that makes one of us
Okay, so we've just taken a spin through Michigan's hitch-susceptible cover three. What's the alternative? The Big Ten Network did give us one replay of something different. It comes on third and long in the third quarter with Michigan on offense. Threet is going to throw a hitch to Darryl Stonum.
(This camera angle is wider so I've cropped the pictures down; it'll be slightly fuzzy.)
This is pretty much the same as the first frame we saw with Morgan Trent earlier, except the defensive back is lined up a little further inside.
The Purdue corner turns his hips towards the receiver and looks directly at him. This is man coverage.
When Stonum starts making his break it's immediately apparent to the corner and he can turn his hips towards the receiver. Trent had to turn around the other way, taking himself away from the receiver until he can get his body around.
The ball from Threet is low and inside and gets marked IN in UFR, but Stonum still has a chance at a diving grab…
…but the Purdue defender is right there making life difficult. If the ball was better thrown he had a chance at a breakup or even an interception.
Here's the video on this one:
So it's pretty obvious why Purdue was able to play this sort of tight coverage on third and long: Michigan's receivers don't seem like much of a downfield threat and this guy has safety help over the top, a luxury not afforded Michigan's corners on the regular.
Okay. Question: does Michigan's coverage style (here's another example from Cissoko) make any sense given the defense they're running? I'm not a coach, but doesn't it seem like Michigan corners should be setting up outside of the receivers and funneling them towards the safety and/or zone dropping flex players and linebackers? Maybe Shafer thought Siller couldn't throw well enough to make him pay. He was wrong.
So, yeah, the 3-3-5 stack was sort of a spectacular failure. Not so much against the run, as Michigan did a decent job against Korey Sheets, but against the pass. Michigan allowed 266 yards—7.8 YPA—to a third string quarterback on a 2-6 team. That is epic failure.
Why did this happen? Well, IMO, the stack had a a major part in it. Here's a typical presnap alignment:
The thing to note is the one deep safety. This is Donovan Warren. With one deep safety the corners are basically on an island; they won't get much deep help from Warren on a sideline route. Fades and fly routes are going to be left up to the corner to defend.
Michigan, then, often sees itself play a cover-three susceptible to little hitches and out-cuts. Here's a closeup of Morgan Trent on one of many, many ten-yard outs that Purdue will find wide open all day:
We've just started the route. Greg Orton is going to run directly at Trent.
Trent is turning his hips inside and looking at the quarterback.
Here we see Orton reaching Trent, with Trent turned completely away from him, not even looking at him.
Orton begins his break with Trent facing 180 degrees the wrong way.
Orton now has two yards of separation and Trent is still facing the wrong way.
The perspective on this makes it difficult to tell, but this is wide open for ten yards.
Here's the video:
Michigan's corners have been coached to do this. Trent would do it all day, and Boubacar Cissoko did it on the long Orton completion down the sideline. This coverage style covers that extremely well and the only thing that made that a completion was a perfectly thrown ball, some bad luck on Cissoko's part not to knock it down, and a great catch by Orton.
However, every wide open hitch or out was because of it, too. It's completely impossible to flip your hips around 180 degrees that quickly, impossible to do anything but tackle after a first down. So Michigan provided this n00b quarterback with the easiest reads in the history of reads most of the day, and ended up paying for it.
Okay, so. On the last play Michigan picked up three yards when Brandon Minor cut behind Steve Schilling and got tracked down by the backside defensive end.
On this play Michigan returns to a more conventional formation. They're going to run the exact same play:
Again, the key block on this is the Schilling-Moosman double on the playside DT. Getting him blocked and Moosman into the linebackers is win.
Okay, Moosman has engaged with the backside DT and is actually driving him back off the LOS a bit. He didn't get this much push on the last play. Schilling, also given the humiliatingly amateurish Paint Arrow, is hustling to get into position.
Schilling's still rushing to get there, but on this scoop we see Moosman already disengaging to get to the second level… look at him facing downfield, not towards the guy he's nominally blocking. You can see Schilling's knees buckle.
Schilling intends to cut the DT. For this to be legal, Moosman can no longer be engaged with him; this block requires precise timing.
Moosman is away, again with great angle to block a Purdue linebacker who's got no idea what's going on. You can see the frontside DT shooting hard to the playside. There is about to be…
…one hell of a crease. Minor cuts up. Note that the linebacker Moosman had such a great angle on has decided to go around the other way. The backside DE is still chasing, but Minor's cut isn't taking him back to the right enough to be caught.
Minor bursts into the secondary for 21 yards. The backside DT, nearest to you in the picture, is still getting up. That linebacker who eschewed contact is waving at Minor's legs as he passes.
Object lesson on this one is: cutting a defensive tackle to the ground is a good idea, if you can do it.
Also, it is sad that Michigan's running game commands so little respect that Purdue has six guys in the box on second and seven. Four wide and all that, I guess, but still.
On Saturday, Michigan faced third and three and, for probably the first time in 20 or so years, called a designed quarterback run. Here it is:
Okay, empty backfield and wide splits on the defensive line. Seems like a pretty good setup, but there is one issue: this play is designed to go between the DT to the right of your screen and the DE to the same side. Without a lead blocker one of those linebackers will nail the slow-ish Threet before the marker.
To allay this, Michigan is going to try a reach block by Molk on the DT, which will allow Moosman to head downfield on the linebacker.
What's a reach block? Uh… well…
Using the left guard again, to “reach” would be to get around the defensive tackle, and use his right shoulder to pin him to the inside, so that a ball carrier can go around you to the left. Again, it is about getting the face mask in “front” or beyond the defender to get the shoulder pad in position. Seriously, line up with a friend sometime and try to reach block to your outside, you will appreciate linemen athleticism much more.
The idea is to get Molk around the defensive tackle so he can seal him, creasing the two defenders, as Moosman heads downfield to take out a linebacker. If this sounds hard, it is. I lost this in the ether, but at one point during my research for Hail To The Victors 2007 I came across one coach's description of a bunch of different blocks, ordered by difficulty. "Reach block by the center" was #1.
Real UFR diehards may remember a common bitch from the last couple years that usually went something like "Kraus attempts to block a DT lined up playside of him, but he shoots into the backfield/flows down the line to tackle/eats a baby." These were all attempted reach blocks gone bad.
(A "scoop" block, as I understand it, is basically an assisted reach block. Moosman banged the DT back and helped Molk get over, that would be a scoop.)
And all this stuff is supposed to be hard when the DT is lined up to your outside shoulder. Here the DT is lined up slightly outside of the guard(!). How is this going to work?
You can see the line shifting to the left here, and you can see that the DT is now between the two OL. The wide splits were a pass rush gambit—tougher to block outside that way—and the first step of the DT is upfield, not down the line.
Molk makes contact and he's in decent position here given the relative momentum here, but he's still got to get his helmet across the player, then anchor as well as he can to preserve the crease between the two OL. Chris Spielman, by the way, is currently doodling on the DE, who is still in pass rush mode.
Molk is now full of win, playside of a guy who lined up a yard outside of him at the snap. Moosman is in great position to block the MLB, but doesn't have to because he's getting cut to the ground. Minor is about to block the safety-type object.
Woop! Open spaces, first and goal, and a one-yard Minor touchdown follow.
Object lessons. I picked this play out of all the various things for a variety of reasons. To wit:
I think Molk might be pretty good once he is enormous-er. I brought this up earlier in the year, but Molk was a fringe top-100 guy who was the only real OL recruit brought in after the shift to zone blocking. He got dinged later in the year for being small, but in a system like this where he's reach-blocking all day his agility is an asset. Time and again against Penn State he successful executed these blocks, springing people into the secondary. Against Notre Dame he did the same thing.
The issues are obvious, though: too many missed blocks, and too many blocks where he's just not strong enough to deal with his man. But he's a redshirt freshman; strength should come.
(This is the long way of saying I think GS was unduly harsh on Molk this week in the Run Chart; he should get more credit for these reach blocks.)
You can only make a reach block if the defense lets you. I'm not a coach or an expert or anything but over the last three years I've watched a ton of stretch plays and have come to the conclusion that if the DL steps the right way and you have been tasked with a reach block, you lose.
And the thing is, either way can be the right way. Last year Penn State's Ollie Ogbu had three TFLs and a half-dozen more plays he forced into unblocked defenders because he was shooting behind the attempted reach block. Penn State slanted their DL all day, and if they got a zone left they strung it out and if they got a zone right they came under it and did even more damage.
Diversity. The reason Michigan's run game was so successful against Penn State was because of its diversity. For much of the first half, Michigan had Penn State defenders expecting stretch and getting something else.
The results are, for the first time, encouraging. The rushing game against Penn State this year and last, sacks excised:
|Year||Carries||Yards||Avg.||Opp Rush D|
Some of that improvement is the decline in Penn State's defense, but raise your hand if you think the Penn State defense declined more than the Michigan offense.
Right, no takers.
How? Well, I found a three-play sequence on Michigan's first touchdown drive interesting. Michigan had been moving the ball and found itself in fourth and one. Penn State slanted into the backfield and should have had Minor(+2!) dead; Minor squirmed out and got the first. On the next two plays, Penn State went back to the slant—back to the successful gameplan from a year ago—and got cut for a total of 14 yards and a first down because Michigan ran the same play you see above and that backside veer play. Michigan had Penn State guessing in a way that Carr never did, IMO, and that's a large reason why WVU's ground game was near the national best in YPC.
Of course, all that died in the second half, but there's only so much diversity Michigan has at this point. If they had a reliable passing game (read: Threet with elbows) or a better offensive line or some rocket quarterback they'd be able to punish Penn State's adjustments to their run game; as it was they just ran out of things to do.
It might be silly to highlight a play that worked in a game where your offense scores three net points but I thought this was an interesting play, and it's one we haven't seen this year. It's third and one on Michigan's second drive of the game; Threet's about to throw the disaster interception.
But first we have to get that first down. Michigan comes out in an I-formation(secondary offense what!) with twin receivers to the bottom of the screen. Also, note the position of TE Kevin Koger: he's covered up by the outside receiver and can't go downfield.*
This next frame is taken right at the snap and exists to show you the presnap motion of the fullback: he took two quick steps to his right the instant before the snap. Michigan used to do this all the time under Debord and it drove me crazy; in this instance the fullback shuffle is 1) much quicker and 2) followed immediately by the snap. It's still a tipoff, but less of one.
The DL didn't pick up on it, obviously: they're slanting inside.
The Rockets are in trouble at this point: check Molk out: he's got a complete seal on the playside DT. This leaves six Michigan blockers (WR, RG, RT, TE, LG attempting to cut the MLB, as you can see) in an area with four Toledo defenders. Five of them can get excellent angles on their players; only the LG has something of a difficult job.
This hole is truly gaping as Grady approaches the LOS, but the LG has failed to chop that linebacker, leaving him to Moosman, and Moundros is kind of running aimlessly outside.
Our final frame shows Moosman having whiffed on the MLB, Grady past the first down marker, and backside pursuit encroaching. Moundros is still looking for someone to block.
So: a great playcall that caught Toledo's defense, particularly the playside DT, in a play they weren't prepared to defend. Good execution by Molk and Schilling creates a cavernous hole in the line with plenty of blockers heading downfield, but poor execution by those downfield blockers held this gain down to about six; club that MLB and he's probably down to the two or three and Michigan grinds in a touchdown.
*(This bugs me way less than the formation where Greg Mathews is split out and covered up because if Michigan passes they can still use Koger as a blocker; the Mathews thing is basically playing 10-on-11 if you pass.)
We're back with a trips formation.
Check out that guy lined up outside of Mathews way out there: that's a linebacker. He is on the LOS outside of a couple wide receivers. Weird.
Michigan will run a zone stretch, and Illinois will do what they did:
Again we see the backside defensive end crashing in with no thought for Threet. He can do this because Miller is shooting out to contain the QB, like we saw on the last play.
Also, check the linebacker, already flowing upfield as soon as he sees the play start.
Check the very top of the screen: if Michigan was to throw a bubble screen Odoms would probably get lit up, as the linebacker has used his advantageous position to hop past the wide receiver. This defense takes away all three prongs of the Rodriguez system.
But it doesn't take away everything. Look at the vast huge gaping hole between that linebacker and Miller. If Odoms was just to not go anywhere, or if he was to run a little hitch to the line of scrimmage, he would be wide open. Illinois's secondary is playing in the parking lot. If Odoms was to come in at the snap and establish a pitch relationship with Threet, there's no one to cover him on the option. This is not defeat for the zone read; it's just defeat for Michigan's zone read at this moment.
Anyway, McGuffie is forced to cut back because Schilling's been driven into the backfield…
… and gets smooshed.
The instant conclusion to jump to is that Rodriguez got beat. Illinois regularly deployed this gimmicky defense specifically aimed the staple of his offense. It has big gaping flaws in it that Michigan did not adjust to. They assumed Michigan would run on first down, put themselves in a position to stop it, and did; Michigan did not go to a throwing mode.
It's not that easy, of course, when you're a coach at a new school and you're using freshmen everywhere. Illinois, like Oregon, has guys who have been in this offense for multiple years and they pulled out a vast array of sleight-of-hand. Michigan probably doesn't have that rabbit in its hat yet.
Still… doing the UFR here was a sad trip into the recent past. At this point we know the line is better at pass blocking than run blocking—though Ortmann was way worse at guard than he was a tackle—and that teams are teeing off on the zone-read/bubble on first and ten; too rarely did we go away from that.