needs moar usage
Last time on Picture Pages, Denard Robinson got Roy Roundtree killed against UConn by being too impatient to throw on a curl-flat combo. This time things will go a bit better.
The setup has Michigan in a four-wide formation with ND in a two-deep shell and a 3-4 defense—this is slightly unusual since ND spent most of the day in a 4-3:
Michigan starts the play with a zone stretch fake, pulling Schilling around to act as the lead blocker Shaw was on the previous play:
Roundtree's come in like he's going to block as Odoms heads upfield:
Roundtree then breaks outside as Walls rubs Odoms, pushing him out of bounds briefly. If this was man coverage Roundtree would be open, but if it's zone it'll be Odoms, or you can also take a look at Grady running well behind the linebackers, all of whom have sucked up to either the zone stretch fake or the threat of Robinson on the edge:
It was right about here, if not earlier, that Denard threw the ball against UConn:
But as you can see from the downfield perspective, that would have been a bad idea since the corner is disconnecting from Odoms and coming upfield. If he had thrown it above, the ball would be about halfway to Roundtree right here:
This probably would have led to another decleating hit. But Denard sees the play developing and waits. A split second later Odoms sits down on a fifteen-yard hitch. He's wide open:
ND's Harrison Smith doesn't know what to do with both Grady and Odoms open in front of him. Robinson zing:
Odoms picks up some YAC…
…and it's 21 yards.
- Most of the same stuff from the last post. Michigan will see a lot of zone. Most of the routes they run will be designed to beat it. Robinson is going to be expected to high-low cover two corners a ton, not least because a hard cover two corner gives the defense much better containment on the edge than a cover three where the corners bail out into deep zones.
- Holy pants wide open receivers again. The Roundtree and Odoms catches were more spectacular and the end result of this play had nothing to do with the linebackers, but Kelvin Grady is screamingly wide open on this play too because Manti Te'o is hurtling towards the line of scrimmage and Calabrese sucked in towards the zone stretch fake. Michigan should have run more play action, though I guess it's hard to criticize what the offense did when it wasn't getting flags in its face.
- Denard is learning stuff. Obviously. I haven't gotten through everything yet and do remember a period in the fourth quarter where he was looking pretty wobbly, but the coaches probably spent a bunch of time this week working on Denard's mistakes and getting his patience right for various plays. So far there haven't been any plays against Notre Dame where I thought "that throw is way too late/early" except one on which Robinson dodged a blitzer and had his timing disrupted.
- Odoms is just fine as an outside receiver. It would be nice if he was a towering colossus of speed but given Robinson's strengths it's better to have a reliable mountain goat and experienced route-runner who can sit down in the right spots and catch the balls zinged to him. It seems clear that going over the top is not one of Robinson's strengths, at least not right now.
This is from the UConn game and has been discussed previously, but here it is in glorious coughing-up-blood Picture-Page-O-Vision. It's pretty simple but I don't think I've spent much, if any time, on the site discussing making reads in the passing game.
It's the start of the third quarter and Michigan is facing second and eleven. UConn comes out in their two-deep look with corners playing off. Unless the Huskies are disguising a coverage this is likely to be two deep, and since opponents are almost forced to play zone against a spread attack featuring one Denard "Shoelace" Robinson, Esq., Michigan has a pretty good idea that UConn is either going to play a standard cover two defense or a cover four "quarters" look.
On the snap Michigan does a half-roll of the pocket, which gets Robinson closer to his intended targets, can delay linebackers uncertain whether it's a run or pass, and opens up lanes for Robinson if his receivers are covered:
At this point it's obviously a cover-two zone with the two deep safeties and the corner sitting about seven yards downfield looking in the backfield. Roundtree breaks well outside of the playside LB, who was held inside by the threat of a run. That guy's not going to prevent him from turning upfield if the ball is accurately thrown.
This is a curl-flat package where the inside receiver runs a very shallow out and the outside receiver heads about ten or fifteen yards downfield, then sits down in what should be the hole between the corner and the safety. The cover-two corner then has to pick whether to sink deep to take away the curl, opening up the flat, or come up on the flat, opening up the curl:
Robinson cocks to throw, but there's a problem:
He's throwing the ball too soon, before the corner has been forced to make a choice. Stonum's not even five yards downfield. The corner is is looking directly at what's going on and can jump up into the route…
…and it's never good when you're catching the ball with your back turned to a blur…
…so Roundtree is daed:
Video of what went down:
- Opponents are going to have to play a lot of zone against Michigan this year. Anyone intent on having base personnel on the field—which both Notre Dame and UConn did the entire game—will be putting linebackers in space against slot receivers if Michigan goes to man, and possibly opening up big plays when those guys read run incorrectly. Also, man coverage against four verticals means a lot of guys are running downfield with their back to Denard. This is not good for a defense.
- Most of Michigan's routes will be zone beaters, then. This may be the source of criticism about Rodriguez's fairly primitive passing packages, but if you've forced the defense into a limited subset of available coverages you can get away with this, as Michigan did all day against UConn and on the final drive against ND, when Michigan ran several variations on curl-flat to march down the field.
- Zone-beating routes endeavor to make one particular zone defender cover two guys. Here it's the outside guy on Stonum and Roundtree. In the snag package Michigan ran all day against UConn it's the playside linebacker and sometimes the playside corner.
- Most of Robinson's reads are simple "if this one guy does this throw it here, otherwise throw it there" things.This is the privilege afforded him by his running ability. Exotic coverages are difficult to get away with unless you're really good. I expect Ohio State to be able to confuse him. Maybe Iowa, Wisconsin, and Penn State will be able to do this as well, though PSU and Iowa are replacing lots of linebackers and are dedicated to base defense, too, so man coverage will be hard to get away with.
- Here Robinson lacks the patience to let the play develop. If he just waits a second or two it will be clear which option is open.
Later today: Robinson learns from his mistake to Notre Dame's detriment.
This is belated, but still relevant since I just saw Michigan run this play with Denard Robinson as the tailback in the Purdue game. This is a staple of the Illinois offense but it's something Michigan hasn't run yet in the Rodriguez era. I assume Michigan decided that the best way to practice it was to install it, and once it's installed you might as well run it.
It's Michigan's first drive of the day. The setup is a standard four-wide set on which Michigan is in a stretch setup with the quarterback behind the tailback:
On the snap it looks identical to the stretch in the backfield, with the running back coming across the quarterback's face to take a handoff, or not take a handoff. The line, however, is doing something completely different. They're blocking down:
On a stretch the line would be moving the same direction as the tailback and leaving the backside defensive end unblocked. On this play they block the opposite direction and leave the frontside DE unblocked. This is a veer.
Michigan's run a different sort of veer earlier that looks more like a traditional stretch with the tailback attacking upfield and the quarterback the player that needs to be contained. That touchdown against Purdue last year where Minor ran untouched into the endzone was a veer. On this play, the upfield threat is the QB and the RB needs to be contained.
On the exchange the Illinois line has slanted in anticipation of a stretch; they're reacting to the line. The backside DE is shuffling out and Koger is immediately releasing to the second level to pick up a block on the MLB:
Forcier keeps it. I think he keeps it incorrectly given the DE's reaction to the play:
Oops. If that DE had taken off for the tailback this is a good gain. Look at that crease up the middle. Since Forcier isn't Denard Robinson he probably gets tracked down by a safety—they're off the screen deep—but Illinois has gotten fooled by this play. Everyone except the DE, that is:
He cuts off that crease.
Forcier's a slippery bugger in space, though, and this DE is not nearly as agile as he is. As we've seen all year, dude can make you miss. He manages to get around the DE and to the outside. This delay has allowed opponents to converge, though:
Forcier gets down voluntarily:
It's four yards thanks to the mismatch between the DE and Forcier, but he had to make a guy miss to get it.
- Forcier's freshman status is much more pronounced on the zone read. Forcier's made a lot of poor decisions this year when it comes to handing the ball off or taking it. Most of the time his error is keeping the ball, but when Robinson came in to run the veer against Purdue he handed it off as the DE was running right out of the play and Robinson got nailed for a three-yard loss.
It's not just the passing game which should improve as Forcier gets more experience. Michigan's run game is being hampered by Forcier's youth as well. This is why the quarterback is even more important in Rodriguez's system than others.
- The veer is an excellent counter to Michigan's usual zone stuff… To the line it looks like a stretch and will draw stretch responses. As you can see in the frames above, the Illinois line has crashed itself out of the play, helping Michigan down-block it. There's a big damn crease if the DE heads out for the tailback. If the DE is on a scrape exchange and crashing for the QB, the handoff read is a potential big gain because the scraper is going to have to deal with a blocker and you have a tailback in a lot of space for cutbacks. Michigan tried it a couple times against Illinois; Illinois, unsurprisingly, reacted well to it. It's their base running play, IME. They've seen it.
- …but it requires far more precision on the read. Watching Juice Williams in detail the past couple years has given me an appreciation for how difficult it is to perceive the DE's intent and momentum, and how your fakes can drag him out of position. Williams gets low and extends the ball and holds it there almost impossibly long, then drags it out after the DE commits. Forcier does not have that patience yet.
This read is also more important to the success of the play. If the DE crashes down on a stretch he may get to the tailback if other people on the DT cut off creases. Fundamentally he's a cutback defender and a play can still work if the QB is not contained and gives it off. Here a missed read is probably going to be a loss, Forcier jukes notwithstanding.
I mentioned this earlier in one of the two instances where I brought up Chris Brown's explanation of the differences between inside and outside zone runs. Here's a play featuring the tell a couple coaches suggested I look for when I was complaining about the difficulty of distinguishing between the two.
Michigan's in a shotgun with trips to the right. Two things to note here are the two deep Iowa safeties, and the shift of the Iowa linebackers outside. Angerer, the MLB, is lined up over Odoms, sort of:
Also, Greece has destroyed Latvia in World Cup qualifying.
The thing to note in the above frame is the position of Forcier relative to Minor. Forcier is a yard or so in front of his tailback. For comparison, here's a play against Indiana that would end up a standard zone stretch:
Forcier is a yard behind the tailback. This allows the RB to come across him at speed and get to the frontside creases the stretch looks to exploit.
Back in the Iowa game, the positioning of Forcier allows Minor to take a handoff already headed upfield, which was one of the adjustments that Penn State struggled with so badly last year. Also note a great oddity:
Michigan is blocking the backside defensive end! Why are they doing this? Well, if you don't block him and he crashes down and you're running a play that's anything short of a stretch play that's running away from him there's a good chance he makes a thumping tackle in the backfield. Michigan did this a lot against Iowa because Brandon Minor's RAGE is most effective when he's heading straight upfield.
Another item to note: at the moment of the handoff, Forcier is staring at the MLB over Odoms, judging whether or not he's coming up to contain.
He isn't. And one reason for that may be that this looks like play action. Odoms isn't running a bubble. The backside defensive end is getting blocked. In the past, this has always been a pass, or an attempted one. So Angerer gets a pass drop. By our next frame he'll be hanging out at the first down line, six yards back from the frame above:
You'll note that Minor is running right next to Forcier; with five guys in the box and no support for a hypothetical bounce, Minor could have made this same run. Iowa's decision to leave two deep safeties back makes it really hard for them to stop Michigan's ground game, though it did prevent Michigan from breaking anything long: their longest run in Kinnick was twelve yards.
At the end of the play Forcier has near first down yardage after having slid to the ground untouched. The Iowa defender does give him his best Cato June, though:
Here's the glorious you-tube-o-vision, in which you can see that the receivers' half-hearted routes. That indicates this was a called run play, not an improvisation, in case you're wondering if this was play action gone awry (awright?):
- Zone runs have a bit of a tell. If your depth perception and processing is quick enough and you see the QB step forward you've got a good idea that it's not a stretch. If he stays back you've got a good idea it is. This is probably not a huge deal since the QB takes up his final position moments before the snap, preventing—or at least hindering—the ability for defenses to key on it. It's a lot to process that when you're trying to time the snap and figuring out your assignments and whatnot. It is there.
- But you, the viewer, have a great view of it. TV angles are great for picking this out, though, and it's simple enough that you can try to pick it out real-time.
- RAGE. Michigan went to a lot of interior, non-stretch runs with Minor and blocked the backside DE. This helped out on a variety of plays and should hypothetically make Forcier's job on the reads easier because the guy he's reading is a lot further away and his motion has to be less subtle if he's got contain. This also brings in some elements of Paul Johnson's flexbone, too. Johnson loves to leave a guy unblocked for much of the game, then crush him unexpectedly for a big play.
- Michigan's mixing up its routes on certain keeper plays. I'm betting that if Odoms ran a bubble route on this play that was a key for one of the linebackers to shoot up for contain against Forcier and for one of the safeties to crash down on the bubble. By just running its receivers downfield, Michigan got Iowa to go into pass drops and opened up tons of space for Forcier.
- Iowa loves them some two-deep safeties. The zone read brings in the quarterback as another runner and has essentially forced its opponents to ditch the two-deep look. In the Rodriguez coaching videos kicking around the web, the implicit assumption is that opponents will usually have a single deep safety because of the threat of the keeper. Iowa defies that, and it worked for them, albeit barely. Michigan racked up almost 200 yards on the ground without its starting center and nominal starting tailback despite seeing five drives end on turnovers. Michigan had similar success against Notre Dame last year when Corwin Brown decided to keep two deep safeties. Once Michigan emerges from its freshman quarterback purgatory I wonder if Iowa will be able to get away with this sort of thing.
I said I'd come back to this when I have video, and now I do. In last week's Iowa game, the linebackers became extremely aggressive against the run. This usually worked out pretty well. Iowa had a lot of problems running the ball, and what success they had was usually due to the NT getting blown too far back off the ball for Mouton—it was usually Mouton caught in the wash, with Ezeh flowing to a point farther outside—to flow to the ball. There were a couple instances in which the linebackers zipped into the wrong hole, but all told it was an encouraging performance, especially for Ezeh. Ezeh picked up a +4.5, his first positive outcome of the season.
Here's an excellent example of the linebacker's new aggression. It's first and ten late in the first half. Iowa's got the ball and is playing conservatively. They start in an I and motion the outside receiver in:
Here's the snap. You can see Martin already off the ball. Woolfolk has gone in motion to cover the receiver who shifted; this is man coverage:
A half-second later, Iowa is shifting the line left and running a zone stretch. It's hard to see the line from this angle but from the top to the bottom:
- Stevie Brown is holding the outside against an Iowa TE.
- Craig Roh has gotten sealed inside by his guy. I think this is because Michigan's line was slanting away from the play at the snap and then had to try to adjust. Look at Martin in the picture above: he's heading straight upfield. In the picture below, he's behind an OL and trying to come around.
- Ryan Van Bergen and Martin are in a big heap of bodies, with three blockers trying to take on two linemen. They don't crease and they don't allow anyone to get to the second level, so that's a win for Michigan.
- Brandon Graham isn't doing so hot but it doesn't matter. He may be preparing to shoot upfield in the event of a waggle.
And then you've got the linebackers, who are moving forward already, well before the handoff point. Both of them are headed outside.
At the handoff point, Roh has gotten himself a tad bit farther in the backfield. There are still no creases and no downfield blockers. Ezeh is heading outside into the crease between Brown and Roh. Mouton's waiting a bit in case there's a cutback; his designated hole is somewhere between Roh and Martin:
The handoff's made, and Mouton reads that there's nothing in the middle and heads outside. Ezeh's already in the hole, about to meet the fullback…
…who he crushes:
The key in the above frame is that Ezeh got outside the fullback, forcing the tailback behind him and into the help, which could be Roh or RVB but in this case is Mouton, who's running untouched into the path of the tailback…
…for a TFL:
Here's the video:
In real time you can hear, and feel, the crunching destruction of the pwned fullback. Michigan's been doing this for a while now. Contrast several plays against Iowa and Michigan State on which the linebackers flow downhill immediately with this, the opening play of the Notre Dame game:
Yes, they're flowing to the ball, but the hesitancy is obvious. This happened a few times.
The problem comes when opponents go to play action and two tight ends get wide open at the same time, but I don't know if that's their responsibility. With Michigan going to more man coverage since the insertion of Woolfolk at corner, Mouton and Ezeh can be responsible for the two guys in the I; the tight ends are not their problem. In an ace set, that's not the case, but at least one of them was innocent on Moeaki Disaster II.
I'm not sure if this is better play from the linebackers or Robinson removing responsibilities from them and telling them to go forth to rampage. The multiple times Iowa got guys wide open on play action waggles, and Michigan State's success with tight ends, suggest that Michigan has traded one problem for another here.
This one's pretty simple because the blocking scheme is "hey, receivers, block that guy": it's the edge pitch Michigan debuted over the weekend.
A standard formation with Koger as the near-side slot receiver. Note Forcier's position: he's a yard in front of the tailback—in this case Michael Shaw. This usually means Michigan is running something intended to go up the middle. On pure stretch plays Forcier will be even with the tailback.
Iowa, for its part, is in the base 4-3 cover two they ran the whole game. More about this in UFR later, but if Iowa persists in running this scheme in the future I think Michigan is going to smoke them when their quarterbacks are freshmen who are freakin' out, man.
Here's the snap:
Forcier's got the ball already and you can see Shaw bugging out to the sideline to get a in a pitch relationship with Forcier. There's no counter action on this play, it's just get to the edge as quickly as possible.
A moment later:
Forcier's still got the ball and has hardly moved; you can see by the clock on the field that this is less than a second later. The only things to note here are Iowa's MLB, who's taken a step to the side of the field a stretch would go to, and the defensive end, who has also stepped inside in anticipation of one of Michigan's plays that attacks the backside DE's usual tendency to either crash or head out on the quarterback. His caution, usually rewarded, pulls him out of this play.
A second or two later, Forcier has ditched the ball and is a spectator:
Odoms has whiffed his cut block, unfortunately, leaving a linebacker in space. Iowa safety Tyler Sash is also filling, and the backside DE has reacted to provide some contain.
Shaw gets upfield quickly before the three Iowa defenders can converge…
…and picks up five yards despite Odoms whiffing on one of the two relevant blocks.
- It's hard for this play to not pick up five yards unless the defense is specifically gameplanning for it (which they probably will at some point). Michigan ran this a bunch and the worst it ever did was two yards on second and two, and that was because a receiver ran right by a safety and that safety bounced Minor—who's not the best guy to run this thing—out. At all other times it picked up four to six yards. Opponents will now start preparing for that, which will open up some other stuff, as the defensive end's tendency to slide down the line in an effort to defend the zone counter dive opened up the edge for this play. Cat and mouse goes on forever.
- It's probably never going to break big against a defense like Iowa's. Linebacker versus slot receiver usually doesn't go well and it doesn't develop fast enough to make a cut block, even a really successful one, more than an annoyance when those linebackers are five yards downfield. Then you've got that safety coming downhill unblocked, the backside defensive end peeling back, and linebacker help from the inside. It's weirdly like MSU's power off tackle game, which is likely to pick up 3-5 yards and unlikely to do anything more.
- It's something I bet they wanted to run against MSU, and might work better against an aggressive defense that's using a corner guy as a scrape exchange defender. Iowa plays two deep on every play, which always gives them a safety who can run to the POA and fill. If the corner guy is charging off the slot and sucks in on Forcier, then Odoms can go block the safety and Shaw ends up with a lot of room to run downfield. Or he ends up with that scrape defender in his face. About that…
- This is step one in the evolution of a speed option game. The solution to that is to turn this into a true option play where Forcier threatens to get upfield and takes that scrape defender before pitching, or turns it up himself for yardage. Right now this is just a safe little pitch play that has no read and is easy to run.