Upon Further Review: Offense vs Florida

SPONSOR NOTE: After we did WTKA today Craig Ross told me that he was in a mediation and for some reason they needed data on a refinance. "What the hell," Craig said, and called Matt. 15 minutes later he had a term sheet and his clients were looking at him like he had some crazy connection in the mortgage industry.

Do you like feeling like a big shot in mediations? Or maybe just your own home, wearing or not wearing pants as you choose, on your own, because you are you and nobody else can tell you what to put on your legs? Homesure Lending can do that.

FORMATION NOTES: uh i forgot to take clips just a sec

There wasn't anything too weird except a few tackle over plays. There was an increase in WR snaps. Michigan averaged 2.1 WRs, 1.4 TEs, and 1.5 RBs per snap. Before the 22-personnel heavy fourth quarter Michigan had 2.3 WRs per snap.

PERSONNEL NOTES: QB was Speight with those two O'Korn drives. RB was Evans primarily, then Isaac, then Higdon, and that's it. FB seemed about evenly split between Poggi and Hill.

OL was Cole-Bredeson-Kugler-Onwenu-Ulizio the whole way except for one brief drive where Runyan replaced Onwenu.

WR was fairly diverse. Black, Crawford, and Perry led the way. DPJ, Ways, and McDoom got scattered snaps behind the starters. TE seemed split almost evenly between McKeon, Wheatley, Bunting, Gentry, and Eubanks. Those snaps are probably in descending order.

[After THE JUMP: it's a long one]

FIRE SPEIGHT

Aw, cumong. Chart:

[Hennechart orientation: mouse over column headers for explanations of the categories. + is handed out for a good throw under duress. * is handed out for a very bad version of a bad thing. Numbers in parens are screens. DSR is an attempt to compress the numbers into one overall number. PFF is PFF's grade.]

WILTON SPEIGHT

This was not quite up to his 2016 standard, though I bet this is a much tougher defense than the average one he faced a year ago. It was not far off, and totally lacked a bad read unless you want to ding him for taking a tough corner route when facing cover zero instead of something in the middle of the field.

He was... middling. There's a lot of garment rending in the aftermath; I'm still Team That Guy. The first interception still looks like a good read and good-enough throw after watching it a dozen times:

Crawford isn't even that far off the ground. Rudock missed Chesson far worse a couple years back against this same team and you don't hear anything about it because Chesson caught some of them and the others weren't returned for TDs.

The second INT was obviously awful, and he had one more seven-point miss:

On the other hand: great check. As I mentioned in the game column, Speight's decisions were uniformly excellent. It's the two maximally bad inaccurate passes that are downers in the aftermath. I'm not saying that's going to stop; I do think they're likely to punish Michigan less harshly when he does miss.

When not throwing away 14 points, Speight started to put complaints about his deep ball to rest by dropping some dimes against the Gators. An early PI was an inch perfect 40 yard throw between defenders. The Black touchdown was appropriately short for such a wide open dude. This post to a bracketed Perry is *kisses fingers*:

And he nailed Eubanks in stride late.

He commanded the team well, chipped in on Michigan's RPS win, and looked like a guy this defense can win a lot of games with. This was fine. I think he'll improve.

O'Korn, meanwhile, had an eventful three dropbacks:

JOHN O'KORN

That's one picture perfect fade to Black, one instasack suffered, and one rare TAX when he showed very poor pocket awareness and took an unnecessary sack. This is not really a competition.

Are we concerned about the all or nothing nature of the offense?

This offseason Bill Connelly published a study about explosive plays, asserting that they were the "three pointers of football." This referenced the KenPom study that found that there was no correlation between a team's three-point percentage defense during the first half of conference play and during the second half. If you're an avid reader you probably know all about this already since we bring it up constantly when talking about basketball.

Connelly found a similar lack of correlation with his "IsoPPP" stat that tries to model explosiveness, thus the post. Maize n Brew picked up on this and fretted over Michigan's boom or bust offense against Florida.

I'm not convinced by Connelly's post like I am by Kenpom's just based on the eye test. Once Kenpom showed that three point percentage was essentially random, that clicked. Once you think about it, yes, almost all three pointers are open shots the offense has generated. Closeouts have limited efficacy, no-assist jacks are rare, and players tend to self-censor if they can't shoot. The average three pointer is very very average.

I don't think "offenses with a lot of big plays and a lot of meh ones are actually bad" is quite the assertion Connelly is making, but it has been repurposed as such through the usual statistical inference telephone. And I can't get behind that kind of assertion for football. This game was a good example as to why. Florida's linebackers frequently engaged in an activity a former Northwestern linebacker derisively referred to as "blitzball" when he saw it in his own team:

There is no "read", only reaction. I've seen all the linebackers do it through the first six games, I do not mean to single Walker out. This is how they played against Stanford and Duke, but they were making plays. When they're not, it will cripple a defense versus efficient offenses.

And lo, Michigan's first offensive snap:

This is inherently behavior that leads to boom or bust plays from the offense. (See also: Michigan D yards per completion a year ago.) Florida has super aggressive linebackers and kept shooting DEs to the interior; it was rare that anyone tasked with a kickout block would actually have a guy to kick out. That led to a lot of thunder up the gut for 2 to –2 yards, and a lot of bounce bounce bounce bounce.

Michigan blockers and occasionally running backs were caught off guard by this. A third quarter run saw 1) Poggi eat a linebacker two yards in the backfield and 2) the backside DE giving up the corner. Evans cut back once but missed the bounce:

That is a lot of space between the hashes and no force player; Evans took two yards up the middle. Poggi is the Michigan player in the backfield.

Michigan's bounce bounce bounce ground game was effective, and the way Florida played the run enabled that by jamming up the middle. That's who Florida was last year, too: #9 in success rate allowed at 35%, #42 in explosiveness allowed. D-I average success rate is 42%. Michigan was at 32%. Is that the thing to focus on, or is Michigan's 5.8 YPP against last year's #4 defense? Michigan backs put up 5.3 YPC; the team put up nearly 450 yards.

Connelly's assertion quoted by Maize n Brew:

“The key to explosiveness is efficiency. The key to making big plays is being able to stay on the field long enough to make one.”

They had 19 first downs, which was just under last year's NCAA average (21), so they stayed on the field, against a defense that was great a year ago and will be either great or very good this year. Worried: no. 

Is that why there were multiple third and long conversions on the ground?

The first one was Florida daring Michigan to do it, so they did it. They showed full on press coverage with two deep safeties, and ran it:

That's five in the box and all four press guys not even looking at a run until it's too late. The poor damn color guy is trying to react on the fly and asserts that Onwenu must have wrecked some guy; replay; Onwenu pulls and literally has no one to block. The rest of it is Super Mobile Mason Cole back at the spot he killed it at during his sophomore year. Isaac barely outruns him by the sticks.

This was kind of a thing on third down, where Florida would straight up tell Michigan what they were doing. If they showed all out blitz zero they ran it. If they showed man under two deep they ran it. This worked okay sometimes; sometimes not so okay.

Isaac's offset draw caught Florida in a twist that should have actually worked for the Gators. Cole normally throws a guy upfield of him and then goes and finds some meat downfield; he could not do that here. That left a linebacker who coulda shoulda tracked Isaac down short of the sticks, but he took a bad angle:

So... kind of? Shannon went with a super aggressive anti-pass defense on the first one and Michigan took him up on his dare, and they went with a pass rush stunt on the second that made it a one on one matchup between the LB and Isaac.

Okay but they really couldn't run at all unless a tailback was doing something weird and good.

Michigan appeared to have a severe Kalis issue where guys would run by dudes it seemed like they really really needed to block.

Sometimes this was intentional. Michigan's TD on the opening drive of the third quarter completed Saturday's Ode To RichRod by running a bonafide zone read belly play:

Cole runs by that guy and that guy runs by Hidgon, touchdown, RPS +2, MGoJen tweet.

And I kind of think that some dorf-lookin' blocks were actually Michigan trying to hit dingers by letting guys go so they could try to get 2-for-1s. This early FB trap is nerfed by the MLB refusing to bite on any of the frippery, but the most interesting bit to me was Ulizio blowing by a guy who is lined up directly over him to go get a second level block. Dude runs free and still runs by Hill:

So is that fine? As designed? Maybe? A couple Wheatley blocks seemed like they were in the "hit dingers" category. This jet sweep features a Very Mason Cole downfield block but for this section keep your eye on 17:

He runs by the DE and at no point checks him, and that DE strings out a play otherwise headed for a huge gain. Is that an error, or is Michigan banking on the DE biting harder? It feels like the former, both because Wheatley executes confidently and holy hell that DE closed down a ton of space. Later in the game Wheatley would basically chip the backside end on a power play and then go look for work downfield:

#17 to bottom of screen

That looks intentional too, and when he releases he does have guys to block. Unfortunately, the DE makes up that ground and tackles. It's like Michigan found that Khalid Hill block against Colorado and asked themselves "how do we do more of that?" Michigan's offset draw is another great example: it's so successful because the left tackle gets a two for one by flinging a DE and then going and getting a downfield block.

I feel there's a "but" coming.

Sometimes it didn't feel intentional but was so consistent it might have been. I had a lot of Ulizio minuses because he'd run by a DE and that guy would subsequently run by a TE with basically no shot at making good. Here there's a decent cutback lane that Isaac can't explore because the backside end is rampant:

This happened all the time.

I didn't get it, what with the other side of the line executing textbook scoop blocks. Best I can figure is that UF's approach was baffling to Michigan and that they failed to adjust. Because they needed to adjust: not one of these blocks got made when a tight end had to check a DE lined up inside of him and slanting away.

Ulizio did stay in to help some later in the game. Here he helps on the DE; once provided with an achievable target Wheatley puts the dude in the next county, opening up the bounce:

The lack of that double above blew up Michigan's ground game more than anything else. I hope that's a very specific Ulizio problem they can get fixed.

Was Ulizio really that bad outside of that though?

Yes. Run chart:

[ed note: we do not have PFF this year, please forgive the goofy table that it's too late to fix due to Acts of God this week]

No run adjustments this week. Runyan was +1 in three snaps and that is included.

So. As mentioned above I did not get Michigan's blocking scheme at all and heaped a ton of minuses on Ulizio for leaving guys at the LOS the tight ends had little shot at. Some of those may be more properly put on RPS or the tight ends. I have a tendency to minus anyone who blows by a dude at the LOS and gives his teammate an impossible job, and Ulizio did that a ton.

As for the rest of it: Ulizio had to be terrible and he was. Redshirt sophomore, first start, barely entered the conversation until a couple weeks before the end of fall camp: had to be bad. Was bad. I saw some hopeful assertions that he was functional on the ground and bad in pass protection, and I don't think those are accurate.

The pass protection bit was. Obviously:

He gave up another sack when O'Korn was in when an edge rusher got around him at 7 yards.

He also had some bad early Braden lunges on the ground. Here he airballs:

It was bad.

Now: first start against Florida and their weird aggressive don't-care-about-leverage defense. Ulizio got picked because he's actually a tackle, unlike Runyan, and isn't a fourth year player who already struck out a couple times against Rutgers. He will improve. Enough? I have no idea. The most interesting thing about Saturday is going to be his performance.

Onwenu has some pretty hefty numbers himself.

Onwenu had a rough, rough start. He was briefly yanked for Runyan after the first pick six; at that point he was –3.5 in my charting with a pass pro minus. Projected over a full game and that's... not good. Michigan pulled him to yell at and/or settle him, and he improved a great deal afterwards. He climbed above zero, and post-pull his ratio hit our 2/3rds Mendoza line.

After the rough start, Onwenu's debut was about as good as could be expected. He moves people. There is no question that any back seven member is going backwards a long way on contact:

He pulled on the Higdon run before his TD and even though he was off balance there was no option there for the DB. Backwards it is, thank you sir. A third and one fullback dive was an easy four yards because Onwenu deposited the nose tackle two yards downfield.

Onwenu suffered early because he and Ulizio frequently got hammered by Florida's awesome three tech; he also suffered late. In between there were some pancakes and a lot of guys moving downfield. I'll take that for a true sophomore, especially because I had him with one pass pro minus. You could maybe give him another one or two for a stunt or two that got through; at no point did he seem out of his depth. You can't bull rush him, and he's agile enough to mirror that three tech most of the time.

You mentioned Cole back at that spot tho?

Last year was a bit frustrating as the excellent Cole things were muted by the center move and Michigan's general zone incompetence. He was hypothetically a great zone stretch C; we never got to see it. Now that he's back at tackle he's once again executing the crazy mobility blocks he did two years ago.

The third and long conversion above is one. He executes the offset draw block, hurling a pass-rush focused DE upfield and then motoring. He checks both sides like he's a six-year-old crossing the street and then thumps a safety ten yards downfield. That was a tiny bit like seeing Denard rip off a 70-yard TD after that Alabama debacle: reunited and it feels so good.

And even though he's at tackle there were a couple of righteous scoop blocks that presage good things. He and Bredeson got after it. This one is totally irrelevant:

It is also impressive. Bredeson gets a thumping hit in and by the time the guy recovers he's been replaced by Cole. That scoop was equally impressive and more relevant on the Isaac run that helped set up Michigan's shoulda-been first touchdown.

Cole loped downfield and wiped a bunch of dudes out in this game; he did not have the same issues IDing who he should hit as the guys on the other side, and he got to zone it up with Bredeson. He was outstanding on the ground.

He was equally good as a pass protector... until the last snap, when he got beat clean and gave up a blindside sack. I'll still take that against a team with DEs as hyped as Florida.

And the other two OL?

Bredeson took a massive step forward. The most important thing for him is the mental side and I only had 2.5 minuses for him. Even last year he was very good when he knew what he was doing; the early parts of this year will be an inflection point in his future. If he's as with it as it seemed in this game he's headed for All American-level play.

Kugler meanwhile was close to the best case scenario: very few biffs that seem traceable back to him, especially in pass protection, and an effective day on the ground. He got UF's nose tackle moving frequently.

Figure out right tackle and this is going to be a damn good OL? I think?

Running backs?

RB chart (WR grades are run only):

This was a good game for Isaac, what with the bounces being open. He's always been majestic when allowed to get up to full steam in the open field—remember Isaac vs Peppers in the 2016 spring game—and he was here as well.

This was the best game of his Michigan career, considering the opponent. He took and hit all the available bounce opportunities and turned 3 yards into 18 on the 4th and one.

I'm willing to upgrade my Isaac take to "I'm saying there's a chance."

Evans is still the guy and got 22 carries in this one; many were doomed and a few rescued. I thought he missed a couple bounce reads, but he shook a safety one on one and also had this tight-area make-em-say-uh:

Higdon had a couple of good runs interspersed with one major WTF cut.

Michigan's fullbacks were good—Poggi was a lot better, and even made some decisions his tailbacks would have been wise to follow. Hill didn't do a ton but did open up one of the bounce bounce bounce runs with a hammer panda block:

That's a 6'6", 230 pound guy per the roster, so that was never going to end well for him. Sadists will also enjoy Hill versus a cornerback on a late crack sweep.

Receivers?

A couple of drops mar their day.

[0 = uncatchable, 1 = circus catch, 2 = moderate difficulty, 3 = routine]

ROUTES: Black +, McKeon +, Bunting -

Perry was the story for me. Perry is extremely legit and Speight trusts him. This was apparent in last year's OSU game and it's apparent here, as there were a couple of third downs where he did his Route Artisan thing and moved the chains. He gets yanked back here and still extends to catch:

That ability to extend and catch at the edge of his radius came in handy later when he was the target on the *kisses fingers* downfield shot discussed in the Speight section. Those are a couple DANG catches. He's reliable, he's able to put some dang on it, he's viable downfield, he's going to lead Michigan in catches this year. Probably.

I wish Speight trusted Perry more, actually. When UF went cover zero on that third down after the fumble I saw Perry in the slot and was like YES HIM YES HIM; Speight chose one of those close but not quite fades. Perry torched his opponent for an easy throw over the middle on one of three or four bad Speight decisions on the day. Notably, the next time that guy matched up with Perry on cover zero he gave him three more yards of cushion. Put some respect on it, son!

Black had a mansome block on the Isaac offset draw first down.

Tight ends?

McKeon had some promising moments; he also looked like a redshirt freshman. That redzone run by Evans looked for all the world like a designed cutback run on which he blew his assignment; Onwenu and Kugler clobbered a guy and Evans was headed for that hole until he saw an unblocked guy in it. Meanwhile McKeon isn't really doing anything as the TE to the bottom of the screen:

Wheatley got a lot of stick from some parts because he was the designated "chase a DL to the ballcarrier" guy on a lot of bad zone decisions from Ulizio.

Gentry wasn't targeted. He did have some plausible seeming inline blocks.

REFS ARGH WHAT

Yessir. We should be deeply grateful that Michigan won this game; if they did not having four points taken off the board because of the worst call in a Michigan game since the Michigan State pylon event would have been a giant raspberry to start the season.

In addition, that delay of game penalty was because the clock operator reset the playclock to 25 seconds and nobody noticed.

On the other hand, Michigan got away with two obvious holds. So it wasn't all bad.

Any Harbaugh stuff?

There wasn't much that jumped out as weird or odd or goofy or even particularly new. My best a-ha moment in this game came on back to back runs towards the end of the first half. On the first Michigan splits Poggi out as a tight end and tries to combo through UF's rampant three-tech. He explodes, as he is wont to do, and both Ulizio and Onwenu get stuck on him. If Onwenu leaves the three tech tackles for loss; if he doesn't there's going to be an unblocked guy in the hole:

watch 50, the right guard

Evans bounces away from this and gets one yard.

On the next play Michigan deploys a slight tweak. They draw Poggi into the backfield and run an iso. Nobody has to combo through anyone. Onwenu has license to focus exclusively on the three tech, and Poggi will take the MLB. Onwenu buries that guy and there's a huge gap for Evans to hit:

I'm pretty sure UF's front was doing the same thing on both plays; this was a pleasing demonstration of how even something so frustrating can be exploited. The only other run stuff was some tempo on the first drive of the second half, and the Ode to RichRod.

There was some Pep Effect in this game, most notably on this second and ten when Florida told Michigan they were running zone and linebackers would have to check Perry. They did not:

An easy, quick eight yards that Speight knew presnap. See also: pick six #2, except with an accurate throw. No throws to the backs yet, though.

Heroes?

Cole and Bredeson; Isaac. Hill and Wheatley.

Maybe not so heroic?

Ulizio gave up two sacks and looked lost IDing his assignments. The second year tight ends all had at least one –2 block. Speight was okay but did dorf 14 points away.

What does it mean for Cincinnati and the future?

Best case scenario for the left side of the line. Cole was great on the ground and only got hit once in pass pro; Bredeson does look like he took the quantum leap his place on the depth chart suggests; Kugler was savvy and powerful enough.

Onwenu got a lot better in the second half. I didn't know you could pull an OL for being rattled and have it work. Bonus take: Runyan is your sixth OL no matter what.

Let's reserve Ulizio judgment. He was a mess; he had to be a mess.

Young tight ends are talented and inconsistent. Eubanks just hauled in a deep ball after outrunning that Florida speed. He also airballed a block. McKeon looked like a very natural receiver... and airballed three blocks. None of that is surprising and most of it will get figured out.

Michigan's two headed tailback might have three heads. You have our attention, Ty Isaac.

Speight is probably fine but may always be streaky. Push the panic button if he keeps giving away huge plays.