Monday, November 29, 2010

Ohio State: Where whatever

Michigan's season is not so mercifully over. There were highs and lows and eventually, Michigan ended up about as good as everyone thought they would: 7-5 with a terrible defense and a great offense. No one expected the offense to be quite this prolific or the defense to be quite this bad, but Michigan fared about exactly as they should have.

Saturday's game, though pretty demoralizing, shouldn't change anyone's opinion on Rich Rodriguez. Three of Michigan's five losses came against one-loss, BCS top-10 teams. One came on the road against a resurgent Penn State team and the other against a middle-of-the-road Iowa team.

Asking that Michigan be more competitive against Ohio State sort of misses the point. Were it not for a 170-pound Ray Vinpoal being a 170-pound true freshman that gets trucked by 200-pound junior running backs despite being in perfect position, Ohio State averaged 3.8 YPC. Terrelle Pryor looked, well, like Pryor, and not some manfreak that the defense turned him into. Michigan's four-man defensive line was functional. The offense only stopped when it killed itself (which was frequent). It was a lack of execution and inexperience that made this game uncompetitive, not schemes, raw talent, or fluke plays. Ohio State is basically perfection in terms of doing what they're trained to do. Michigan isn't yet. Whether that's a function of starting mostly underclassmen or poor coaching is a debate, but I'd be willing to bet it's 85% inexperience and 15% coaching.

I have very little else to say on this topic right now and am in the middle of a conference for work, so content on the game will be scarce until Thursday or Friday.

  • I feel at least somewhat confident saying this: Greg Robinson will be your defensive coordinator next year. It won't make you want to die. I think there's been enough improvement on the defensive side of the ball in the last few weeks that Robinson, working at the talent disadvantage he is, may be given another shot.
  • Roy Roundtree will be wearing #12 next year.
  • Vincent Smith will be your starter in the bowl game and 2011.
  • The defense looked good in this game until it was clear that all of the momentum was going Ohio State's way. Michigan failed to score in the red zone twice early, and that's when it all started going downhill. You just can't give a team like Ohio State more chances than they already have.
  • Denard finally scrambled and it was awesome. Except when he fumbled and it wasn't.

Wednesday, November 24, 2010

Holidays and real-life work

It's going to be a slow week of posting despite the fact that it's Ohio State week. I'm traveling all day today to spend Thanksgiving in Chicago, followed by an imaging trade show that starts Sunday for work. I'll try and post things when I can, but Internet will be limited as will time. I've got one more play to add to the Swallowing Kenny Demens debate, but that likely won't be posted until tomorrow.

UPDATE: The play I had planned on discussing tomorrow was actually Picture Paged by MGoBlog earlier today. I disagree with Brian's diagnosis of this play as well (at least slightly), and voice most of those concerns in the comments section.

Tuesday, November 23, 2010

Swallowing Kenny Demens addition

Today, MGoBlog Picture Paged a Wisconsin run demonstrating a flaw in the Michigan defensive formation:
by lining up his MLB just behind his nose tackle he dooms that guy to take one step to the playside, whereupon he is eaten by a guard who has no one lined up over him. Even if Michigan successfully plugs that hole they are crazy vulnerable to cutbacks and counters.
I don't think this is exactly right. The problem with the play is not necessarily that Demens is too close to the line of scrimmage, rather that Michigan's defensive numbers and zone coverage dictated not only a decent gain but a tackle by the safety.

MGoBlog's complaint is that Kenny Demens is lined up directly over Mike Martin and only two yards off the line of scrimmage. The bigger problem, however, is that Michigan is playing cover-3. Like Chris Brown of Smart Football likes to say, "football begins as arithmetic" (with depressing Michigan/Indiana tie-in!). Before this play even starts, Michigan is committing three deep safeties to an I-formation with two receivers--Courtney Avery, on the weakside of the play, is rolled down toward the line of scrimmage but because there are no receivers on his side of the field. Though it might look like more, Michigan is committing only six defenders to the box against seven Wisconsin blockers and a tailback. Cam Gordon and Jordan Kovacs are outside the box and covering flat zones (sort of, about which more later).

Important to note: on the strongside of play and inside the box, Michigan has four defenders (Martin, Roh, Ezeh, and Demens) versus four blockers (center, right guard, right tackle, and the fullback). The numbers are decidedly in Wisconsin's favor.

After the snap, the center and left guard double team Martin. The right guard is given a free release where Demens and Ezeh are attacking. Van Bergen is being double teamed by the left tackle and tight end. He gets sealed to the outside but understandably so. He has no shot here. There are no problems with Michigan's defense yet, but Jordan Kovacs standing with his hands on his knees becomes a huge problem in a minute.

Demens and Ezeh are running to the point of attack. The right guard, who was given a free release will be hit by Demens a yard past the line of scrimmage and Ezeh will attack the fullback.

This is where my understanding of this play varies from MGoBlog's. Brian argues that, were Demens further off the line of scrimmage, somehow, the offensive guard who was given a free release wouldn't find him or would miss his block. Frankly, that's wishful thinking. Demens plugging the line at the point of attack by taking on the offensive guard is actually the best possible scenario. Ezeh will be charged with stuffing the fullback, which will necessitate either a cutback or a bounceout. Structurally, Michigan is still sound.

The bigger issue with this play is that Michigan is playing zone coverage. Mouton and Kovacs are still frozen in place because they have to account for the pass. In Mouton's case, this is a function playing zone coverage and being assigned to the middle of the field. Kovacs, however, should be ripping down the line of scrimmage behind the offensive line right now. He can do this because Wisconsin didn't put any receivers on the weak side of the field allowing Avery to roll down and cover the overhang. So when the left guard releases his double team on Martin and starts to head upfield for Mouton...

...Kovacs needs to be rushing into the backfield to make the play on a cutback that has been forced by Ezeh and Demens stuffing the Wisconsin blockers at the line of scrimmage. Instead, Mouton has to take on an offensive guard, leaving the middle of the field open for Wisconsin:

Now look at Kovacs and Avery. What is Kovacs doing here? He's unblocked, isn't defending the overhang (that's Avery's job), and watching the play move away from him. If Kovacs had started attacking the backfield when he should have, this would've been a two yard loss. Michigan's defense is still sound if he misses: even a miss would necessitate a cut to where the play eventually goes, and Avery has overhang coverage. I think Kovacs just flat missed his assignment here and is frozen because he thinks he has a responsibility in the passing game (he doesn't as there are no weakside receivers and Avery has coverage on the weakside TE should he release).

Kovacs begins to attack the backside of the play when the ball is already beyond the line of scrimmage. Mouton has been successfully blocked. Avery, who was in overhang coverage is late getting back to the point of attack, which is OK because he needs to be behind the ball. And eventually Ray Vinopal will come downhill from the free safety spot to make the tackle.

The point of this is that it's not the alignment that's the problem. Moving Demens further from the line of scrimmage does not somehow remedy the fact that Wisconsin has a decided numbers advantage. Nor does it mean that he'll be able to read and react to a play that's designed specifically to flatten him. (What more could Demens do on this play from four yards deeper? Expecting him to completely shed a free-releasing offensive lineman is unrealistic and assumes that Michigan's players are supermen.) For this play to work, Wisconsin actually only had to block six of Michigan's defenders: Roh, Martin, and Van Bergen (all on the line), and Ezeh, Demens, and Mouton (the linebackers). Michigan essentially committed five players to the pass (the three deep safeties, along with Kovacs and Gordon at the Spur/Bandit positions). All of this despite the fact that Wisconsin had seven blockers.

This could have been remedied had Kovacs realized the shift in coverage when Wisconsin didn't put a receiver on his side of the field. But he failed to acknowledge it and instead completely took himself out of the play, acting as a redundancy overhang defender. So while I agree that this is probably an "incoherent defense", it's not the alignment of the linebackers on this play that causes the problem. Michigan is unwilling to test man coverage and, because of that, find themselves spread out and susceptible to the rush.

Zone blitz: Proof of defensive... coaching?

Michigan's defense didn't do much right against Wisconsin. They couldn't stop the run despite Wisconsin passing a grand total of one time in the second half, and couldn't stop either in the first half. And while some of this can be attributed to the dearth of talent on Michigan's defense, a fair share of blame needs to fall on the defensive coaching staff (per usual). But at least once, the coaching staff designed a play to get a free blitzer on Scott Tolien.

Wisconsin is in 3rd and nine on their first possession of the game. They're showing a four-wide set with a running back in the backfield to help block. Michigan is in their 3-3-5 front and showing blitz. Thomas Gordon, Kenny Demens, and Jonas Mouton are all on the line of scrimmage showing blitz. Cam Gordon and Jordan Kovacs are both two yards from the line of scrimmage and look like they're going to blitz. James Rogers, Courtney Avery, and Ray Vinopal are showing cover-3.

Immediately after the snap, all three down linemen, as well as Mouton, Demens, and T. Gordon show blitz. Kovacs is also blitzing from the second level. Cam Gordon is sitting in underneath zone coverage.

A moment later, Mike Martin (red) and Kenny Demens (blue) are dropping out of the blitz and are going to settle into underneath zones. Still blitzing are both DEs (Roh and Van Bergen), Mouton, T. Gordon, and Kovacs. Most importantly though, Van Bergen (bottom of the screen) slanted into B-gap (between the offensive tackle and offensive guard), forcing both of them to block him. This helps free up two blitzers (Kovacs and Mouton) on only one blocker (the running back).

You can see here that the running back has taken on Mouton, giving Kovacs a free rush up the gut. The offensive tackle realizes that he missed the Kovacs block and was successfully baited by Van Bergen. He's reaching back to get a hold of Kovacs. You can also see Demens and Martin settling into underneath zones just beyond the line of scrimmage. It's important to note that Michigan is rushing five men against six blockers and is still able to get a free run at Tolzien.

Kovacs is untouched.

Sacksauce. Martin and Demens are sitting in underneath zones. Mouton has pancaked the Wisconsin running back. And Wisconsin is using four blockers to stop three blitzers while the left tackle is lost in space, and Kovacs has a free run at Tolzien.

The defensive coaching staff deserves credit when they do something good. This is just such a play. While a lot of the defensive struggles this year can be attributed to schematic breakdowns brought on by poor play design, this is a play in which the coaching staff was able to manipulate the offensive line with a zone drop and get a free release on Tolzien despite blitzing five men against six blockers.

This worked because Michigan was willing to sell out on the blitz and live with whatever happened, good or bad. The only way Michigan can realistically run plays like this is when they get teams in third-and-long situations: if Wisconsin gets a big completion here, it's just a first down, something that Michigan's base defense is unlikely to stop anyway. But given the potential that the blitz brings, this play becomes a really good call in this situation. Why the defense doesn't try similar plays more often or simply blitz linebackers more frequently is beyond me. Sitting back and waiting for quarterbacks to pick apart Michigan's defense is just as dangerous as selling out on the blitz--it should be noted that if Michigan had more talent on the defensive side of the ball, this likely wouldn't be true. If Michigan wants a chance at beating Ohio State, running complex zone blitzes like this, and landing them, will be essential.

Monday, November 22, 2010

Wisconsin: Where it went about as expected

Were it not for a ridiculous implosion against Michigan State, Wisconsin would be sitting alongside undefeateds Auburn and Oregon, hoping that they'll be in Arizona for the national championship game. Wisconsin is that good. It's not surprising, then, that Michigan struggled so mightily against an oversized, veteran unit that is unquestionably the best outfit in the Big Ten. That doesn't make Michigan's complete inability to stop the run or their general defensive incompetence any less disappointing, but it does mitigate it somewhat.

By the second half, Wisconsin had completely abandoned the passing game, but it didn't matter. The Badgers stable of backs averaged 6.2 YPC on an astounding 58 carries. Scott Tolzien, meanwhile, went 14/15 for 201 yards--his only incompletion was an interception to James Rogers which sort of stopped the bleeding at the end of the first half. But again, none of this surprised anyone. Tolzien is a senior quarterback with all the talent to torch Michigan's secondary. And most of their starters are 4th- and 5th-year seniors. If you could describe the perfect storm for Michigan, it was Wisconsin.

The more bothersome aspect of the game was Michigan's offense getting shut out in the first half. Denard overthrew a clear touchdown pass and Broekhuizen missed yet another field goal. So there were points to be had. But in general, Michigan's offense was thoroughly stymied in the first half. I'll have to rewatch the game to see if Michigan made any changes in the second half or just executed better, but the offense finally got going behind the legs and arm of Denard.

  • Speaking of Denard, this is the best he's looked since before Michigan State. He finally got the ground game going and his passing looked much improved. The interception was decidedly not his fault. JJ Watt had basically a free release to bat the ball into the air. Not to mention that the play before Roy Roundtree had an egregious first down drop. (This is the second time that the play directly following a Roundtree first down drop that Denard has thrown an interception.)
  • Receivers: Darryl Stonum had, by far, his best game this year and possibly his best in a the Winged Helmet. That over-his-head catch that he made at the four yard line was utterly ridiculous, and he was getting separation on cornerbacks all day. Roundtree, on the other hand, is proving a bothersome trend of dropping sure catches. I don't know why this has started to happen, but every week, it seems like we can count on at least one Roundtree drop.
  • Injuries to Roh and Martin are bad news for Michigan's defense. Losing one or both for the Ohio State game spells certain doom.
  • Vincent Smith looked like he had another good game, but the box score begs to differ (50 yards on 11 carries). Stephen Hopkins, meanwhile, should probably start losing some of his carries. This is the second straight week that he's looked overmatched and unable to make proper cuts.
  • Missed tackles, argh.
  • Jeremy Gallon, argh. Yet another game that Gallon looked terrible on kick returns, punctuated by the injury/fumble. It was amazing to see the different between Gallon and when Stonum/Drew Dileo started returning. The latter two were both able to make good cuts and get upfield in a hurry. I have no idea why Gallon is allowed on the field.
  • Michigan's handling of the end of the first half didn't really bother me. Given how the offense was moving the ball in the first half, not taking shots downfield was fine. Especially when it was clear that, if given any time at all, Wisconsin's offense would march down the field on Michigan. An interception probably heeds more points for Wisconsin.

Friday, November 19, 2010

Wisconsin preview

After locking up a bowl game against Illinois and a winning record against Purdue, Michigan heads into its unquestioned most difficult two-game stretch of the season against Wisconsin, followed by The Game. If things seem grim, it's because they are:
Passing28th nationally
190.50 YPG
74th nationally
200.30 YPG
Rushing25th nationally
125.10 YPG
12th nationally
228.20 YPG
Scoring23rd nationally
19.50 PPG
8th nationally
40.20 PPG
Passing114th nationally
270.20 YPG
31st nationally
255.50 YPG
Rushing75th nationally
163.70 YPG
9th nationally
266.30 YPG
Scoring93rd nationally
32.10 PPG
14th nationally
37.70 PPG
The thing that sticks out in these numbers is Wisconsin's passing ranking. Though Scott Tolzien has been incredibly efficient this year, Wisconsin is not necessarily putting up the bulk numbers you might expect. And while that may be encouraging, Wisconsin is still Wisconsin, and their running game is as strong as ever.

Wisconsin boasts a stable of freakbacks like they usually do. The headliner is John Clay who, despite missing the thrashing against Indiana last week, is averaging 5.4 YPC and has 13 TDs. Clay is every bit as good as we expected him to be, and if he's healthy, Michigan is going to have a really hard time containing him. If he's not, he's backed up by freshman "speed back" James White and sophomore Montee Ball. Both of these two would end up as starters on most Big Ten squads, averaging 6.8 and 5.6 YPC respectively. Wisconsin's run game is no joke.

Chances are that Michigan's defense is going to struggle to contain them. Without a healthy Mike Martin, Michigan is beyond sunk. With Martin, Michigan will still probably have to go to a four-man defensive line to stop free-releasing offensive linemen rumbling downfield. If Michigan can find some way to slow the run game (they probably can't), they might, might, have a chance in this game.

On the other side of the ball, Michigan is coming off of two horrendous games, combining for 10 turnovers. Denard has returned to his status as Turnover Machine and will need to do a significantly better job holding onto the ball if Michigan is going to have any chance to win. Taylor Lewan was concussed against Purdue and likely won't see the field in this game, meaning Mark Huyge will have the task of blocking Wisconsin's all-world defensive end JJ Watt. Mike Shaw was also concussed and probably won't see any playing time, which means that Vincent Smith, Stephen Hopkins, and (probably) Fitzgerald Toussaint will have to carry the load at tailback.

Prediction: This one is bleak. Michigan's defense doesn't have the firepower to hang with Wisconsin's beefy offensive line and stable of backs. Neither do they have the secondary to compete with an experienced and talented quarterback. Chances are high that Wisconsin will start rolling and never slow down, much like they did last week against Indiana. They won't hang 83 on the Wolverines, but scoring on 75% of their drives wouldn't surprise anyone. Wisconsin will score early and probably get up 14-21 points in the first half and try to run out the clock.

Michigan will lose the turnover battle again. Denard throws at least one pick and Michigan probably fumbles the ball a few times. I doubt if they gain any turnovers as Wisconsin is pretty judicious with the ball. Michigan will lose the ball at least once on their side of the field and it'll probably be a bullet in Michigan's chances to win. If they can find a way to shift this tide, they'll stand a decent chance to win, but in all likelihood, Michigan gets crushed in the turnover department.

Offensively, Michigan will probably put some points on the board. They scored 24 on the road last year against Wisconsin and are a wildly improved outfit. Michigan will miss at least one field goal in this game and the cameras will pan to a clearly distraught Rich Rodriguez as everyone in the stadium wonders why he didn't just go for it on 4th and 12. Michigan will probably have to rely a lot on their passing game against an oversized Wisconsin defensive front. How well Denard handles his reads and throws will be crucial.

There are just too many factors working against Michigan in this one. Then again, the same could've been said about Michigan's improbable win against Wisconsin in 2008. But expecting, or even hoping for, a similar outcome is just denying the mass of evidence that says Michigan is going to get bulldozed in this game. Wisconsin scores on their first drive, forces a Michigan punt, and takes firm control of the game in the 1st quarter. Wisconsin 52-35

Thursday, November 18, 2010

Purdue loading the box

Yesterday, MGoBlog Picture Paged how Purdue was loading the box against Michigan and leaving man-to-man coverage with no safety help for much of the game. Despite this, Michigan just kept running into a wall of defenders to middling effect. Michigan did very little to counteract this until the final drive of the game in which they went exclusively to their heavy package (with both Martell Webb and Kevin Koger on the field in the H-back positions). In response to Michigan's added blockers, Purdue committed yet another defender to the box (nine in total) and left the cornerbacks on an island against Michigan's two wide receivers. After seeing how unflinchingly Purdue had committed to stopping the run, Rich Rodriguez finally called a play to burn Purdue on a third and six that helped ice the game.

Here's the aforementioned formation. Purdue has nine players in the box with cornerbacks covering Michigan's receivers man-to-man. Denard has Vincent Smith next to him in the backfield, with Webb and Koger lined up as H-backs.

After the snap, Michigan leaves the weakside defensive end (Ryan Kerrigan) unblocked. Koger and Webb are rolling out to the weakside like they would on a regular zone read play.

Wednesday, November 17, 2010

Reading the linebacker

One thing that teams have started to do against Michigan in an attempt to stop the zone read is run their defensive end directly upfield and blitz a linebacker in the resulting gaps. We first saw this last week against Illinois. MGoBlog Picture Paged two different plays in which Illinois showed the defensive front and how Michigan adjusted to it. The first instance ended in a rush for zero gain. The second time, Michigan completely disregarded Illinois' defensive end and sent H-back Martell Webb to block the blitzing linebacker. It looked like this:

That's Michigan's triple overtime touchdown. The Illinois player who looks wildly out of position is their defensive end that ran directly upfield. Webb is about to block Illinois' blitzing linebacker. Touchdown.

Purdue deployed a similar tactic, but rather than blitzing the zone vacated by the defensive end, they blitzed up the middle of the formation, opening up the 19-yard Vincent Smith rush. The really interesting thing here is the read that Denard makes.

Michigan is in a three-wide set with Vincent Smith and Michael Shaw in the backfield. Purdue is in their base 4-3 with a linebacker over Roy Roundtree in the slot. Ryan Kerrigan is in a 9-tech defensive end position (outside the tight end, which in Michigan's formation isn't actually on the field).

Tuesday, November 16, 2010

Late reads = sadface

In the last few games, Denard has had accuracy problems in the passing game, resulting in multiple interception efforts against Michigan State, Illinois, and Purdue. After rewatching each game, it seems as if his mistakes are a combination of both bad reads and poor technique. Never was this more apparent than his first interception against Purdue on Saturday.

Michigan is in a 4-wide set with Vincent Smith in the backfield next to Denard. Purdue is in a 4-3 front. Michigan was just called for a holding penalty and is put in a 2nd and goal from the 18 yard line.

On the snap, Purdue rushes their four down linemen. Smith stays in the backfield to help in pass protection, which Michigan had struggled with all day. Purdue is playing zone coverage with the three linebackers playing underneath zones and the corners and safeties playing quarters coverage over the top.

The two receivers to Denard's left are both running slants. Denard's first read is Roy Roundtree in the slot who is slanting just beyond the underneath defenders. Unfortunately, Denard seemed to lock onto the outside receiver and is staring outside rather than making the proper read to the inside. The area in red is soft spot in the zone that Denard should throw into.

Right now, Denard needs to throw the ball, just before Roundtree crosses into the soft spot in the zone. Unfortunately, he just got his head turned around to his first read and recognizes where he should throw the ball. As you can see, Denard is holding the ball below chest level and won't be able to release the ball quickly enough. I think this is a fundamental problem in Denard's throwing motion: he holds the ball too low and has a wind-up that takes a bit too long. Now, so did Tim Tebow--and we all saw how that turned out--but a higher hold might help get the ball out a little quicker.

Denard has now released the ball, which is behind the line of scrimmage. Roundtree is running through the soft spot in the zone. and has basically no chance to catch the ball. Then...


Denard's issues on many of these throws have been poor reads, not inaccurate passes. This is another instance of what Rodriguez talks about when he says its important that the quarterbacks have their eyes in the right place. On the positive side, this can be fixed. With more time in the system and more time playing in general, these plays will become instinctive. Unfortunately, this play can't be attributed to the awful weather on Saturday. Making the proper read will be crucial if Michigan wants to have any chance against Wisconsin and Ohio State.

Monday, November 15, 2010

Purdue: Where Michigan looked awful

Let's be clear about one thing: Purdue is a terrible football team. Despite my assertion that their hodgepodge quaterbacking troupe would outperform their season averages against one of the worst defenses in college football, they looked lost, hopeless, and overmatched in almost every regard. Were it not for a 39-yard screen pass that resulted from a missed Michigan tackle, Purdue's quarterbacks were 16/32 for 93 yards (2.9 YPA!!!). Their running game was just was bad averaging 3.4 YPC. Purdue is bad; possibly the worst team in the Big Ten.

With that said, Michigan's 27-16 victory was... uninspiring. Purdue's defensive line thoroughly destroyed what has been one of the best offensive lines in the country this year. Part of that was due to Manimal Ryan Kerrigan, who made any Michigan lineman who tried to guard him look like a high schooler. His stat line was ridiculous: 4 sacks, 2 forced fumbles, and 10 tackles. Kerrigan is a legit top-10 NFL pick this year, but his end-to-end destruction of the offensive line is disturbing, both on a coaching standpoint (no H-back to chip block him?) and performance outlook. On Michigan's last drive, they went exclusively to their heavy set with both tight ends in the game and only two receivers, which had mixed, but generally positive results.

It stands to reason that both Michigan and Illinois were worn out from last week's marathon. Both teams faced wildly overmatched opponents (Purdue and Minnesota), and both struggled significantly. At least Michigan came out on top. The other mitigating factor was Michigan's inability to get anything going during the game due to turnovers, some of which could be attributed to the foul weather. Denard was constantly sailing deep passes--worse than he has all year, likely because of the rain--and holding onto the ball was a problem for both teams. It was an ugly win, but it was still a win. Michigan should regroup this week and get ready for the two-week gauntlet against Wisconsin and Ohio State.

  • Though Denard looked bad all game, he didn't get much help from the offensive line that was getting constantly beaten at the line of scrimmage. Vincent Smith had a good day because Purdue was over compensating for Denard, and when Denard did pull the ball, the offensive line looked sluggish. His passing wasn't great either. Anything that he threw downfield sailed well over receivers' heads and his short passes were often inaccurate. This was probably Denard's worst game of the year, at least some of which would be attributed to the rain.
  • Forcier looked just as bad as Denard did when he came into the game. Again, I think this falls on the weather and offensive line.
  • Michigan's defense looked functional. They played in almost exclusively three-man fronts, and showed a lot of 3-3-5 with the occasional 3-4 look. It was encouraging that the defense performed the way it did without Mike Martin and Jonas Mouton, caveats about level of competition noted.
  • I'll have to watch more closely on the tape, but I think Courtney Avery had a really good game aside from the 39-yard screen pass on which he was the primary culprit. Ray Vinopal also seemed to have another good day. And Mark Moundros appeared to play really well at both outside and inside lineback positions.
  • I don't really have much else to say other than let's be thankful Michigan came out on top. I'm sure there will be plenty of stuff to break down this week to try and make sense of Michigan's awful day.

Friday, November 12, 2010

Purdue preview

Michigan is now 1-1 against inexperienced starters, with a shredding at the hands of may-not-be-terrible Matt McGloin and an up and down day against Nathan Scheelhaase. Next up, Michigan gets to take on what's left of the Purdue quarterbacking depth chart in Sean Robinson. But before we get to that, the numbers:
Passing115th nationally
285.56 YPG
28th nationally
262.44 YPG
Rushing84th nationally
168.11 YPG
9th nationally
273.44 YPG
Scoring104th nationally
33.89 PPG
12th nationally
38.89 PPG
Passing64th nationally
218.22 YPG
114th nationally
137.33 YPG
Rushing49th nationally
142.78 YPG
39th nationally
173.22 YPG
Scoring74th nationally
27.67 PPG
111th nationally
17.56 PPG
So... Michigan needs to win this game; Purdue is thoroughly awful. Their defense is in the middle of the pack nationally, and given what Michigan did to Illinois' top-20 defense last week, that shouldn't be a concern at all. On the other side of the ball, Purdue has little or no offense, ranking in the bottom 10 of passing and scoring offense.

Freshman Sean Robinson didn't even see playing time until four weeks ago when he came in against Minnesota. This is because he's a freshman and if Purdue had any luck at all, they wouldn't have to be playing him this year. But their quarterback position has been thoroughly destroyed by injuries and now this is where they find themselves. On the year, Robinson is 32/68 for 233 yards (3.3 YPA (!)) and a 2/4 TD/INT ratio. On the ground, he's no more impressive, boasting 29 carries for 58 yards and no rushing touchdowns. But those numbers need to be tempered a bit as his only three career starts are against Illinois, Ohio State, and Wisconsin. We can expect Robinson to be better than those numbers (especially against Michigan's horrid defense) but he's the least likely of anyone to truly pick apart Michigan's defense.

Speaking of which, despite a fair amount of statistical analysis that says Michigan's defense was terrible last week, I was pretty encouraged by the improvements of the D against Illinois. The move back to a base 3-4 look was largely positive and if Michigan sticks with it, I think the defense could be whipped into functioning, non-crater shape by the year's end. With Craig Roh's move to the defensive line hopefully permanent, Michigan could actually turn in a decent performance, so long as the secondary avoids injuries and doesn't totally implode.

Michigan's offense is going to move the ball. Easily. This shouldn't even be a contest. Although Purdue may have the best defensive lineman in the Big Ten in defensive end Ryan Kerrigan. If Taylor Lewan can effectively shut down Kerrigan, Michigan should have no problem in this game. I also expect Michigan to largely run away from Kerrigan's side of the field, and we'll probably see the return of Denard optioning off of the playside defensive end. Otherwise, the Purdue defense shouldn't provide any resistance to this offense.

Prediction: Michigan wins. This should be the easiest game Michigan has all year in the Big Ten and should look more like UConn than Indiana or Illinois. Purdue's offense is anemic behind what is essentially Michigan's secondary: a young prospect that's being forced into the lineup way before he should. Robinson showed good things in the first half against Wisconsin before properly imploding. So while I expect Robinson to have a far better day than he's shown all year, his inexperience and youth will cost the Purdue offense a few too many times.

I expect to see Michigan come out in primarily a 3-4 set again, and I expect it to perform pretty well. I was encouraged by the defensive performance against Illinois and think that with a few more games in the 3-4 system, we start to see some of the wrinkles ironed out. Michigan will hold a decent running attack mostly in check and Purdue's passing game will be thwarted on bad passes or negative plays. Though I do expect to see Michigan victimized on more quarterback roll-out passes than we've seen the last few weeks.

Offensively, this should be a game of little resistance. Michigan's offense will only stall when they do it to themselves, and we should see a similar onslaught like we did against Illinois. Denard will have a huge game 200/200 and should account for three or four touchdowns. We will even see him break a long run again.

Finally, Michigan wins the turnover battle though not by much. The defense forces maybe one turnover but Purdue's defense doesn't provide much fight as Michigan's offense gets it clicking early. Greg Robinson will show another game of defensive improvement as Rich Rod and company allow him to continue running his schemes, but there are still a few infuriating breakdowns. Ultimately, though, this should be an easy one. Michigan 42-24

Thursday, November 11, 2010

"Just put points on the board"

It's nearly blasphemy to criticize the Michigan offense this year. By all accounts, it's one of the most prolific in the country. But there's one area in which it fails miserably: getting up by more than one score. I don't know if it's a play calling/coaching issue, an opponent's will to live, or just bad timing, but repeatedly this year, I've found myself saying, "Now just put points on the board" only to see a punt or turnover.

I went back through Michigan's games this year and, excluding the two blowouts (UConn and Bowling Green) tabulated how many attempts Michigan had to take a two-score lead, how many times they succeeded, and the game time that Michigan held a lead of 8+ points. The results are terrifying:

Opponent Attempts Success Game time with 2-score lead
Notre Dame 10 1 4:09
Umass 1 1 25:24
Indiana 5 0 0
MSU 1 0 0
Iowa 1 0 0
Penn State 0 0 0
Illinois 4 0 0
Totals 22 2 29:33

The most egregious example in the chart is the Notre Dame game, where Michigan had countless chances to put some distance between them and the Irish but failed to do so time and again. Some of these instances were the function of turnovers: against Indiana, Denard fumbled at the one yard line, and the Illinois game saw Denard throw an INT on the first play of one such drive, as well as a fumbled kickoff. Just as bothersome, however, is that in every game aside from Penn State--even Michigan's other two losses--Michigan had a lead, forced a defensive stop, and then had the ball with an opportunity to put more points on the board and almost always failed to do so.

There's something a little more revealing though: You can probably blame the three losses pretty heavily on the defense. Through two games, Michigan was only afforded the chance to go up by 8+ points twice, which means the defense was completely unable to stop the opposing offense. But the 0/4 and 0/5 efforts against Illinois and Indiana, respectively, paint the picture of a team forcing defensive stops but an offense that seems to slow down when they have the chance to put their foot on another team's throat. So the games become either a direct back and forth or a game of catch-up for the offense.

I think part of the blame here has to fall on Rodriguez. When Michigan has a chance to go up by two scores, playcalling immediately becomes QB sneak, QB sneak, pass. Rodriguez looks like he's trying to play ball control and run out the clock without actually having the comfort zone to do so (not unlike 3rd and short situations where Rodriguez goes into the I-formation). Other times, like against Illinois, it's just bad timing (Denard's second INT and the kickoff return fumble accounted for two of the four missed opportunities). Regardless, Michigan's offense looks to stall when it should really kick into gear.

For what it's worth, against UConn and Bowling Green, the offense was given one chance to go up by two scores, did so, and never looked back. But against everyone else (seven games), Michigan has only been up by 8+ points for 29:33. Part of that is Michigan's defensive woes, but even that doesn't explain the woeful 2/22 success rate. For a team that scores as often and as much as Michigan does, this stat is simply mind boggling. Not only will getting up 8+ points give a little breathing room for the offense, it'll help the defense by forcing opponents to get out of their gameplan. Whether or not that will translate to more wins is yet to be seen, but if Michigan is going to have any chance going forward, they're going to have to capitalize on opportunities like these.

Late post today

I spent 1.5 hours at the DMV yesterday and then 2 hours in traffic trying to get home. I was in no shape to put anything together for today. But I'll probably get something up around 4 pm EST. In the meantime, enjoy this billboard of Jack Johnson, part of an advertising campaign for the LA Kings. Apologies about the image quality: this was taken on the fly during my commute home sometime last week.

Wednesday, November 10, 2010

Mixing up zone coverage

Coming into the season, we heard a lot of talk about how the defense was going to show multiple fronts capable of confusing quarterbacks. The results have been... uninspiring. But against Illinois, with Greg Robinson's back against a wall, he finally started to show some of the things that the defense promised before the season began. Part of this is because Michigan was daring Nathan Scheelhaase to throw the ball, about more which later, but Michigan's defense looked like a completely different unit against Illinois than they have all year. On one third-and-short early in the game, Michigan disguised their defensive coverage in an attempt to bait Scheelhaase into a bad throw while staying fundamentally sound and forcing an Illinois punt.

Michigan is in the wildly effective 3-4 over that they had shown for most of the game. Cam Gordon is rolled up to the line of scrimmage at the top of the screen. Michigan has its corners and safeties all off the line of scrimmage. I believe Michigan is showing a cover-2 look pre-snap. The person to watch is Jordan Kovacs (in red). Illinois has a bunch receiver set with a split receiver to the other side of the field and Jason Ford in the backfield next to Scheelhaase.

After the snap Michigan rushes their three down linemen as well as Gordon who had walked up to the line of scrimmage. Kovacs has continued to come down the field from the safety position he showed before the snap. Michigan's corners and Ray Vinopal will drop into cover-3. Illinois is running the snag route with the x-receiver (top of the screen) running a hitch and Ford running a flare out of the backfield.

Kovacs' zone is in red. As the x-receiver starts to turn his shoulders on the hitch route, Kovacs is undercutting him. If Scheelhaase throws here, which he might--this is his first read--Kovacs has a sure interception. Gordon is coming unblocked off the edge forcing a quick throw from Scheelhaase. Meanwhile, Michigan's linebackers are playing underneath zones and sorting out the bunch formation Illinois is running.

Scheelhaase either has to throw or take a sack from Gordon who is still untouched. Kovacs is in perfect position and has come directly under the x-receiver. He also has his eyes in the backfield and can see Ford running the flare route into the flat. He'll get a good break on the play and have a chance to make a tackle for loss.

Ford catches the ball behind the line of scrimmage and turns upfield. Kovacs is heading directly downfield at him with a chance to make the tackle. He'll eventually whiff, but Ford will be wrapped up by rallying Michigan defenders short of the first down marker and because Kovacs was able to slow him down.

One of the reasons Michigan succeeded here was that they brought pressure off the edge with Gordon. Michigan showed a 3-4 front for a lot of this game and it was really effective. They frequently brought blitzes off the edge and were able to force quick throws or scrambles from Scheelhaase. Part of this is because Michigan was confident that Scheelhaase couldn't beat them through the air consistently. Though I think this is probably Michigan's most effective defensive approach (bring a lot of pressure, mix up looks, and try to get to quarterbacks), against better quarterbacks, Michigan probably won't be so cavalier.

In the end, this play wasn't really any different than the same zone coverage they've been showing all year: a four-man front, four underneath defenders, and cover-3. What changed was the pre-snap alignment of Kovacs and Michigan's ability to disguise the coverage. By bringing Kovacs down from a pre-snap safety position, the x-receiver was unable to read the zone and settle into a soft spot. The Gordon blitz confused the Illinois offensive line, and Gordon was able to get into the backfield untouched, forcing a quick throw. Scheelhaase's first read is the x-receiver, but with Kovacs running directly into the passing lane, he's forced to check it down to the flat where Kovacs' momentum will naturally carry him.

This just looks like a more competent defense than we've seen all year, and much of it is because of a change in base personnel/formation. The 3-4 front that Michigan showed was really effective in this game and should be Michigan's base defense going forward. It's plays like this and the general success Michigan showed against non-option runs that makes me think this defense isn't completely hopeless and Greg Robinson hasn't already been handed his pink slip.

Tuesday, November 9, 2010

Credit where it's due: Demens zone coverage

After Kenny Demens' first start against Iowa this year, I was skeptical of the amount of praise being heaped upon him. What I saw was a raw young talent that showed some problems in pass coverage. One play in question was an Iowa touchdown that came because Demens failed to properly cover his zone, allowing a dragging receiver to run unabated into the endzone. The incriminating moment:

The receiver skating past Demens was the slot receiver from his left. Demens failed to diagnose the play in time to cover him, and the receiver ran to the corner of the endzone with the help of Courtney Avery who had turned his head away from the play while re-routing a vertical receiver.

But against Illinois on Saturday, Demens showed that he's learned from his mistake and properly diagnosed a dragging receiver in what is a shockingly similar play--very likely designed by Illinois to exploit the same pass coverage problems Michigan's linebackers showed against Iowa.

There are a few subtle differences between the Illinois (bottom) and Iowa (top) plays in pre-snap alignment. Against Iowa, Michigan showed a four-man front, with four underneath defenders and cover-3. Iowa had three receivers, a tight end to the left of Stanzi, and a running back in the backfield, while taking the snap under center. Illinois had a similar three-receiver setup, but took the snap in the shotgun with a running back next to Nathan Scheelhaase. Iowa dragged its tight end across the formation, running a mesh route with the weakside receiver. Illinois runs a crossing pattern between the weakside, dragging receiver and the tight end.

On the snap, Michigan rushes the three down linemen and drops eight players into coverage.

A moment later, there are two things to watch. First in the crossing pattern by the tight end and receiver on the bottom of the screen. The receiver is crossing under the tight end and beginning to drag across the formation. The other thing to watch is Cam Goron on top of the screen, who is charged with re-routing the slot receiver, much like Avery was supposed to do in the same play against Iowa.

Monday, November 8, 2010

Illinois: Where Michigan's bowling

"It's been a good week. Mr. Brandon and I both talked about that on the way over here. It's been a good week." - Rich Rodriguez

It's hard to sum up Michigan's win against Illinois any more succinctly. Maybe if I could find a picture of the smile Rodriguez gave when he said "It's been a good week."

This game played out like basically all of Michigan's this year: crazy offense, lots of defensive lapses, Denard injury, down-to-the-wire shootout, turnovers, etc. Except Michigan actually won this game. Illinois' offense was just a little too sputtering to overcome Michigan's, which was able to dispense Illinois' defense--that had won them five games already this year--with ease. Now, Rich Rodriguez can look ahead to Purdue and, with any luck, 2011, a year that should see Michigan contend for a Big Ten title.

It's worth mentioning that there's a non-zero chance that Greg Robinson is retained at the end of the year. In my Where Are They Now posts during the bye week, I said this,
However, there's 0% chance [Greg Robinson]'s fired at the end of the year granted that Rodriguez is kept on board. Rich Rod realizes the talent issues Robinson is facing and will afford him at least one more year.
So maybe 0% was probably a little overzealous. But given what appeared to be improvements from the defense in this game (about more which in a minute) and the fact that Robinson has gained Rich Rod's trust the past two years--running the 3-3-5 at his behest, et al--there's a distinct possibility that he sticks around for another year. I wouldn't be all that against it, given Michigan continues to improve like they did on Saturday.

  • 65 points be damned; Michigan's defense looked better in this Illinois game. And all it took was doing away with the 3-3-5. Of Illinois' 16 possessions, six of them ended in punts, one in a turnover, one in a missed field goal, and the rest in touchdowns or field goals (forcing field goals, go figure). I'll take a 50% success rate with this Michigan defense. "Slightly below average" is way better than "worst in the country" if you ask me, especially when Mike Martin isn't playing for much of the game. Now if we could just do something about those...
  • Turnovers. Denard's first interception should be credited to Roy Roundtree who, on the previous play, dropped a first down pass that hit him square in the gut with no defenders around him. Denard's second was both his and Martell Webb's fault. I think Webb was probably a hot read who didn't get his head turned around in time. Forcier's fumble was... bad. The others were not great.
  • Michigan's kick coverage was awful in this game. Kickoffs--and there were a lot of them--were routinely returned to the 40-yard line or beyond. I don't know if this was something that Illinois was doing or something that Michigan was doing, but it was ugly.
  • Nathan Scheelhaase is definitively the worst starting quarterback in the Big Ten. Dude simply can't throw. His stats probably speak for themselves, but to reiterate what I said in my preview: daring him to throw downfield, even with Michigan's defense, is probably a good idea.
  • Argh halfback wheel routes. Halfback wheel routes are Michigan's new MOEAKI!!!
  • Courtney Avery was functional-to-good. Maybe it was because of Michigan's overall scheme change, but dude looked like an upgrade from Floyd. He made a huge first-down-saving tackle in the open field and generally played well. Ray Vinopal also had a few big plays in the game and looks like he might be a functional answer to Michigan's free safety problem.
  • Michigan's successful two-point conversion was the same play that Forcier threw to Matthews to beat Notre Dame last year. Michigan gets man coverage on the outside and runs a slant-out play that gets the receiver open in the corner. Forcier made a better pass to Hemingway than he did to Matthews, but both ended the same way.
  • Roy Roundtree getting pulled down from behind was... ominous.
If Denard has a concussion (and he likely does), Michigan should be able to start Forcier against Purdue and still come out with a victory. I'd rather see Denard get healthy for the Wisconsin/Ohio State gauntlet than risk more serious injury against a team that Michigan should be able to dispense of. But for now, I'll just enjoy this victory.

Friday, November 5, 2010

Illinois preview

I mean, whatever...
Passing19th nationally
183.88 YPG
111th nationally
143.13 YPG
Rushing26th nationally
117.50 YPG
24th nationally
194.38 YPG
Scoring12th nationally
16.75 PPG
63rd nationally
26.88 PPG
Passing117th nationally
290.50 YPG
43rd nationally
242.88 YPG
Rushing57th nationally
149.75 YPG
8th nationally
275.50 YPG
Scoring89th nationally
30.00 PPG
35.38 PPG
I honestly don't know what to say anymore. Michigan's defense is the worst in the country and just lost its starting cornerback, which may or may not make any difference. The Wolverines are hosting the upstart Illini led by redshirt, arm punting freshman Nathan Scheelhaase. If Michigan's defense was even remotely competent, they would stomp an Illinois team that has trouble moving the ball. But with  Michigan's ability to turn any walk-on who thinks he's Brett Favre into Brett Favre, this is another game where the opposition will greatly exceed their season averages and we'll all stand, holding our breath, hoping that Denard doesn't intercept the team to defeat. I'll try and be more constructive, though:

Scheelhaase has actually gotten a bad rep and has been a fairly effective quarterback aside from two games (MSU and Missouri) against defenses that Michigan cannot replicate. Those two games--one of which was his first as a starter--accounted for six of his seven interceptions on the year. Otherwise, Scheelhaase has thrown nine touchdowns to just one interception. But part of that is because he's averaging a gimpy 6.62 YPA. And his reputation as a dual threat quaterback is a little overblown: he's averaging 4.8 YPC on 93 carries, averaging just 56 yards per game. More threatening than straight runs are the rare plays in which Michigan covers for long enough that Scheelhaase has time to scramble out of the pocket. If Matt McGloin can get 11 yards on QB scrambles, Scheelhaase certainly can.

Though a little counter intuitive, the way to beat Illinois is to dare Scheelhaase to throw down the field. When he has to make a long throw, he often throws ducks that flutter into the secondary. He does the same when pressured in the pocket. If Michigan feels good enough with their corners, blitzing more frequently and expecting Scheelhaase to throw the ball beyond the sticks may be a viable way to beat them. Though Michigan can't forget about running back Mikel Leshoure who is averaging 4.9 YPC.

On the other side of the ball, you know what to expect. Illinois has a good defense, but so did Iowa and Michigan State, and Michigan only slowed down when they beat themselves. Expect Michigan to run basically everything we've seen them run with maybe one or two new formations/plays. Mike Shaw didn't show up on Michigan's injury report, so expect to see a lot of him in the backfield with Denard.

Prediction: I really don't know. After the last three weeks, it's difficult to imagine Michigan winning against a pretty good Illinois team. But then again, it's still hard to pick against them because, well, Illinois isn't that good. Hasn't this defense hit rock bottom? Is this the week that Greg Robinson simplifies schemes and just runs a 3-4 or a 4-3? Is this the week that Rich Rodriguez takes over the defense and demands better production? If Michigan's defense is really as porous as it seems, wouldn't we rather see Leshoure than Evan Royster?

I expect Michigan to be able to move the ball just as well as they have all year. Denard has another 150/150 day and may even push 200/200. Mike Shaw also has a decent performance but nothing to write home about as Michigan will probably end up passing a lot in the second half to try and catch up. Denard will throw at least one pick in this game--probably early on--but he'll account for at least three touchdowns. Michigan will try kicking at least one field goal in this game.

Defense is a wild card. With Greg Robinson's obvious pink slip on the table, Michigan's defense will be changing things up to try and make something work. And with the reports that Craig Roh has complained about his move to linebacker and has been playing as a defensive end, this may be the first time that we see the 3-3-5 mostly scrapped in favor of more traditional 3-4 and 4-3 sets. JT Floyd is injured, so expect a lot of true freshman-ness and the mistakes that come with it. If Michigan moves away from it's hackneyed 3-3-5, this might be the corner turn we've waited for. If we see a lot of stacked sets, expect a similar performance to the Penn State game.

So this is another game that I'd like to skip a prediction on because I can see this game playing out a lot of different ways. One thing we know for sure is that Illinois is going to move the ball. Whether or not Michigan can find a way to force field goals, punts, or the occasional turnover is yet to be seen. Although banking on any of those happening is fools gold. This has the makings to be a Notre Dame-like shootout (quarterback incompetence interspersed with massive defensive breakdowns, versus, ya know, Denard). I'm still not sold on Scheelhaase's ability to throw the ball downfield, and if Illinois sticks with underneath routes, Michigan might be able to force a few stops. And despite his status as a scholarship player, I think Scheelhaase is a worse passer than McGloin, so third-and-long situations might (might) work out OK.

I have literally no confidence in this prediction. Michigan 28-27

Thursday, November 4, 2010

Floyd and the NCAA (unrelated)

Watching the Penn State game any more than I already have this week would end in unceremonious sepuku, which would usually be fine, but I'm more or less house sitting right now and I'd hate to ruin someone else's carpet. Instead, I'll touch on some of the odds and ends floating around this week.

Like everyone with a clue had expected, the NCAA accepted Michigan's self-imposed penalties for the "major" violations. I don't have much to say about this besides, "well, duh" and other things that have been more eloquently expressed elsewhere. Michigan's crimes were minor in nature but major in name, and only surfaced because of an obvious agenda by the Free Press. Anyone trying to deny that anymore is of the same mind and probably still complains that Rich Rodriguez has a West Virginia accent.

Michigan had a press conference just now and David Brandon predictably dropped bombs on the Free Press,
We believe that a thorough investigation would show the allegations to be false and misleading...

There was nothing found that even remotely suggested that our players welfare, safety, or wellbeing was at risk.
which was basically Brandon giving the middle finger to the Free Press. The only bits of relevant information otherwise were that Michigan was given a third year of probation, and that 32 of the 130 practice hours Michigan will sacrifice have already been completed. Michigan plans on getting through these hours in 18 months rather than the 24 months they were allotted.

Now we can go back to being casually uncompetitive at football in peace.

Floyd and the depth chart
With the loss of JT Floyd for the season, Michigan now turns to three true freshmen to try and carry the secondary load. This may not be as harrowing as it sounds because Floyd hasn't been particularly effective and, well,

What exactly are we so afraid that the freshmen will do, fail to cover people and miss tackles? That's what we were getting with Floyd. And now we'll gain at least 30 yards a game back in facemask penalties. At least now we'll get to find out what we've got in this freshman group. It's sink or swim time.

My guess is that Courtney Avery is the first off the bench to replace Floyd. Aside from my general excitement about Avery in the face of his obvious coverage miscues, I don't know if anyone else has been taught the position. Michigan's secondary moves around a lot during games: corners drop to safety depth, they move into linebacker-type positions, and they play corner. But the concern is that left corner and right corner are not mirrors of one another. It has seemed like Avery has come in to replace Floyd and the platoon of Cullen Christian and Terrence Talbott have relieved James Rogers when he comes off the field. So starters of Avery and Rogers is what I'm anticipating with Rogers being replaced sometime in the game with one of the other freshmen. It's going to be terrifying.

The talent drop off from Floyd to the freshman should be basically zero, and this will give the young players a chance to really get their feet wet. The other bonus is that with the younger squad, Michigan will probably have to simplify its schemes in order to protect them, meaning the defense may very well find a formation and stick with it for most of the game. This would be a big step up from the mish-mash of formations that they brought out against Penn State.

Then again, Michigan's two corners and free safety are true freshman (or James Rogers. Their other safety is a sophomore walk-on. Their linebacker/safety hybrids are freshmen. So... Nathan Scheelhaase is going to pass the ball a lot.

Wednesday, November 3, 2010

"This is not a stack" addition

MGoBlog had a Picture Pages post today addressing the 3-3-5ness or lack thereof of Michigan's defense this year. As I rewatched the game, I saw similar problems but didn't quite know how to explain it. What I saw was Penn State frequently getting a blocking advantage with motion and designed cutbacks. That's beside the point, however, and I think Brian explains Michigan's schematic problems really well there. I did see one thing in his comparison to a similar WVU defense that he failed to mention--and it's something I actually loathe to bring up because this horse is dead and I'm tired of kicking it.

One of the reasons Michigan can't show this front is because they can't/won't play man defense. WVU is showing a six-man rush with man coverage on the receivers, the middle linebacker on the running back, and presumably, the safeties covering the tight ends if they release upfield (otherwise, they're just giving cover-2 help over the top).

One of the keys to this play, as Brian says,
At the snap six players attack the line, giving all but one WVU DL a one-on-one matchup
With Michigan's insistence on playing zone coverage, the linebackers/spurs/bandits are unable to attack like this because of play action. Since each player has a responsibility in the passing game, attacking the line of scrimmage aggressively on a hand off will open up gaping holes in play action. Michigan's linebackers have to wait until they're certain that the running back has the ball before they can attack the line of scrimmage, at which point, the offensive linemen have been given a free release to the second level. Not to mention that when the linebackers start flowing down the line, it creates gaps in zone coverage on the backside of the play, something that Penn State saw a lot of success on by sneaking tight ends and full backs into the flat.

So when Brian says things like,
Michigan is a passive three man line with guys you can easily single block (but get to double if you want) and linebackers who are living a nightmare.
it's largely because Michigan physically can't attack the line of scrimmage more aggressively while also requiring second-level defenders to be responsible for the passing game.

This, of course, comes back to the Read and React principle that you expect your players to be able to perform, because not all zone plays get gashed by halfback runs. However, with only a three-man defensive line, Michigan is giving up far too many free releases to the second level where, when players have had a chance to react to the play, they'are being blocked by free releasing linemen.

I brought up a similar concern after the Indiana game, when I started my man vs. zone tirade:
Rogers, Floyd, Avery, and Talbott can all cover for at least limited amounts of time (Avery might be the best cover corner on the team that's currently healthy, about which more this week). Never did they get blown off the line and completely lose coverage [Ed. Whoops, wrong here re: Floyd]. As such, Michigan can blitz more often. I think part of the reason Michigan hasn't blitzed as much as we'd all like to see is because, as porous as the zones are with eight defenders, they'd be even more open with five or six.
I was arguing for, essentially, the same as above: with the ability to cover receivers one-on-one, a defense has more opportunities and flexibility in the blitzing game. Michigan can't bring more than three or four blitzers--Michigan's "blitzes" are equivalent to a four-man defensive line--because they are terrified of being beaten in the passing game. As such, Michigan has settled for these three-man rushes in the hopes that the zones behind them will hold up, which they haven't.

This defense is broken and more bothersome, it appears that Robinson is fine with the status quo. This is the first week where people are actively coming after him rather than Rodriguez, who has taken the brunt of the media's force this year. If Robinson doesn't change anything this week against Illinois, Michigan will lose in the same fashion that they've been beaten the previous three weeks.

Pulling the ball on the zone read

It's no secret that teams have been coached to force Denard to hand the ball off on the zone read. But despite Michigan's numbers advantage after the handoff, the running backs have only been gaining minimal yardage and occasionally tackled by the defensive end that was keeping contain on Denard. Part of the reason that the backs were having so much trouble is that just before Denard hands the ball off, the defensive ends are already starting to attack the running backs. Against Penn State, Denard broke out some freaky Juice Williams play fakes on the zone read, where he held the ball long enough to catch the defensive ends cheating like this. Denard's ability to execute this frequently will be key to keeping defenses honest and opening up running lanes for the halfbacks. One such instance was on a 2nd and 13 in the third quarter following a clipping penalty that set Michigan back.

There's nothing special about this alignment. Michigan is in an four-wide ace set with Vincent Smith in the backfield. Penn State showed this defensive front a lot during this game, and I believe that it's a 4-2-5, but without more closely analyzing their defensive personnel, it may just be a variation of their base 4-3. But for the sake of argument, we'll call this a 4-2-5.

On the snap, Michigan leaves the playside defensive end (red) unblocked. Steve Schilling is pulling across the formation to the playside of the play and will be responsible for the playside linebacker. Denard is reading the defensive end.

Denard is still reading the defensive end who is showing contain, signaling that Denard should hand the ball off. Schilling has pulled across the formation and is getting to the second level to block the playside linebacker.

This is the moment where Denard really sells the fake. Smith is now beyond Denard but they are still engaged; Denard can still keep or hand the ball off. The defensive end begins to move with Smith because he thinks that Denard is handing the ball off, but he's held it long enough to see which way the defensive end is going to commit.

Now the defensive end has completely committed to the Smith run, as has one of the Penn State defensive tackles who Patrick Omameh has completely whiffed on. Despite the missed block, Denard is still able to skirt ahead through the gap because the defensive tackle was sold on the play fake. Penn State's playside linebacker and nickelback are also crashing on the Smith run, opening up the field for Denard.

The defensive end (red) who had been keeping contain on Denard is now three yards away from him laterally with no shot of catching him. Denard was able to make sure he committed to Smith before pulling the ball. The defensive tackle that Omameh missed is also diving backward to try and tackle Denard after he bit on the play fake. With good blocks down the field, Denard is able to pick up the first down and more.

I think a lot of people have been especially negative on the running backs this year when they've been pretty functional, at least on par with the production of the receivers. But despite teams nominally forcing Denard to hand the ball off on the zone read, they've still been cheating and finding a way to get to the running backs by reading Denard's play fake. Michigan has seen opponents try a couple of different things like this to contain the hand off. Notre Dame blitzed their linebackers to plug the holes at the offensive line in order to give the weakside defensive end enough time to come down the line and make the tackle on the halfback. But Michigan was able to counter this by using the play-action seam route that has become a staple of the offense. In this case, and presumably going forward, Denard's play fakes are going to take longer and force the defensive ends to truly make a commitment to either the running back or quarterback, opening up lanes for both players.

Tuesday, November 2, 2010

2nd and 21? No problem

Going into this game, I thought that any time Michigan got Penn State in a long-yardage situation, it would almost certainly end up in a Michigan stop. PSU was starting a walk-on quarterback and Michigan's defense couldn't be that bad. Well, I was wrong. On 2nd and 21 in the first quarter, Michigan failed miserably in every single regard and gave up a wide open reception against two vertical receivers while playing cover-3. There's no way to watch this play and think of it as anything other than a complete and utter failing by the defensive coaching staff. There are so many fundamental mistakes in this play that it's difficult to image this team has ever practiced together. Masochism:

As I mentioned, it's 2nd and 21 after a Michigan sack. Michigan is showing a cover-3 look with a three-man rush and five underneath defenders. Penn State comes out with three receivers in an I-formation. They'll run play action, send their wide receivers down the field, and drag the slot receiver across the formation. James Rogers, Ray Vinopal, and JT Floyd will split the field into deep thirds. In what is very likely a passing down, Michigan has essentially eight men in the box.

Penn State runs play action. You can see Michigan's underneath defenders are all frozen in place, and all of which are between the hashes (!!). PSU's outside receivers are running straight downfield and the slot receiver is starting to come across the formation. While Michigan is susceptible to an easy 10-yard hitch route on this play, that's actually what they're hoping for: force PSU into a third and long and look for a stop then. This is where things get really ugly though.

On what is now obviously a passing down, Michigan's underneath defenders are still all in the box and none is deeper than four yards from the line of scrimmage. PSU is in a max protect set and has seven players blocking three Michigan defenders. To make matters worse, Kovacs and Gordon are both staring aimlessly into the flats where there's absolutely no PSU players. Michigan's linebackers are within five yards of one another, and all three are guarding the slot receiver dragging across the field.

At this point, Kovacs realizes how poorly he's executed this play and has turned his head away from the quarterback to try and catch up with the outside receiver. Michigan's linebackers are literally running into each other and covering no one (What's happening here?). Michigan's underneath defenders are all still essentially inside the hash marks and none of them are deeper than five yards beyond the line of scrimmage. McGloin has all day to throw because, well, seven blockers against three rushers won't ever get to the quarterback.

The ball is released (in red) and three of the five underneath defenders have now gotten good depth--the problem is they should have gotten this deep shortly after the ball was snapped. Instead, they're terribly out of position and no where near any of PSU's receivers. Then...

...there is a wide open PSU receiver just past the sticks with the closest Michigan defender 10 yards away. You might be asking yourself, "How did that guy get so open on a two-receiver route against cover-3?" Let's look:

Floyd is in pretty good position as the deep safety. Despite not having any underneath help because his underneath defenders are woefully out of position, he's keeping the play in front of him and in position to force the play toward his help defense (Ray Vinopal in the deep middle of the field). Then...

...the PSU receiver takes one stop toward the inside of the field and then breaks to the outside. Floyd immediately turns his back on the play to run with the receiver to the inside, despite having help defense toward the middle of the field. By the time Floyd gets his head turned around, the receiver has settled into a completely vacated zone 22 yards down the field and makes the first down catch.

Absolutely everyone involved in this play is to blame, but almost all of it has to fall on Greg Robinson. These players have absolutely no idea what they're doing here. The underneath defenders get no depth. Floyd doesn't know where his help defense is and once again shows how poor his technique is (like on those wide-open slants against Iowa) by turning away from the play. I mean, this is 2nd and 21 against a walk-on quarterback in the second drive of his first career start. How does this happen?

You can say that the players didn't execute here, and that you can't blame this on Robinson when we don't know exactly how they were coached, but the massive amount of basic error is unfathomable for a team in its 8th game of the season, regardless of whether or not these are freshmen or seniors. I honestly don't know how a defense can be this clueless. No one on the field appears to know where they're supposed to be or how they're supposed to defend this. When this play was taught, what exactly did Robinson see that made him think, "Yes, that will work"? This is not a complex play with pattern reading or man coverage or stunts. This is a three-man rush with five underneath defenders and three deep safeties. Penn State sends only three receivers. And this isn't even one of those broken plays where defenses are forced to cover for too long. PSU's receiver ran 22 yards, stopped, and caught the football.

There's no way to watch this and conclude anything other than massive failing on the behalf of the coaching staff and that Michigan has the worst defense in the country. 2009 looms.