to thump the RB when he cuts back. Ryan getting blown inside on his blitz doesn't help things, but it must be frustrating for linebackers to not react to the play you've called. That is why they went YOINK.

So what do you do with Jake Ryan then?

Both Michigan and App State played a number of snaps in something I called "30 slide" where the line is in a three-man front with the fourth lineman standing up on the edge. A reprise from above:

From top to bottom, those players are Clark, Beyer, Glasgow, and Henry. It seems like it would be very natural for Ryan to become this standup end. It is basically spread SAM. You could then have Clark put his hand down and significantly beef up the 30 front.

That's a chunk of playing time and then you can rotate him in for Beyer and Clark at DE, especially Beyer, on passing downs. Everyone is going to rotate on this defense, Michigan is going to play a ton of nickel, and they can find him his snaps at spots he is more likely to make a positive impact than he did at MLB in this game. This still-theoretical Morgan injury obviously puts a big monkey wrench into that plan, but if you're in a nickel James Ross is still available, and Gedeon and McCray may round into viable options over the course of the season.

We like Pipkins?

Yeah. He was the most effective NT of the backups by some distance, and while I'm willing to believe that Glasgow is doing things in practice with enough consistency to deserve to get a lot of playing time it seems like a hard case to make for either Mone or Hurst. Pipkins took doubles and made tackles:

I think he will come on as the season develops.

Are we concerned with the relative lack of pressure in this game?

Hard to say. After some thoroughly failed experiments early…

…ASU barely threw a pass longer than eight yards. It wasn't until the middle of the third quarter that they had a completion longer than six(!) yards (they had a nine yarder called back for a chop block), and for the game they had two completions that travelled more than eight yards downfield in the air, the second coming against the third stringers on the 19-play TD drive.

Given the lack of opportunity, both in terms of sheer number of plays and the dink and dunk nature of them, two sacks, four hurries, and a number of crowd-alarming thunks to the QB's chest…

excellent YOU GOT JACKED UP "aaaooooowwwww" from the crowd

…are good. I mean, I kind of felt the same way but when I looked at it either the QB was about to die or at least get annoyed on route to a +18 pressure metric. Jury remains out. Ask again next week.

What was with that play where they rushed ALL THE PEOPLE?

Michigan surely did not intend to fling seven(!) guys at the QB on this play:

This whole play was borked—the guy running a wheel to the top of the screen was wide open for a touchdown—because it looks like half the defense is playing one thing and the other half is playing another.

Wilson is motioning Clark into man on the slot receiver who gets targeted, but it kind of seems like Taylor is running a three-deep zone (I can't tell because of this game's stupidly tight camera angle). If you're running man free then having a corner blitz, as Peppers does, is bad, and then they have no one for the tailback if he leaks out to the right. Bolden's certainly not checking him.

Michigan's lucky that didn't hurt them in an important game; hopefully we don't see that happen again. This was –3 cover and –3 RPS, so when M was actually trying to play the same defense those numbers are even a bit better.

In the picture pages you mentioned playing "games" to stop the belly play. Have an example?

Meant to include this:

Corner has the QB, DE slides inside and must be picked up by the tackle, Morgan is there to shut it down for one yard.

Even in this case you would like Henry to not get moved so much. He's got an issue understanding when the OL is using his momentum against him. If they actually want to stop him, that's when he's good.

SIDE NOTE: This is Michigan's 30 slide package. Watch the play again focusing on Beyer, who gets blown up by a double team. Not relevant to this play but 1) that's why Michigan doesn't want him playing SDE, and 2) that's why putting Clark down seems like a better idea. It's weird that they're aligning this way when Beyer is a former SAM and 25 pounds lighter than Clark.

What happened on the long completion?

So I think in certain situations Michigan is running either pattern-matching concepts or hybrid man/zone stuff. (These are the same thing.) It is otherwise hard to reconcile what exactly is going on on certain plays.

Here Clark rolls to the line of scrimmage and runs man to man with the slot receiver, who goes on a wheel route. Meanwhile, the rest of the defense is definitely in a zone. Lewis's drop is pure cover 3, the linebackers are clearly dropping into zones, it's a zone. Except one of the slot guys is getting manned up by Clark.

This is pretty well defended before it even gets to Lewis, as Ross gets a fingernail on the ball to change its trajectory. It may still be a completion even so but that'll happen if a QB just gets it over a LB. Instead Lewis undercuts it and nearly picks it off after coming inside. That feels like he knows that if the #1 WR cuts inside that Clark has that and he can play that guy in quasi-man.

Michigan has been tracking vertical slot receivers deep for a while now but it often felt like there was a point where they were supposed to come off and pass it to the safety level; this feels like they've ditched that and now if you are the slot defender in a zone and that guy goes vertical, he's your guy. This is an attempt to deal with one of the main problems of a 3-deep zone: four verts.

Now the table is set. On the 34-yarder that set ASU up for their first touchdown, Wilson rotates down to the interior receiver of a trips set and tracks that guy the whole way as he starts cutting inside. This is four verts.

Wilson is step for step with the guy he's covering all the way as they pattern-match on vertical routes from slot receivers. Clark is… also with that guy. At worst he should be to the center of the field, if not shaded to the trips side; if he understands this coverage against four verts he should be laser-focused on the #2 receiver to the trips side.

This probably could have been a Picture Pages.

But the secondary was impressive?

Yes. Downfield separation was nonexistent except on the above play and the rush-seven incompletion above.

While I think Jeremy Clark busted just above and he got a silly PI on the final drive, I still liked his debut. He is rangy and seems athletic. His man to man coverage was generally high quality. If he stops busting, he'll be good.

Also in corner depth for days, it says something that when Michigan went to dime early that it was Delonte Hollowell who came in instead of, say, Countess sliding down to a spot he played all last year. They clearly like him in that role. It didn't feel uncomfortable at all with him out there after Peppers's absence. While he didn't get many opportunities to make plays, his coverage seemed good and the one time they tested him he delivered a wallop:

Shame that was already incomplete, as there's a good chance that knocks the ball out. I will take that kind of contact all day on a six-yard slant.

Meanwhile I don't think either starting cornerback got a throw in his direction all day except for the Countess stick on a WR screen. When a starting CB doesn't chart, as Taylor didn't, he is doing work.

Peppers?

Had one nice open field tackle and was not otherwise tested. This section only exists because I clipped a three yard punt return that seems like it heralds a butt-ton of fun from that guy over the next few years.

Those first two moves are enticing.

Heroes?

Frank Clark and the DL in general. The secondary in general save for I guess Jeremy Clark.

Maybe not so heroic?

Jeremy Clark did bust on the long completion and his PI was silly. Has to clean that up. I didn't think either starting LB did well.

What does it mean for Notre Dame?

We're going to miss Morgan if he is indeed out. Maybe the starters get better; not having him as an option in case they don't is bad.

Jake Ryan to MLB doesn't look so hot after game one.  I don't see him getting better to the point where it becomes one. That's just not his game.

The secondary lived up to the hype insofar as it was possible. Press coverage and zero separation. All day.  Even got an unexpected additional contributor in Hollowell, who was good and fiesty, like Brandon Harrison (hopefully without the comprehensive inability to blitz).

The line is super deep. A two deep of starters, basically. Nose isn't as deep as we might want right now but Glasgow/Pipkins looks just fine.

Frank Clark looks like end of 2013 Clark. This is muy bueno.

I really wish we didn't have duplicate names on defense again.

It could be happening. Check back Saturday to see.