Upon Further Review 2014: Defense vs Utah

FORMATION NOTES: Not much of note. M was in a 4-3 over most of the day with few deviations. Miami did do some of that late-model RR/Baylor stuff by putting their WRs way far away from everything. I called this Shotgun 4-wide far. (This looks like pistol but the QB is 5 yards deep and the RB will pick a side/position presnap.)

SUBSTITUTION NOTES: Line starters were the usual. Backups saw some changes, though. The backup ends were still Charlton and Ojemudia; the second 3Tech was Godin—I don't think Wormley even played—and Hurst got in for some snaps that had previously gone to Pipkins. Again, don't think Pipkins played.

Linebacker was all Ross/Bolden/Ryan, and Michigan spent a big hunk of the game in a 4-3 playing zone so Ross got a fair share of time. When Michigan went nickel they slid Countess inside with Peppers and Lewis on the edge, and then Peppers disappeared and they stopped playing nickel. Hollowell didn't get on the field.

Safety was Jeremy Clark the whole way and Hill; then Hill left and Thomas came in.

APOLOGY NOTE: I used a different video file format and my conversion process balked at it, so the audio is messed up on the clips. I just didn't execute. Sorry guys.

[After THE JUMP: this went well.]

This is late.

Yes, but not nine months late!

Okay, point. This bit was rather good.

It was. Now, Utah's offense isn't spectacular—they finished last year 73rd in YPP—and their 50-point bombings to open the season likely say more about the opponents than any surge from the Ute offense.

Even so, drives before it was over:

• 6 three and outs, one ending in a long FG attempt because of a M INT
• A defensive TD
• 79-yard field goal drive that was basically that one screen
• 54-yard FG drive
• 67-yard TD drive
• 32-yard FG drive
• 14-yard FG drive

They made one big mistake and then they had three separate incidents where whatever pattern matching scheme they were using was falling apart on mesh routes; once they got burned on a run where they ended up with five in the box. Everything else was snuffed or the kind of dink stuff you give up so you can defend actual offense.

A lot of Utah's passing offense was the nothing hitch for five yards. One of a half-dozen examples:

I usually go "eh, push" on an immediate tackle for five yards because getting more aggressive than that is asking for trouble and it's really hard to go down the field in five-yard throwing increments. When it's third and four I do want Michigan to be tough on anything near the sticks.

Hooray.

Any time you give up about 10 points worth of field position over the course of a game you expect to win.

Sometimes you don't.

chart

okay

M straight up dominated the Utes OL with their DL, allowing Michigan to spend almost all of the game with six or even five in the box against a rushing O that uses their QB. That is extremely impressive, and allowed Michigan to cut off the WR screen game and effectively cover deep for the most part.

What happened on the screen?

Michigan got RPSed, for one, with a bunch of guys screaming at the QB and nobody peeling back. That is likely a missed assignment for someone who's supposed to check the back, but who I don't know. If I had to guess I'd say Glasgow, but major grain of salt there. I attributed the points there to RPS.

Then Jeremy Clark and Blake Countess both took very bad pursuit angles. A safety going upfield of a block on a run that's at the sideline is just… well, it isn't good.

The RPS (and possible busted assignment) got Utah 20 yards; Countess and Clark going parallel to the LOS on a textbook example of that pursuit drill you see everybody run in practice is the other 47. Both of them need to avoid that blocker by going around downfield and then try to mitigate damage already done.

And the repeated mesh whatevers?

The touchdown felt like a pretty massive breakdown in pattern matching. Not only did Utah have a drag route open for 28 yards, but they also had a wide open wheel route down the sideline:

When I look at these things I am trying to figure out how to make this coverage make sense, which is difficult. Bear with me; here's how I think this is supposed to work:

• This is quarter-quarter-half coverage. Craig Ross asked Mattison about what he was running and this was the answer, which clears up a lot of things I thought last week.
• Thomas has a deep half with just the one WR to his side of the field. Lewis and Clark have quarters to the field.
• Lewis should pass the WR off to  Clark when he shows post route and prepare for a wheel.
• Ross should break off his coverage on the out way before he does and he should definitely not take it vertical, instead recovering to find an underneath threat.
• Ryan shouldn't be dragging out of his zone at all and should in fact be looking for the mesh route as soon as the #3 WR breaks inside. 

If these things happen Michigan covers everything at least okay and if they give up the drag route it's not for a touchdown. The one guy I don't blame at all on the play it's Thomas, who doesn't have anyone in his zone and starts moving to the other zones to clean up. A guy with a deep half is not responsible for a drag route to the other side of the field.

Lewis was terrific.

Yes. He showed his speed on the screen and against a guy who had a bunch of yards playing in the Pac-12 a year ago. He put the clamps on Dres Anderson here:

When they tested him again he was running Anderson's route for him:

He did get beat for a first down thanks to his eagerness to get over the top of a fade that Anderson sold nicely; that was the only flaw on the day.

Mailbag reprise: where did Pipkins go?

Man I know we're complaining about everything but this is not a place in which the coaching staff's decisions look anything other than logical. Glasgow turned in another very solid game. He rarely gets sealed, he takes on doubles and runs through them, he seems to understand where he's got to be a lot. And because he's a nose tackle none of that shows up in a box score. The inverted veer is pretty tough to stop and Michigan hammered it most of the day, with contributions from everybody. Here's Glasgow burrowing through a double to constrict space:

That's only a half point in the system, as constricting space usually is, but it's an important part of the play since it sends the QB further outside than his blocks would like him to be and gives Ryan an opportunity to make a stick in the hole.

Here Utah motion gets Michigan down to five in the box and Utah still gets a blah gain because Henry and Glasgow both win blocks:

Also Beyer. If you're going to have five guys in the box you better win multiple DL blocks and Michigan did that for the most part.

Henry had an equally impactful game; here he surges up the backside of a power play to ruin Utah's angles and set up a TFL.

And even though Glasgow ends up with no part in the play watch him feel the block, start to rip away from it, delay as he figures out where the ball is, and run down the line to provide support if the D needs it. The guy does not stay blocked. Whatever you think about Hoke, give him this: guy can coach up a DT. Hurst got in and had a few nice snaps. Hell, one of Ryan's play-both-guys-on-a-mesh events saw both Mone and Charlton discard their  guys to provide ultimately unnecessary assistance.

I mean, take out sacks and Utah ran for 3.3 YPC in this game. Last year they were at 5 even with sacks included. Michigan held ND to under 60 rushing yards. Whatever is happening at defensive tackle is working.

Jake Ryan back?

Ryan had a bunch of tackles and not one but two instances where he played both guys on a mesh point, and he does feel a lot more comfortable as an MLB as the season goes along. He is getting significant help from the DL winning those one on one blocks, but on that first Glasgow clip above you can see him run through contact instead of trying to avoid it like he often did at SAM. When provided a play that should put the ball in spot X, he showed up in spot X. Here a slot blitz from Ross is paired with a slant to the play by Michigan, so the ball should pop up behind Beyer, and when it does Ryan is there:

That's not hard, except insofar as all linebacking is hard; it is a thing where you need to understand the playcall and feel where the ball is likely to go, something that didn't seem to be the case in the opener.

That tendency is not gone. Ryan seemed to drag way out of position on two of the three mesh routes. Michigan was playing zone hypothetically suited to defending those but Ryan wanted a chuck so bad that he ran after one guy and opened it up for the other twice.

There were also a couple instances when Ryan sat still because he didn't understand what he was seeing. Here he makes this block really easy:

And I thought he had the right idea on a similar IZ play that ended up on the backside but didn't trust his read enough:

With the line slanting hard playside he should know where to go as soon as the center releases without hitting anyone, because now Willie Henry is going to be ripping up into the frontside. At this point he's got the information he needs:

If he hits it hard this is a tackle at the line; he slows it up and it turns into a big gain to kickoff Utah's touchdown drive. (I also think with a slot guy coming in off the edge and a QB who's meh in the athleticism department that Beyer might need to be more aggressive crashing down on the back.) It's not just not getting to the line, either: he gets caught between blowing past the OL and hitting him hard and ends up getting hit, sealed, and moved, making the cut to the secondary an easy one for the back.

It is getting better. I've seen a lot worse instincts from Michigan linebackers who have been playing ILB for years.

And you're on board with Bolden now?

I'm giving the offseason hype train an 8/10 so far. He is much more decisive and this has led to a large reduction in the plays where he gets hit instead of vice versa. Even when he doesn't make the play he's kind of making a play, like this dangerous seeming screen he raided:

That shot upfield draws the attention of two blockers and forces the receiver to Lewis, so when Clark comes up he's unblocked and attacking a guy with a CB wrapped around his ankle. Again, nothing in the box score, but play made.

He scored negative in this one as he fell down a few times and had some tackling issues, but overall I think he's made good progress.

I told you Frank Clark was going to be fine.

Heyyyyyy. I told you.

Heroes?

The DL won all day. The good bits of Jake Ryan were real good. Jourdan Lewis had a breakout day.

Maybe not so heroic?

The bad bits of Jake Ryan were bad. Jeremy Clark had that boggling pursuit angle, which was just one bad play but it was 40 yards worth of bad play on a day where Utah got fewer than 300.

What does it mean for Minnesota and beyond?

The DL is ready to manball it up. This will be an interesting matchup against a very solid, veteran interior line that has a potential All Big Ten player in G Zac Epping. Michigan has given every indication they are up to the task so far.

Jourdan Lewis locked down a starting spot, Taylor and Peppers or no. Best secondary performance of the year, hands down, and one that promises more down the road.

Ryan is making progress, but is still prone to breakdowns. Especially in coverage. He remains at his best when asked to reprise his edge terror role, and Michigan is smartly asking him to do that about as frequently as possible without becoming predictable. Those tackles on mesh points are excellent stuff.

I want Jarrod Wilson back please. I don't think Dymonte was a big issue but with M going with a lot of single high safety looks I am much more comfortable with Wilson in that spot than Clark.

This may be the defense we all hoped for. Get the secondary healthy and this is a unit that has answered a couple of the questions asked in preseason by getting excellent play out of Glasgow and pretty good play from the LBs. Get Peppers or Taylor right and put 'em opposite Lewis and it feels hard to move the ball on these guys.