Picture Pages: This Is Not A Stack Comment Count

Brian November 3rd, 2010 at 1:50 PM

Note: no UFR today, as the torrent got down late Monday and I couldn't do the first half then. Hopefully both halves tomorrow.

You'll have to forgive the picture quality on this one—both of these are low-quality torrents. Just like Michigan's defense. AMIRITE!

So in the game column this week I complained about the alignment of the middle linebacker in this bastardized version of the 3-3-5. Michigan has him maybe a yard behind the nose tackle, like so:


This creates a major vulnerability against misdirection, as we'll see. This play is a first and ten on Penn State's first drive. They've driven it into the Michigan half of the field because of depressing things, and more depressing things will happen. This isn't one of them. Michigan shows a two-deep with six in the box, but moves Kovacs down late to add a seventh guy, which gives Michigan the formation above versus Penn State's ace 3-wide.

At the snap the offset fullback heads inside the tackle to his side. You can see the handoff is going to be made to the right side of McGloin. Linebackers start scraping as each and every DE attempts to take on two blockers:


Here's the handoff point. The fullback is hitting the backside B gap, which makes me think this is a called counter play. Where's Demens?


Demens has taken a step towards the line of scrimmage and has hit a guard. Now… he hit the backside guard, the one that PSU is cutting towards. He read the play, but he's a linebacker two yards from the LOS meeting a guard with a free release who's much bigger than him. Momentum means that the best he can do is bounce off it and attempt to flow down the line. (This is much more apparent in the video below.)


The play cuts back as designed. Roh has attacked a frontside gap. Martin and Demens are caught up in the wash on the interior, and Mouton, who was scraping along well back of everyone else, is going to eat the fullback four yards downfield:


The saving grace here is Kovacs, who sifts through the blockers and makes a mediocre ankle tackle that the RB (Royster, I think) steps through:


Demens and others finish it off but after four yards:


Michigan got away with this by putting an extra player in the box late. When Penn State was not caught in a bad playcall, counters like this gashed Michigan all night.

Here's the video:

I don't have an exact replica of this from Rodriguez's WVU days but here's an inside zone Rutgers ran in their 2007 game. Rutgers was no joke on the ground in '07. Ray Rice was around and the Scarlet Knights finished 26th nationally.

The first thing that's obvious is that the MLB is six yards off the line of scrimmage, not two. Also despite playing against a bigger set—Rutgers has a tight end on the field instead of a third wide receiver—West Virginia maintains two deep safeties:


At the snap WVU has shifted to an aggressive look with the OLBs and the spur at the LOS; the MLB has moved up a yard:


At the snap six players attack the line, giving all but one WVU DL a one-on-one matchup:


This is a similar setup, really: inside zone. Main difference is that there is an inline TE instead of a fullback on the backside, but they block the backside end above. The playside end is about to beat a Rutgers tackle to the inside. Note the MLB two yards away from the LOS now—where Demens started the play—after the handoff. He's scraping to the hole. A Rice cutback would be somewhat problematic for him but he's not likely to get a lineman in his face:


MLB has now engaged an OL at the LOS. Rutgers tackle is totally beaten and forces Rice to start cutting:


There are four WVU guys in the area:


And Rice goes down shortly after he crosses the LOS:


On the day Rutgers would get 183 rushing yards, but Mike Teel completed under 50% of his passes and threw two interceptions on a 128 yard passing day because WVU left the safeties back the whole time. West Virginia won 31-3. Their rushing defense was 18th nationally.

So, things:

  • It seemed like Michigan was using Jonas Mouton like WVU used their MLB in the 3-3-5. Except Mouton was four yards off the LOS, not six, and not aligned in the middle of the field. So if he's going to get to anything on the frontside he has to run hard, which means he is susceptible to cutbacks.
  • I don't think Demens ever had a prayer of dealing with a cutback or counter because of his alignment. One step to the playside and he's a yard away from the LOS about to get swallowed by a guard.
  • Michigan plays Demens at the same depth in their other line alignments. 3-4:
    Paired with the disconnect in WVU's 3-3-5 this signals shoehorning to me. Demens should be at a certain depth in more conventional sets and putting him six yards back would confuse him in pass drops, run fills, etc, but in the 3-3-5 he takes one step and there's a lineman releasing free into him. In these sets he's got a chance to scrape without dealing with an unblocked OL all the time. So…
  • Michigan's deployment of the 3-3-5 isn't really a 3-3-5. I don't know what it is, but that whole attacking from everywhere, making different fronts, blitzing, getting guys through the line unblocked thing is something you can see on a fairly typical WVU play above. There are six guys on the LOS threatening and a dedicated cleanup guy behind them with the space and time to get anywhere along the line. Michigan is a passive three man line with guys you can easily single block (but get to double if you want) and linebackers who are living a nightmare. It's incoherent, and Michigan going back to it after having a fairly solid day against Iowa basing almost exclusively from traditional fronts is a miniature version of what happened against Purdue in 2008. Michigan's 3-3-5 is a 3-4 with linebackers in places that don't make sense.
  • Michigan only escapes the above play by outnumbering the offense. No one on the defense beat their counterpart. Everyone was blocked out of the play, which means you can't win unless you've got an extra guy, which means you can't play two deep without getting smashed.
  • I have no idea what Greg Robinson is trying to accomplish. This puts me in the same situation as Greg Robinson.


Yard Dog

November 3rd, 2010 at 3:42 PM ^

The coaches are not getting the kids in position to make plays.  It also appears that against a known running team (us vs. PSU, WVU vs. Rutgers), if you run the stack you need to have six guys attacking the LOS with the MLB running free and not straight into a block.  As Brian points out, Demens has no momentum and is getting swallowed by a guy 60-70 pounds heavier.  At least if he has a few more yards to work with he can properly react and potentially crush the blocker if he has enough momentum.

The kids can only make plays if the coaches put them in the right position - I've heard this from so many coaches I know it makes me crazy, and in this case, it appears to be completely true.


November 3rd, 2010 at 3:46 PM ^

No. 27 was a miniature Ron Dayne for Rutgers.  The Knights had a banged up offensive line that season (who's OL is ever 100% anyway)?

I have to accept all that you've written about the LB formations here, Brian.  I probably need to buy this book and study it.

Something's really effed up with this.  Greg Robinson clearly doesn't know the tactics and itty bitty details behind the 3-3-5, and sadly It's as if UM's defensive assistants (former colleagues of Jeff Casteel at WVU who are familiar with the 3-3-5) only understand half of the story as well, and that they are relaying things piecemeal to Robinson over time (i.e. reasons why to line up at x, reason why not to do y). It's all half-assed, delayed delayed and impromptu.

The way things are playing out Michigan is probably going to be 5-4 heading down to West Lafayette in a must-win game.  It's de ja vu 2008 Purdue, where we have a DC who doesn't know what in the hell he is doing or why he is or isn't doing it, and a bunch of fledgling defensive assistants who will be pointing fingers at their boss for the loss.

I think Michigan beats Purdue though and finishes 6-6.


November 3rd, 2010 at 3:54 PM ^

Or would Michigan just be better if they just played man to man with some safety help deep?

At least they would be in the same zip code when a pass is thrown. 


November 3rd, 2010 at 3:57 PM ^

can you please just post something that will make us all feel not so sad about the series of dongpunches that have come our way the past few days?


i have faith in you fearless leader, and ill be refreshing mgoblog every few minutes, waiting hopefully in anticipation.


November 3rd, 2010 at 4:09 PM ^

The players have hearts, and brains, and ideas about how winning football is supposed to be played.

Almost all of these kids were very, very, successful at football in HS.

They decided to come to Michigan.

They put their trust in these coaches, worked their ASSES off, and expect to be rewarded with success.

When they experience nothing but constant failure how can any football player who knows how to play the game ever since Pop Warner not have their faith shaken?

I'm wondering why there hasn't been even MORE attrition than we've already seen?

 How many of these players secretly think GERG and Gibson are blithering idiots?


November 3rd, 2010 at 4:10 PM ^

if we could get someone from the program to explain just what the hell is supposed to be going on here.

Never gonna happen for a lot of reasons, I know, but I desperately want there to be some reasonably straightforward explanation for this.  Somthing like blown signals such that half the guys were doing one thing and the other half were doing something else would be maddening, but fixable.

As it is, we're left guessing and those guesses are pretty unpleasant.


November 3rd, 2010 at 4:16 PM ^

So Rich wades in, we hear about a bunch of player position moves during the week, and we play a different scheme than we did against Iowa. Also, GERG is forced to a 3-3-5 based scheme because that is what Rich and GERG's assistants know ....

GERG might be as no-longer-capable as you're suggesting Brian, but how can we really tell when Rich has forced mid-season experimentation ala-Purdue '08, the players are inexperienced, and he's already adapting his defensive plans to their 3-3-5, etc.?

I know this seemed to be the case in '07 under Lloyd as well, so maybe it is normal, but it just seems like the teams we have field these past few years are not really capable of developing in season. They get in the weekly cycle of prepping for the next opponent, so once you get through fall camp, trying to make mid-regular-season adjustments to scheme and position assignment are a recipe for more mistakes not less. (As we saw in '07 you can develop in post season bowl practice though.)

Like I said before Saturday last week - I was hoping the changes were around getting some guys game experience where they are to play next year, versus trying something new mid season, less a Purdue '08 result ensue. Well, the latter occurred.

This forgetting about the past / only looking forward thing, seems to be at the expense of overlooking valid lessons learned from the past ...


November 3rd, 2010 at 4:26 PM ^

I have the exact same concern.  I am all for creativity and doing something - - - anything - - - to try to plug the hole.  But, as Brian has pointed out in so many season and game previews, position switch players starting is a sure negative sign.  Well, that is exactly what we are seeing. 

In RR's defense (something I have NOT been doing lately), if he just kept the defense the same, we would all be screaming our heads off about his stubborness and lack of care / innovation on defense.


November 4th, 2010 at 4:36 AM ^

And Ezeh would still be starting instead of Demens.  When nothing is working, change is needed.  Maybe the changes that did not work last week (with Martin injured and out) will work a little better this week and a little better the week after.


November 3rd, 2010 at 4:23 PM ^

"This puts me in the same situation as Greg Robinson."

Brian, in no way shape or form are you GERG.  Your hair is not as stylish.  You actually know football.  And, you are not facing almost certain termination.


November 3rd, 2010 at 4:48 PM ^

says we'll be schizophrenic with our fronts and formations again this week.  This team has no defensive identity and because nothing has worked the coaches are grasping at straws and trying new things every week.  They're on tilt.

Once Royster gashed us early, we freaked out and were susceptible to screens, etc. for the rest of the night.  Penn State's FB screen was to Michigan what Air Force

I was having nightmares about defending the option this week but in reality our troubles all boil down to an inability to cover and we'll get picked apart in the air once we sell out to stop the run.  Of course, selling out to stop the run means we'll still give up huge chunks on the ground as well.

P.S. One minor quibble I have with many of the above comments is the contention that the kids aren't being put in position to make plays.  This is true but only to a certain extent.  Talbott should have had a pick against PSU.  Vinopal could have broken the TD up on the fade.  Kovacs should have had a Rick-6.  Cam Gordon should have blown up TJ Jones.  There are tons of examples from every game where even decent (more experienced? more talented? less confused?) players make big plays to kill drives or create turnovers.  It's not all on the coaches.


November 3rd, 2010 at 5:24 PM ^

I was having nightmares about defending the option

IIRC, Michigan always seems to have trouble defending the chuck and duck. This week, I fear, will be no different.

Blue in Seattle

November 3rd, 2010 at 4:51 PM ^

Before I get to my point about Casteel, I don't see anything in this post from Brian that indicates there is a dramatic change from the Defense of the Iowa game and the defense of the Penn State game.  That to me would be an indication that the coaches are tinkering too much and throwing stuff at the wall to see what sticks.  What seems more likely to me is that the Michigan D is facing progressively more talented offenses coupled with increasing injury to key players.  And what is wrong with Penn State other than the QB.  And actually this game granted was a walkon, but how hard is it to handoff?

In any case, if you are dead set on dreaming about he next DC, I doubt that it's going to be Casteel.  Based on this article, I think Casteel is licking his lips at the chance to become the HC at West Virginia.  And I think he had his sights set on that the day Rich Rodriguez left for Michigan.  I mean what would you do?  Leap at a lateral move to a talent depleted team in a glaring spotlight? or convince the regents at WVU that you've demonstrated both your loyalty and talent, so just give Stewart the boot and assign me HC?

If he wanted to be the DC at Michigan he would be here already, burning in the same purgatory the rest of us are enjoying.

Go Blue Beat T…

November 3rd, 2010 at 5:19 PM ^

It's hard to play 110% when you consistently get knocked off the line at the snap. Fuck alignments and schemes and position and lateral movement-- run defense is about one word, "penetration."


November 3rd, 2010 at 7:02 PM ^

I agree with one major point in the post. I'm confused as to why M went back to the 3-3-5 after having a decent, for them, week against Iowa. The insistence of RR and defensive staff to go 3-3-5 is not working. I agree that G. Robinson is not getting it done, but I'm not sure, just like with Shafer, that the staff is doing anything to help. Frustrating, especially for my nine months pregnant wife who has to listen to me rant about four man fronts and misuse of Craig Roh for three hours. 

Ben from SF

November 3rd, 2010 at 7:09 PM ^

Congratulations to y'all for your new addition!

My wife actually demanded that my brother and I go to a bar to watch the game due to my excessive negativity.  Imagine trying to find a free table and TV when the local baseball team (Giants) plays a world series game!


November 4th, 2010 at 11:35 AM ^

I'd put this guy out of my mind entirely- he was in the Memory Hole, never to return.

Thanks a lot, arod.  Many fond memories of the day this jabroni was fired.

As to the picture pages, Brian's analysis gives food for thought, but I think the real problem with the play is the D-line getting blasted off the ball by an injury-depleted Penn Schtate O-line.  Multiple free-release second-level blocks and combo blocks- they got everything they wanted on that play, and got Kovacs on Royster 1:1.  Kovacs, to his credit, slowed Royster down enough to allow tacklers to converge, but Penn State still had to be happy with that matchup.  A missed tackle by Kovacs and this play could go a long, long way.


November 3rd, 2010 at 7:16 PM ^

This to me just illustrates why RR must go.  I'm not at all sure how anyone (including Brian) could look at the defensive philosophy on this team and not draw the conclusion that Rodriguez is woefully inadequate in his position as head coach.  

Logic would dictate that if the defensive scheme is incoherent, but the head coach fails to draw that conclusion from obvious facts, then the head coach may not have the requisite knowledge of schemes he's previously run to continue in his position.  You simply can't reconcile one with the other.  You might disagree with Brian's analysis of the 3-3-5 (Michigan v. WVU).  But if you agree with Brian's analysis, then I don't see how anyone can't think RR just doesn't understand one major aspect of his job.

ngandu.com: the case against Rich Rod


November 3rd, 2010 at 8:05 PM ^

Brian, I know everybody thinks I'm a Devil's advocate around here, but I have to say...

I'm not quite sure what you're criticizing.  Maybe I'm just a failure at reading.  But the point of the 3-3-5 is to have 8 guys in the box (in this case 7, since they spread us out) and outnumber the blockers.  On this play, every player maintained his gap responsibility, at least well enough to have a successful play.  There were guys who got manhandled a little bit, but they're all in their gaps.

Meanwhile, the fullback leads up on Mouton, which leaves Kovacs completely unblocked.  His job is to help to the inside, which he does.  If our DL holds their ground a little bit and squeezes down the hole, then Kovacs has a smaller gap to defend.  It would be great if Kovacs could step up in the hole like a full fledged linebacker and stuff Royster, but Kovacs isn't very big nor a great athlete.

This is a gap control defense.  The middle linebacker isn't meant to roam from sideline to sideline.  He's meant to have a gap (in this case the playside A gap), which he fills nicely.  Royster bounces it to the B gap, through which the FB is leading up on Mouton.  Okay, fine.  There's another guy in the box to make the play, which as I said, is the point of the 3-3-5.

I'm not saying these things just to be a Devil's advocate, and I really think I must have missed something in the post.  You seem to be complaining about this defense, even though a better tackle from Kovacs would have limited the play to maybe 2 yards, which is a win for any defense.  It's not run perfectly because some players didn't execute, but the scheme here seems to be fundamentally sound.


November 3rd, 2010 at 10:57 PM ^

I'm not good at analyzing a defense myself, so I have to depend on others.  That this was a basically sound defensive play with some poor execution makes a hell of a lot more sense than Brian's claim that it's a complete dumpter fire.

I fear Brian is getting a bit too emotionally involved in his analyses.  The two of you both know a hell of a lot more than me, but it would be nice if there could be some collaboration on these picture pages and other things so that there is less jerking us poor saps around with varying (and contradictory) assertions about what it all means.

I don't know what is actually possible.  For myself, I want rational, insightful analysis and I'm beginning to doubt that's what I'm always getting.

Space Coyote

November 4th, 2010 at 12:11 AM ^

The scheme isn't flawed, it's the execution.  

It should also be noted that the 3-3-5 doesn't need to be aligned the same way for every team.  Just as there is a 4-3 under/over, there are variations for how to align in a 3-3-5.  While I have never taught a MLB to align this close, I'm sure there is reasoning.  I'd be interested to hear it, but either way I don't think the MLB depth pre-snap is the reason this defense is bad rather than good.

Edit: It is my opinion that it is simply Demens job to smash the A-gap, as Magnus said.  A reason for lining up so close could be to prevent combos off the NT to catch the MLB 4-5 yards off the LOS.  You notice how Demens lines up close, the combo can't get to him deep and his gap is filled (as has been the case often this season, read: every complaint made ever about Ezeh getting blocked 4 yards off the LOS).  It appears the other MLB (I believe Mouton) is scrapping over the top initially, but then as he sees the play bouncing back he goes to hit the play side B gap.  It actually looks to me like Michigan motions into a 3-4.  The D-line getting pushed off the ball and making the gaps larger than they should be appears to be the big fault on this play.


November 4th, 2010 at 4:46 AM ^

If Martin were in, and not injured, perhaps the D-line would not have been getting pushed off the ball, the gaps would not have been larger than they should be and Kovacs would have been able to make the tackle.  Everyone, or almost everyone, seems to be forgetting that our one really good defensive player was out with an injury.  That had to hurt us, particularly since Royster seemed to be really on his game last Saturday.


November 4th, 2010 at 5:13 AM ^

And exactly what I was thinking.

I do think that Brian's broader criticism is that this defense is so static that it allows offenses to pick their point of attack because the linebackers don't have room to react to the the play. I think that's an interesting perspective, but assumes the linebackers are experienced enough to make those calls consistently.

Brian also argues that the only reason this works is because Kovacs was there in the first place; on the other hand that might be an example of RPS +1.

So this is a chicken/egg, or XO/Jimmy Joe discussion, AGAIN. These players may not be capable of reading defenses well enough to be put in positions where they have to make more than one or two choices quickly. At least with this scheme each player should, in theory, know exactly what to do and do it decisively.


November 3rd, 2010 at 9:01 PM ^

Great breakdown and shows what happens when you have a defensive scheme in place that nobody on the staff, you know, knows how to run correctly.  And you don't have players who can fit the system well.  And you have 8 freshman on the 2-deep.

It is sad, but the only thing that shocked me in these picture pages was the ticker below the WVU-Rutgers game that listed UVa as the #15 team in the country.  What happened there?


November 4th, 2010 at 7:32 AM ^

I know the origins of it, in a typo at Syracuse, adopted by exasperated fans there. But at Michigan it has generally been used as a term of ironic affection (especially in all caps), and I'm just not feeling it.

Although I think Magnus' critique of Brian's thinking is probably right. Michigan had a player in position to make the tackle, which is about all you can ask of a defensive scheme when you don't have players (i.e., Martin) in the game who can disrupt the offense.

Penn State Clips

November 4th, 2010 at 11:52 AM ^

The saving grace here is Kovacs, who sifts through the blockers and makes a mediocre ankle tackle that the RB (Royster, I think) steps through.

You are correct, sir, that was Royster.

I'm looking forward to the UFR. Not in a bragging way, but because I don't have the expertise to break down the game (I'm just a fan) and appreciate those who can.