1. Is the defensive line going to survive?
Son of a bitch. I told you not to ask that. I don't know, man. I don't know.
On the face of things it's not completely dire. Michigan starts two seniors and two juniors. They're big. The backups aren't freshmen, for the most part, and when Michigan's in the nickel package they'll lift the dodgiest parts of the line for what promises to be a stunting, slanting, pressuring Ryan-Roh-Black-Beyer/Clark group. The starters are all touted recruits save Black. Meanwhile, Michigan has three DL coaches and coached Will Heininger up like whoah last year. BONUS: If you squint it kind of looks like "QWash" looks like "quash."
They're unproven, and the lack of playing time last year is a cautionary note. Defensive linemen rotate, and rotate a lot if their coaches have faith in them. Washington hardly existed last year. Campbell did, though, and to a lesser extent so did Black.
A potential problem is the swing in strategy Michigan has to undertake as they transition away from the best penetrating nose tackle at Michigan since NTs ballooned into the 300 pound range. Quinton Washington may turn out all right; he's not going to be Mike Martin. This means the linebackers have to take big steps forward, beat guys who are (hopefully) releasing late after Washington and Campbell shove them back, and fill impeccably. The linebackers' jobs should actually get easier since Michigan has a pair of guys who can demand doubles (hypothetically); they'll have to make a quantum leap in consistency if the rushing defense is going to tread water.
Add to that a non-nickel line that looks like it's not going to get anywhere near the quarterback and you've got a recipe for frustration, or at least a lot of bending as Kovacs and company make tackles to extend drives and the front four tries to put opponents in passing downs.
Verdict: meh, but no worse.
[After jump: more defense, more Mattison, more PANIC?]
2. You're going to bring up the fumbles again, aren't you?
Yeah, I am. It's been a topic of dissention all offseason and you've heard it. I'll be brief. ZooWolverine put up a great diary that analyzes year to year correlations in the turnover realm and came back with the following results:
- Nothing is that predictable.
- Interceptions are more predictable than fumbles.
- Fumbles forced and fumble recovery rate didn't correlate year to year at all.
- Michigan didn't do well with interceptions.
- They did do well with fumbles—very well.
This was his conclusion, which I agree with totally:
The last category is randomness, which appears to have a very large impact on even the most skilled category, and complete control over a couple of them, meaning any real prediction is fairly foolish. To be a little foolish, then, I’d guess that interception categories improve to above average (say low 40s), but overall turnover margin gets worse, dropping to the 50s. However, I have only slightly more confidence than I do when calling a coin toss.
This is going to be a drag on any improvement the defense may make (unless it's not, small sample sizes and all that). We're just talking about fumble recovery rate here; obviously forcing fumbles and interceptions (usually with QB pressure) are things you can actually do. You cannot hustle to the ball any better than anyone else, and Michigan is going to be walking uphill as they try to replicate last year's 80% performance.
3. Surely there's something else to PANIC about?
Yeah. A major reason that Michigan's defense was so good last year was their crazy good short-yardage defense. Seth checked it nine games in and again after OSU and found that Michigan had reduced real* opponents' success rate on third and one from 78% in 2009 and 2010 to 35% last year.
That is a staggering one-year turnaround, and it's not just good relative to Michigan's awful GERG defenses. In 2006 when Alan Branch and company were busy annihilating the opposition, they gave up third and one conversions at a 33% rate. What Martin, Van Bergen, Heininger, and company did last year was capital-e Elite. Three of the four DL starters and the two guys who drove more of that success…
…than anyone else are now gone, leaving questions in their wake.
If the fumbles get recovered at a more realistic rate and Michigan isn't booting opponents off the field on third and short, you're dealing with a lot more long drives, and that makes your shaky DL depth more of a factor, and that leads to more long drives, and… yeah. Real pounders could be an issue. Like, say, Alabama.
Michigan will have more success against spread-type teams as they go to an aggressive nickel package for which they have good personnel. Teams that can force them to use their big package will have the twin luxuries of probably doing well on third and short and probably getting lots of time when they pass on first and ten.
Unless, of course, they don't because leaps happen.
*[IE, not MAC or I-AA]
4. This is not doubting Mattison but what's so special about the guy that we can replicate?
You probably think Mattison is super aggressive, right? There's a stupid prediction below that Michigan's sacks would almost double and hit around 30th that hit almost on the nose. This is because Mattison, like all defensive coordinators ever, is much more aggressive than the last guy, the toff.
In Mattison's case, this is in fact emphatically true since the guy he's replacing was a milquetoasty has-been who thought sending four guys across the line was the ZERG RUSH of defensive strategies. This was obvious from the drop:
Last year I started tracking the number of rushers M sent at the quarterback because Greg Robinson kept sending three, which I defended as not totally insane at the time. Like everything else, it was totally insane.
The number of three-man rushes against WMU? One. That stuff about being aggressive that every defensive coordinator says? 100% valid. The really cool thing about being aggressive? Mattison is doing it while often getting seven guys into coverage by bringing zone blitzes.
But how many times did Mattison send more than the baseline four? A lot, right?
It's less than you'd think, I think. Via Seth:
|Opponent||Rush 3||Rush 4||Rush 5||Rush 6||Rush 7|
|San Diego State||3.8%||59.6%||21.2%||13.5%||1.9%|
Michigan sent a fifth guy about 20 percent of the time and more than that just 10 percent of the time. The other 70% they were just hanging out, and by "just hanging out" I mean "often still getting an unblocked rusher."
Mattison's near-constant zone blitzing is an NFL approach to defense, and once he figured out that Michigan couldn't do certain things quite as well as the Ravens it was a highly effective one. Despite having not much in the rush department other than Mike Martin, Michigan's pressure metric was on the whole a good one:
|1||WMU||15||5||10||Alex Carder is still coughing up blood.|
|2||ND||7||13||-6||Front four not getting anywhere.|
|5||MINN||15||4||11||NO BLOCKY FOR YOU|
|7||MSU||8||5||3||Not a lot of deep passes this week because of wind.|
|8||Purdue||9||1||8||Most of this a four man rush.|
|9||Iowa||8||6||2||Decent job; few blitzes.|
|10||Illinois||31||18||13||Stunts and okie annihilated OL.|
|11||Nebraska||9||2||7||Doesn't even count lets kill Martinez time|
|12||OSU||13||14||-1||Erratic, usually based on blitzes.|
And that's with no DL getting more than 5.5 sacks. Give Mattison some guys who can beat OTs one on one and those numbers will be more consistent—and very good. Viva okie.
That okie package was lethal, wasn't it?
breakdown of the different blitzes run from it. Chris identified six, seemingly all of which ended with Scheelhaase running for his life or losing it. Por ejemplo:
Now that Michigan is keeping a deep safety on these things and not offering free touchdowns—Mattison learned that lesson in one try—they are increasingly difficult to deal with as new players and stunts get added to them. It's almost like Scot Shafer was on to something.
This was all in year one, and now Michigan gets a year two.
Michigan's insane year-to-year improvement in 2011 leads to surprisingly moderate expectations in 2012. Michigan returns eight starters (three of them freshmen), and yet I get the vibe that no one expects improvement commensurate with what that would normally imply. The gap is even greater once you consider this will be Michigan's second year in a new system.
To this point, I haven't questioned that. The question marks on the line are severe and I've focused on that, especially in re: Alabama's offensive line, but… man, the difference between this defensive line and the kind of DB crater that submarined a bunch of previous Michigan seasons is enormous. If the line's even okay—like absorb blocks mostly—you have to think that improvement from the entire back seven will keep things level even accounting for the fumbles.
Quinton Washington is the key. If he won the job because he's pretty good, everything gets easier. If he won the job because Michigan has zero three-techs, it'll be a rough road heavily dependent on Blake Countess becoming a war daddy a year earlier than seems reasonable.
- JT Floyd > junior JT Floyd
- Blake Countess >> junior Blake Countess
- T. Gordon > younger T. Gordon + some Woolfolk
- Sophomore Jake Ryan >> freshman Jake Ryan
- Senior Kenny Demens > junior Kenny Demens
- Sophomore Desmond Morgan >> Desmond Morgan
- Will Campbell > Will Heininger (right?)
- Black/Clark/Beyer == junior Roh
- Kovacs == Kovacs
- Craig Roh < Ryan Van Bergen
- Quinton Washington <<< Mike Martin
That looks pretty good, right? Except for that last bit, yeah. But…
Last Year's Stupid Predictions
Courtney Avery busts out. Going into next year people are talking about him as an All Big Ten performer.
Kenny Demens leads the team in tackles with Northwestern-MLB-type numbers.
Yes on the leading the team in tackles, no on the NW LB numbers.
Brink is a legitimate player, better than Greg Banks was last year.
I meant JT Floyd, you guys. Srs. I meant Will Heininger, you guys.
The biggest source of pain on the defense is the WLB.
Craig Roh leads the team in sacks with eight.
Roh had four, which was tied for second with Kovacs. RVB had 5.5.
Sacks almost double from 1.4 per game to 2.4. That would be a move from 98th to around 30th.
Accurate. Michigan fished around with Herron and Hawthorne before settling on Morgan, who was iffy as true freshmen usually are. Michigan had 2.31 sacks per game last year, 29th.
Turnovers forced go from 19 to 27.
Michigan got 25 turnovers, though lady luck had a lot to do with that.
Michigan noses just above average in yardage allowed. Advanced metrics have them about 50th.
Exceedingly pessimistic, somehow. Michigan was 17th—Jesus—in yards allowed. 17th. Advanced metrics had them around the same spot.
EVERYTHING SEEMS WONDERFUL
I'm giving myself two points for this one.
This Year's Stupid Predictions
- Washington and Campbell are functional with Campbell leaning towards actually good as long as he's at three-tech and not tired.
- Fumbles are recovered at a 50.3% rate.
- The ILBs take a major leap forward as they understand the defense much better; UFR complaints about slowness are mitigated and evaporate by midseason. Morgan seems a little better than Demens by year's end.
- Ryan hits eight sacks and twice that many TFLs.
- WDE production is a major sore spot. Black ends up not very effective at either of his positions.
- JT Floyd is better tackling on the edge and otherwise static.
- Countess is a breakout star. Corner stats are dumb, so no predictions there, but his UFRs are an obvious step forward.
- Total defense (17) and scoring defense (6) drop somewhat; advanced metrics like FEI (16) hold steady, if not rise a bit.