3-3-5 defense

Earlier in the year I took a cue from Michigan's odd announcement of Adam Braithwaite as an OLB/safeties coach to theorize that Michigan was adopting something half 3-3-5 stack, half 4-3. You can put whatever label you want on it, but it's apparently similar to what Virginia Tech runs. After yesterday's press conference, though, the prevailing opinion is that Michigan's defense is going to be half 3-3-5, half 3-3-5. This, for the Ohio State fans stopping by, is 100% 3-3-5.

Wha? Aigh! Justin Siller! No—

Right.

Evidence for the switch is plentiful. In this episode of "Inside Michigan Football," Troy Woolfolk talks about "the new defense":

In yesterday's press conference the players all made references to the 3-3-5. The usual array of practice reports coming from shadowy trenchcoated internet folk all say that not only is Michigan running the 3-3-5 in practice, that's all they're running. This is no longer in the realm of rumor.

Is it in the realm of sense? I don't know. The major reasons I and other tea leaf readers were banking on an aggressive 4-2-5 were threefold: it's basically what Michigan was trying to run most of last year, available bodies on the defensive line point towards an undershifted four-man front, and Michigan's latest recruiting class features a zillion guys who were told they would be "quick" ends a la Roh.

The 3-3-5 as a base set obliterates the quick. Michigan cobbles theirs together by dropping Roh back to one of the outside linebacker spots. The defensive end spot not occupied by Ryan Van Bergen is now going to be a Banks/Patterson platoon or a 294-pound Mike Martin. Since 3-3-5 defensive ends are not lumbering quasi-DTs like 3-4 defensive ends (more about this later), Martin seems like a questionable fit at end; the alternative is platooning Martin and Campbell, two of the most physically dominant players on the team.

The Unresolved Questions

Is this an alternate look or a base set? If it's a base set, how often will they deploy a four-man front?

Early indications are that Michigan will use it as a base set. One theory out there is that Michigan is running the 3-3-5 to the exclusion of other defenses because Mike Martin is out for spring. I don't think that makes sense. A team that spends all its time learning one set of responsibilities because one player is out for spring practice only to switch to a considerably different set in fall is a team that is going to get its coach fired at the end of the year. Teams don't devote the entirety of spring practice to a "new defense" that is then a changeup when the season comes.

Michigan used the 3-3-5 from time to time last year, most prominently in the Ohio State game when it was an effective base set that shut down Ohio State's I-formation running:

This is actually more of a 5-3 since the DEs are lined up over the guards and the box safeties are rolled up tight to the line of scrimmage, FWIW, but that's a matter of alignment against a run-heavy team. Note that Roh is an outside linebacker here.

This forced OSU into some bunch formations that forced Michigan out of the stack; OSU also attacked it by running single-back formations that are inherently strong against single deep safety defenses because of old-timey football wonk stuff. Buckeye Football Analysis has a deeper analysis if you're in the mood.

When OSU went unbalanced, Michigan responded by putting Roh's hand down and going back to their usual undershifted four-man line. For Michigan the personnel will be exactly the same, allowing them to shift between fronts at will. So if the 3-3-5 isn't working in a particular game or just turns out to be a bad idea, they aren't totally screwed.

They would be at least partially screwed, however, since they're piling more and more on the plates of linebackers who spent a lot of time last year wondering what to do (or decisively doing the wrong thing). The way West Virginia ran a 3-3-5 allows linebackers to be blitzing players who have to do a minimal amount of reading, but if it doesn't work then all that time will be time that could have spent fixing what ails Ezeh and Mouton in a 4-2-5.

I'm not thrilled that Michigan seems to be changing its defense again, especially since I've been pitching defensive coordinator continuity as a major reason Michigan's defense will improve in 2010, but given what they ran most of last year the only players who will be making major changes are the linebackers. In the West Virginia version of the 3-3-5, defensive ends are basically the same as they are in a 4-3. The nose tackle is more of a two-gap player if you can make him one, but that's not something that requires a lot of reading. So… yeah. Maybe it will work.

I bolded this in the announcement about Adam Braithwaite's hire but failed to grasp its oddness and potential significance:

University of Michigan football coach Rich Rodriguez announced Thursday (Feb. 11) the hiring of Adam Braithwaite as the program’s safeties/outside linebackers coach. … Assistant head coach Tony Gibson will coach the cornerbacks and free safety position.

 Your response in the form of a cat—which they should totally do on Jeopardy, BTW:

Braithwaite now coaches the safeties and the outside linebackers. Greg Robinson now coaches the inside linebackers, emphasis on the plural. Tony Gibson now coaches the cornerback and… uh… one safety position. What sense does it make? It makes none. It makes less sense if you believe the premium moderator folk who have been asserting that Michigan is in the odd habit of calling its deep safety "strong" and the guy who rolls up to the line of scrimmage on occasion "free."

So what the hell is going on here? First: however deeply screwed up Michigan's internal lingo about safeties is, my assumption is that the free safety is your deep-zone ballhawk and the strong safety is the guy who rolls up to the line as a semi-linebacker. It would be totally insane to give Gibson the guy at the line and some cornerbacks and Braithwaite one guy at the line and one guy in a deep zone. (Wags may joke here about Rodriguez's previous defensive hires. Take my defensive coordinator… please!)

So that means Gibson is the secondary coach and he is a man in charge of three people. Meanwhile, the outside linebackers coach has a safety or two. Hmm…

Now, I know what you're thinking: ack Donovan Warren is about to get an eight-yard screen in his grill. Or ack Andrew Quarless is about to run straight downfield untouched for a 60-yard touchdown. Or ack… well, we were all around last year. The walls have dents to prove it.  The only thing you're not ack-ing is the defensive line. This is a digression.

The particulars of Braithwaite's hire indicate the eight-man front Michigan ran much the second half of last year was not an immensely unsuccessful attempt at emergency triage on a walk-on-laden matador defense but rather the intended base defense going forward. With so many bodies ticketed for Craig Roh's "quick" position, I don't think this presages a move to a straight 3-3-5 like West Virginia ran. The 4-4-ish set above is likely to be Michigan's most frequent alignment, with quick-as-linebacker sprinkled in as a changeup.

It's half 3-3-5, though. Aside from the disposition of the line and the middle linebackers, that's what it is. The secondary aligns like it's a 3-3-5. The "spinner," while technically a linebacker, was safety Stevie Brown last year and will be either Josh Furman, Mike Jones, Isaiah Bell—though he may have moved inside—or Brandin Hawthorne this year. All of the candidates were high school safeties tagged as tweeners except Hawthorne, who was a safety-sized defensive end. Last year the strong safety was Jordan Kovacs (tweener safety), Mike Williams (just a safety), or Brandon Smith (tweener safety), but this year it's likely to be Marvin Robinson, Carvin Johnson or Vlad Emilien: more high school safeties tagged as tweeners. The distinction between the OLB and safety is also in keeping with 3-3-5 principles: most teams have wacky names for the strongside (spur, spinner, ferret) and weakside safety types (hero, bandit, saber-toothed dragon) because they have different roles. As Jeff Casteel explained so elegantly on the incredibly expensive and totally useless (at the time) 3-3-5 DVD I bought, the weakside guy "gets his meat cooked"—does not have to deal with lead blockers—and the strongside guy "gets his meat raw"—oh God, that's Owen Schmitt and I weigh 210 pounds.

If I'm right, this is one hell of a bold experiment for Rodriguez. His ass is in the wind right now and last year's attempt to implement this was a flaming wreck unprecedented in the modern history of Michigan football. I'm not a coach but I do watch unhealthy amounts of college football and I don't think I've ever seen anyone try this 4-2-2-3 style of D—please correct me if this is not the case. It's a gamble.

[Update: Corrected. Virginia Tech does this plenty. Tyler Sellhorn:

Dear Brian,

Please don't freak out about the defensive changes.  Mostly it seems like a move to VaTech's defense.  The high school where I coach plays a very similar scheme to VaTech (visited Blacksburg twice to commune with Bud Foster and his staff) and has a similar scheme to what you are suggesting M is declaring. 

I think the best terminology for the setup is the 4-2-5 is "multiple" meaning that most of the defensive calls are intended to trick the opposing QB/coaches at the snap.  The OLBs are really more like SSs and rotate up and back based on d-call and opposition formation/personnel, and the "FS" is nearly an identical player who can roll down to play the OLB spot as well based on motion/personnel.  Up front, lots of stunting (lining up in one gap, crossing into another), lots of gap exchanges, lots of rolling coverages where the OLB/SSs will drop into deep coverage.  The scheme is sound.  Maybe you should do your HTTV tape study this summer on 4-2-5/VaTech stuff?

The biggest reason teams use the scheme? OLB/SS away from two-reciever side/strong side plays tight behind backside ILB so that he can flow hard to action away, OLB/SS has what you have called the scrape exchange.

Two notes: "flowing hard to action ", whether it was away or not, is definitely what Michigan was doing with Jordan Kovacs when he was the box safety and Woolfolk was deep, and who was the other team hard after Josh Furman? Virginia Tech.]

On the other hand, going from year one in a system to year two will be a rare privilege for these Michigan defenders—it will be the first time in anyone's career other than fifth-year seniors this is the case—and I'm heartened that amongst the flaming wreck at the end of last year there was a semblance of a long-term plan. By God, they were terrible, but they were terrible with purpose. This passes as optimism for the 2010 Michigan fan.

Anyway, we'll find out in spring if my speculation here is correct. I think they've tipped their hand. I have no idea how it's going to work out but at least they're sticking to one thing for the first time ever-ever.

Side note on Gibson-related special teams bitchin'. I've seen or heard a lot of this over the past couple weeks, and have to provide a "dude wait what" to that, too. Michigan's coming off a year in which they finished third in net punting, 23rd in kickoff returns, 62nd in punt returns, and was 11 of 15 on field goals. Taken together, the metrics indicate one of the best units in the conference, if not the country. Critics are likely thinking of Michigan's persistent inability to field a punt in the Rodriguez era, Zoltan Mesko going blue screen of death on his rollout option punt against Michigan State, and a couple roughing the kicker penalties.

Those things do detract, but they're offset by the kickoff coverage—I can't remember a big opposition kick return—and a few punt blocks. Special teams were the least of Michigan's problems last year. They have to replace both specialists this year so there might be a hiccup. Even so, I'm baffled by special teams criticism outside of HOLD ON TO THE DAMN BALL issues that I think are a little fluky. (Yes, even now.)