Upon Further Review: Defense vs Northwestern

no video; sorry.

Aiigh 300 yard first half burn!

We should really look at the chart.

B... burn?

No.

Ch... chart?

Chchart it is.

Are you high? We gave up 300 first half yards to a team that lost to Duke and ran for six inches against Ohio State and you hand out pluses like it's ecstasy at a rave?

Uh. Right, this was my impression as things progressed: "why are all these numbers so high?" It's like I'm suddenly dependent on assistant coaches for the information I receive. Some possible explanations:

• 64 of Northwestern's yards (ok, 56, we can give them an eight-yard out there) came on that fluky he's-down-no-he-isn't opening play. I didn't assign blame here.
• Michigan came out in this funky 3-3-5 that was not what reader DanK suggested in that guest post about slowing down the spread running attack. (More on this later.) Then they stunted out of this thing; several times when they did this Northwestern played paper (hey, no fair!) and Michigan was screwed from the snap. Several gashing runs have no minuses, the occasional five-yarder has a plus, etc. This is a sign you got outcoached. Maybe I should start tracking some RPS stats. I should.
• The D forced five turnovers. This tends to inflate numbers.
• Bacher had a weird day where he was either throwing something horribly behind his intended receiver or feathering a beauty sideline pass. A couple completions actually came coupled with pluses: the slant Harrison jumped that he deflected to a Wildcat, the sideline completion against Warren that was hugely difficult for both WR and QB, a few short throws that were snuffed out immediately upon the catch.

So... yeah. I think the defensive line and secondary both had excellent days. The linebackers, no so much, but whatever.

Speaking of that sideline completion late in the first half...

Right, we challenged it, and I'm totally on board with that.

The addition of the challenge has been mostly superfluous in college. Since every play is reviewed (and referees have apparently taken to standing over the ball waiting for the booth to get a look after potentially controversial game-changing plays, a change I am totally in favor of) anyway, the coaches' challenge usually sits unloved on the sideline, deployed as often as Notre Dame scores a touchdown. I believe this is the first time that Michigan has ever deployed a challenge, and it was a good use of the resource. 1) We were highly unlikely to use it in the second half. See the previous couple sentences. 2) We didn't need the timeout for anything else. 3) It was a close play.

So what about this 3-3-5 thing?

That was ugly, especially since this space ran (and endorsed) a guest post claiming that a 3-3-5 would be a more effective run defense against the spread option that shredded us so in the first couple games. A recap: since defensive ends usually run themselves upfield in a 4-2, you're asking both DTs to have two-gap responsibility. This is kind of tough. A formation like this:

Gives you a better shot at stopping a spread run game, or at least that's how the theory goes. An embarrassing instance of being wrong, wrong, wrong? Maybe. But...

...this is a snapshot from the 49-yard touchdown run. Note the splits between the DT and the DEs. In our diagram above, the DEs are lined up between the guard and tackle. This is 3-technique (I believe, correct me if I'm wrong). Here the DEs are lined up outside of the tackles, which is 5-technique. This is basically a 4-2 with one DT lined up over the center. Also one of the DEs is Shawn Crable and the DT is John Ferrara (I think). This is not what is suggested above. We stunt off this, slanting Crable inside and looping Graham around...

...Crable gets ridden out of the play and Graham gets picked off by the pulling tackle; Conteh into the second level...

...and from there Jamar Adams misses a tackle and he's off. This was the most disastrous instance, but several other times we lined up in this on second and long and got burned by it. You can't have one guy between the tackles and expect to defend anyone's run game.

Yeah, what is the deal with all the stunting and looping and stuff?

I don't know. It seems too cute, especially on downs the offense can run against. We did this frequently against Appalachian State and paid for it. (Oregon, I believe, was more conventional but against a really fantastic trio of runners.) Meanwhile, whenever Taylor is in the game he's playing Gabe Watson and driving someone's ass into Bacher. Like... line up and out-execute them.

You did not just say that.

I did.

Hypocrite.

Well... I was being sort of ironic. What I don't like from the defense is its slight case of Weis-itis. The 3-3-5 with canyon splits on second and long. The third and ten from NW's one on which we assumed they would run and gave up a long conversion. The stunting around an Appalachian State line that should have spent the day eating Taylor's facemask five yards in the backfield.

There is one aspect of this I do like: when Michigan has gotten its opponents in situations that have no threat of a run, they have been deadly. See either NW drive where they trailed late, the final PSU drive, and the entire Notre Dame game. When permitted to tee off on an opponent that can't threaten a run, Crable and Jamison and BGraham have been dominant, often because of the stunting.

My problem with the defensive playcalling is game theory based. Sev
eral times in this game Michigan put Northwestern in a second and long. In this situation you want to lower variance, since most second and longs result in a punt. Michigan frequently deployed that 3-3-5 with huge splits and stunted; this is a high variance strategy that either results in an unblocked guy or two or a gaping hole. Similarly, on third and ten from the one you want a low-variance play call that results in fourth down from the one or the five or the nine... just fourth down. Instead, Michigan puts eight in the box and bites on play action, leaving Ross Lane open for a twenty-yard conversion.

(Note that you want to avoid predictability here, too, so the occasional unexpected/"wrong" sort of call is okay. No one call can be wrong, like no one raise can be wrong in poker, but over time exploitable patterns emerge.)

Goats?

Adams was the big culprit on the long touchdown run. As per usual, Michigan got little production from the linebackers. Chris Graham is an opponent big play waiting to happen.

Heroes?

Crable, obviously, but also Taylor -- finally as dominant as we expected -- and Harrison. I thought the coverage in the secondary was very good all around; hopefully that signals improvement and not merely "we played Northwestern."

What does it mean for Eastern or, more likely, Purdue?

We will get an immediate check to see whether or not the secondary is actually getting better with the Boilers rolling into town. They're like an upgraded version of Northwestern: scatter-shot QB in a spread offense that's more pass than run. Except Purdue has a better RB with Sutton out, way better receivers, and probably a better offensive line. We'll know lots more after this weekend versus Ohio State.