Previously: Podcast 11.0A, Podcast 11.0B, Podcast 11.0C. The Story. Quarterback. Running Back. Wide Receiver. Tight End. Offensive Tackle. Interior OL. Defensive End. Defensive Tackle. Linebacker. Cornerback. Safety. Special Teams. 5Q5A Offense.

1. Are slot fades better or worse than slant vulnerabilities?

[Patrick Barron]

Michigan played slot coverage with outside leverage last year. This meant slants happened:

And after the slants happened Michigan spent a lot of time and energy making them stop happening without eating more slot fades. This necessarily meant taking the linebacker level away from git-em duty and using them as underneath zone defenders:

But mostly it was the coverage. Michigan shut off Wisconsin slants before they even began by playing man with linebackers in short zones. Bush gets a deflection on Wisconsin's first drive to boot the Badgers off the field:

That was Wisconsin's only slant of the game until their final play. That comes at a cost—Bush is not rushing, and he's not able to fire hard against the run—but it was clearly a good tradeoff to make.

That tradeoff did make slants a bad bet. By Northwestern this space was crabbing at folks to "cut that out" about slants since the Wildcats managed 5.6 YPA on them. Even so it was no longer clearly good by the end of the season. Perpetually-annoying Indiana was the tremor before the quake:

…non-Dwumfour DTs were complete non-factors. Four man rushes were set up to single up the DTs and then double-team Michigan's most dangerous rusher, which was Winovich for a while and Gary later. The Winovich cheapshot was an excellent example:

Both Gary and Winovich get outright doubled. Michigan has nothing resembling a complete DT this year so on a play like second and six Indiana can single up the DTs with impunity and Michigan can't play Dwumfour.

The UFR pressure metric was a slim +1, well below Michigan's standard under Brown. Indiana was not particularly good at avoiding sacks last year—41st in sack rate allowed. When the above cheapshot hobbled Winovich against OSU (the #17 pass-pro team last year) Michigan lost its best bet to Carter Coughlin OSU LT Isaiah Prince, and they almost didn't lay a hand on Haskins all game. You know what happened after that.

[After THE JUMP: Uche wally]

Outside leverage on slots had major knock-on effects. The impact on Michigan's linebacker havoc was apparent:

• Hudson/Bush/McCray TFLs, 2017: 44.5.
• Hudson/Bush/Gil/Ross TFLs, 2018: 21.

You can delete all eight of Hudson's TFLs from the 2017 Minnesota game and still come out with a near-halving of Michigan LBs' ability to chop guys down behind the LOS. Large chunks of the Hudson preview talk about how he's amazing as a blitzer and human as anything else, and how he didn't get to fling himself across the LOS nearly as much this year.

So, if you've got…

• Brad Hawkins, a really big, leaping slot defender
• Khaleke Hudson, a blitzer par excellence
• Shaky DTs
• A lethal rush package featuring Josh Uche

…doesn't it seem like the move is to live with some slot fades in exchange for a lot more second and ten? The fades irritated because there were about the only thing teams could get on Michigan. In a context where the D takes a step back they might be an acceptable cost for avoiding 15-play drives.

There's no indication Michigan has come to the same conclusion. Hawkins was in outside leverage in the spring game, and practice talk never gets in the weeds like this. Here's hoping that Michigan gets back to the LB havoc of 2017.

2. How many snaps can Josh Uche get?

this was a passing down [Patrick Barron]

There is a vision of the Michigan defense that dodges most of the defensive tackle issues in exchange for Josh Uche wreaking havoc. That is the most appealing way to put together Michigan's personnel: 4-3 defensive ends, a surfeit of linebackers, a deficit at DT. It's just that it doesn't work:

Is there one tiny thing that drives you completely insane beyond all bounds of reason?!?!?!?!?!

Michigan only did this once but the 3-3-5 on a standard down again got gashed. Marshall ate a –2 here but I sympathize with him because he's not a nose tackle in any way.

Michigan's defensive tackles were worse last year than they're probably going to be this year, and Uche was clearly a dude, and they still couldn't use a 3-3 stack on standard downs with any consistency. They haven't since the shocking Florida game where it debuted. And that was with Mo Hurst manning the nose.

This doesn't mean it can't. Michigan's half-assed the stack, and that's not something you can really do. Their approaches have not been dynamic enough; they haven't gone ass-backwards into hell like 3-3-5s have to. As this space noted after Northwestern, Michigan's stack "[isn't] delivering guys to the backfield like it has to if it's going to survive."

That doesn't mean it can't work. From time to time Michigan has delivered a 3-3-5 run D snap that makes sense. This is against Penn State:

The debut of the 3-3-5 was a skunkworks project that Michigan successfully hid from the world until gametime. It's possible that they've embarked on a similar effort this offseason. Likely, even. Offseason talk, discussed at length in the linebackers post, has held that 1) Uche is still a SAM first and 2) he's getting a lot more snaps.

Michigan's personnel may be more suited for the 3-3-5 this year than it was last year. The problem with previous editions is that they always used Noah Furbush as the designated bonus DL the 3-3-5 inserts into the line from the linebacker level. He was really effective as Devin Bush's fullback until opponents figured it out. Then he was not. Ross is bigger than Bush was and is more plausible as the bonus DL; Uche is up to 250 and will bring oomph if he's called upon; both DEs are 280 pounds and have a better shot at holding up when they have to slant to the interior; Hawkins, a burly safety, can spin down into the "bandit" role much more effectively than Watson hypothetically could.

If Michigan commits to it, it could work. Not against Wisconsin or MSU, but against a lot of teams. Watching it develop will be one of the most fascinating subplots of the season.

3. What are we supposed to do about crossing routes from 4.3 guys?

Play inside leverage, vary coverages, and get to the quarterback.

The fatal trap Michigan fell into last year was that Chase Winovich would make everything right on standard downs. Winovich got jumped on against Indiana and was not right. Nobody touched Dwayne Haskins, and this was the context in which Brandon Watson's lack of athleticism—previously a minor issue overcome by chipper effort—was ruthlessly exposed.

I'm torn between WE DIDN'T LISTEN because when Watson didn't get all of a guy at the LOS it was over…

CB #28 to top of screen

…and the shape of the pass defense before OSU. On the one hand, it's difficult to get worried when you're the best pass defense in the country. On the other, some asshole on the internet after the Indiana game:

I did get a bit frustrated by Michigan's seeming over-reliance on man coverage. There were a number of long-ish catch and run events that were super easy because Michigan showed man coverage and ran it.

CB #28 to bottom

That is a super easy third and 15 conversion, and a bit ominous for OSU. This is a super easy third and seven conversion, same:

Haskins hit those crossing routes late in last year's game and their offense this year is Indiana with faster WRs: Haskins has the second-shortest average depth of target in the Big Ten and OSU has the most yards after catch in the league. As PFF details, this matchup is all about yards after the catch. But also Michigan is 4th in the country at preventing YAC despite the above. Our perceptions are all out of whack after watching most of this season.

Wee bit frustrating, that. And it was all the more frustrating because there was a solid chunk of the season when it looked like the year three version of Brown's defenses was developing into a man/zone hybrid chameleon. A big chunk of Michigan's early struggles against ND were because they were predictably in single-high man coverage:

…Notre Dame [attacked] Michigan's man coverage early. Live I thought someone had busted a coverage on ND's first third down, but it was an RPS play where Hill gets (legally) rubbed off. There were a couple more early pass patterns that picked or sought to pick off M.

Another version of this attacked Michigan's tendency to go single high. Here Hawkins is the free safety; ND runs one of those zone-read-to-speed-option plays that are becoming the rage against shuffling DEs. Winovich gets edged here and that's all she wrote:

DE #15 to bottom, also #20 FS

You can defend that from single high but you have to get a good play from your DE and that'll weaken your ability to defend zone read belly stuff, which ND ran with some frequency. That's a luxury Michigan had last year with Mo Hurst. Less so this year.

Several other ND plays attempted to attack this same area, but after Michigan had moved to two high looks they got about three yards each. The RPS in this game started out like –7 but after Michigan went to more zone/two high looks it got back to balanced by the end of the game.

Michigan seemed to take this to heart and spent a big chunk of the season mixing their looks up. After Nebraska I asserted the "main development this year has been the increasing deployment of two-high zone looks"; after Northwestern I thought it was "another 60-40 man-zone ratio… it's getting mixed up." And the Wisconsin game featured a critical interception off Michigan's zone:

Michigan's other tactical win in this game was trap coverage, which forced the first interception…

…and also got Long another PBU. Those are huge wins that most QBs probably fall for. You'd have to be a savant to spot the trap presnap. Hornibrook could never be sure what he was looking at and frequently had to hit a small spot as man coverage was fronted by Michigan's ILBs. This led to a period of indecision from Hornibrook, and led into the final thing that blew his day up.

And then Ohio State has an opening drive that is approximately 500% uncontested crossing routes, leading to disaster. It was the most dispiriting thing to happen in a long, long history of dispiriting things against Ohio State: literally the top defense in the country rolls into The Game and gets hammered like Greg Robinson is still running things.

Hopefully that's a wake-up call. OSU probably can't replicate last year since Michigan will get more pressure and Fields is not likely to play like a first round QB as a true sophomore who just transferred, but the defense should be structured around combatting whatever OSU's doing, and if that's suboptimal in other circumstances the offense will make up for it. So: make Fields make tough throws to the outside, give him a bunch of looks, and stick Josh Uche in his teeth.

4. Does it make sense to, you know, back off?

the heck you say?! [Patrick Barron]

Secondary concerns have caused some Michigan fans to ask me if it makes more sense to go to more of a bend-but-don't-break approach this year. This is never going to happen under Don Brown. Put it out of your mind. And I'm not even sure that would be a wise approach in the event Brown's brain underwent Gephardtization.

Hudson is best blitzing. Uche is best blitzing. Michigan's issues late last year were largely because the defense had boxed itself into a situation where they were unable to force their way to the quarterback. The secondary has some questions but still has Hill and Metellus and will get Ambry Thomas, who is super fast, back pretty soon. Dax Hill's probably going to emerge midseason.

So the secondary is fast and the second level is best blitzing. Bombs away, I say.

5. Well?

I want to say this will be a clear step back from last year, but when I say that I'm thinking about the defense pre-OSU. Once the last two games were taken into account Michigan's defense went from vying with Alabama and Clemson for the #1 spot to ninth. It's hard to think about a baseline for the Michigan defense because there was no middle ground.

So: the defense will be better than it was against OSU but worse than it was for much of the rest of the season. The DL should be about as good, with improved play from the DTs offsetting a step or two back at DE. The linebacker level will miss Bush but only take a step or two back itself since it's got two guys who are clearly players and four candidates to fill the WLB spot productively.

The secondary? Er. A ton depends on Hawkins, or failing that, Dax Hill coming through. Two shots are better than one.

Michigan's not going to hold everyone under 200 yards again; there will be blips and bumps and the offense will have to do some outpacing in certain games. But there are few teams in the country who can match Michigan's front seven, and that should be enough to keep them within shouting distance of elite.

BETTER

• Healthy, older Dwufmour >> Hurt, terrible run D version
• Older Gil or someone replacing him > Gil/Ross platoon
• Jeter/Mason/Hinton/Smith > Mone/tiny bit of Solomon
• Carlo Kemp > Carlo Kemp
• Dax Hill/Hawkins > Tyree Kinnel

SAME

• Khaleke Hudson == Khaleke Hudson
• Lavert Hill == Lavert Hill
• Josh Metellus == Josh Metellus

WORSE

• Paye/Danna < Winovich
• Hutchinson < Hurt Gary/Paye
• Josh Ross < Devin Bush
• Vincent Gray/colitis Ambry << David Long
• ??? < Brandon Watson

LAST YEAR'S STUPID PREDICTION

Michigan has the #2 S&P+ defense.

Guh. This was on point until the OSU game and subsequent who-gives-a-crap bowl performance that shot them down to 9th. Half-point.

Josh Uche has a breakout year with 7 sacks in relatively limited snaps.

I get two points for this one because his snaps were incredibly limited and he still led Michigan in sacks.

Gary has his Ndamukong Suh year, but there's a fairly large chance he sees his counting stats stolen away from him by his teammates too much to really shock the world.

Injury prevented this from happening but even accounting for that this was nowhere close. Zero points.

Hudson is a better player with less statistical impact because he blitzes less and drops into coverage more.

"Better" didn't bear out in UFR grading but the overall shape of his season was correct. Half point.

Bush, Hill, and Gary are All-Americans of some variety.

Bush, yes. Gary, incomplete, missed too much time to be an AA. Hill, no. Half-point.

Solomon is better than Dwumfour; both guys have slightly worse 2017 Gary-ish seasons where they do many productive things and seem poised for a monster 2019.

Incomplete due to injury for both players.

Metellus and Kinnel still come in for a lot of criticism largely because the rest of the defense is hard to critique. Both are noticeably better; Kinnel still doesn't get drafted.

Pretty much. Metellus is badly underrated by the fanbase. One point.

THIS YEAR'S STUPID PREDICTIONS

• Aidan Hutchinson has the most TFL output amongst Michigan DL.
• Gil gets passed by McGrone midseason but hangs on to a rotation spot. This space grumbles about it.
• Metellus matches Delano Hill as a third round pick.
• The 3-3-5 is a regular defense on standard downs in addition to its omnipresence on passing downs, allowing Uche to hit double-digit sacks.
• Dwumfour takes a big step forward; Kemp is pretty static. Jeter feels like he should displace Kemp as a starter but settles for being a major rotation piece.
• Dax Hill has to settle for being the nickelback.
• Michigan hangs on to a top 10 SP+ rating by the skin of its teeth.