Full breakdown: Detailed analysis of Michigan's play selection against Army
There's a lot of questions about what, precisely, Michigan's offensive gameplan was against Army, who was directing the playcalling, and what the results were.
So I felt it would be good to move past the anecdotes and analyze the data. Here, then, is a complete rundown of Michigan's play selection for regulation time (OT is a weird animal and is not included), as taken from the espn.com play-by-play tracker. I omit dead ball penalties and Michigan's fake punt. Note that this does not fully appreciate RPOs, nor does it account for called pass plays in which Shea may have scrambled for yardage. Plays listed as sacks are considered passing plays.
First Half:
Drive One:
7 plays
5 passes, 2 runs.
4 first down plays
2 first down runs, 2 first down passes
Result: sack/fumble
Drive Two:
11 plays
6 passes, 5 runs (fake punt not counted, PI call counted)
5 first down plays (including 1st & goal at the 2)
3 runs (including 1st & goal at the 2), 2 passes
Result: TD
Drive Three:
1 play
1 pass
Result: sack/fumble
Drive Four:
6 plays
1 pass, 5 runs
4 first down plays
4 first down runs
Result: fumble
Drive Five (2-minute offense, sort of):
12 plays
7 passes, 5 runs
4 first down plays
1 first down run, 3 first down passes
Result: Missed FG
Half totals:
5 drives, 20 passes, 17 runs, 37 plays total. 1 TD, 3 TOs, 1 missed FG, 1 (successful) fake punt.
18 first down plays: 10 runs, 8 passes
Second Half:
Drive Six:
3 plays
2 passes, 1 run (one McCaffrey pass)
1 first down play
1 first down run
Result: Punt
Drive Seven:
12 plays
5 passes, 8 runs
5 first down plays (including 1st and goal at the 2)
5 first down runs (one may be a scramble on a called pass)
Result: TD
Drive Eight:
11 plays
5 passes, 6 runs
4 first down plays
1 first down run, 3 first down passes
Result: Downs
Drive Nine:
8 plays
8 runs
3 first down plays
3 first down runs
Result: Downs
Half totals:
4 drives, 12 passes, 23 runs, 35 plays total. 1 TD, 2 TO on downs, 1 punt
13 first down plays: 10 runs, 3 passes
Game totals: 9 drives, 32 passes, 40 runs, 72 plays total. 2 TDs, 3 TOs, 2 TO on downs, 1 missed FG, 1 punt, 1 (successful) fake punt.
31 first down plays: 20 runs, 11 passes.
Analysis:
Michigan got noticeably more conservative in the second half. I would suggest that things may go further: There appears to be a significant change after the strip-sack in drive three, Michigan's second turnover. On the subsequent drive (drive four) Michigan runs 5 out of its 6 plays, including every first down. The following drive was a "2 minute" drive that included a surprising amount of running. Michigan then produced the startling second half play selection that is recorded above.
Prior to the second lost fumble, Michigan's offense was actually quite diverse. There was a very close balance between running and passing, including on first down, there was lots of yardage gained, and the second drive produced a touchdown. The 2-minute drill balances the numbers somewhat, but even then, Michigan ran an unusual number of times for a 2-minute offense.
Michigan's only balanced drive in the second half was drive 8, which ended in the first turnover on downs, but proceeded as far as the Army 19 prior to whatever that mess of a play was supposed to be.
Whatever prompted it, Michigan contracted into a massively conservative shell after the second fumble, and only tentatively emerged from that shell a couple of times. Zach Charbonnet is a good player, but he is not the second coming of Adrian Peterson; nevertheless, he received 30 touches in regulation, an astonishing total for a freshman. One would normally suspect such numbers only on teams for whom the RB was the only solid offensive weapon on the roster, rather than a team with three or four potential NFL receivers.
It bears noting that this playcalling philosophy does not resemble the philosophy used at Alabama last year when Josh Gattis was on staff. Something caused someone on Michigan's staff to radically alter the gameplan mid-game. Whether it was Shea's health, turnover concerns, or something seen on film is unknown. Also unknown is who made the decision. However, the numbers make it clear that it was not just an accident or a coincidence. Michigan's offensive contraction was intentional.
September 8th, 2019 at 12:30 AM ^
Who has the guts to ask this on Monday presser to Harbaugh ?? DO IT.
And if Harbaugh dictated this pace , we are never winning anything anymore of significance. Of this I am certain. It is over.
I want to hear what Brian has to say about this. Hard to spin this performance other than crap.
September 8th, 2019 at 12:42 AM ^
I was at the game and I concur. As odd as this will sound, Army’s best Hope was Michigan turnovers and a shortened field. It makes their brutal efficiency almost lethal. Michigan just ran inside zone so often after the third fumble that it felt like they were just trying to power their way to the finish line. Almost as if someone just said,”f this - we are bigger than them. We will slow the game down and limit THEIR opportunities”. I feel like Gattis lacks credibility with Harbaugh so the minute it gets tough, they fall back to what they know. I will also say Shea looked visibly injured. Passes sailed or were off target a lot and he NEVER ran the keeper except once which he came off the field after being tackled. If you can’t threaten the edge with the runner, then all of the RPO action suffers.
September 8th, 2019 at 1:07 PM ^
FWIW, he came off the field because he was cramping. He had been stretching between plays the entire series. I don't think this issue was related to his injury.
September 8th, 2019 at 1:19 PM ^
I mean, you solve the mystery of this OP in your post:
1) Army’s best Hope was Michigan turnovers and a shortened field
2) Shea looked visibly injured
There is no unknown here. Army, as Seth outlined all the way back in Hail to the Victors, has an endless playbook of blitzes that their very smart football team will throw at you. They hope you fuck up with mistakes and turnovers, their offense possesses the ball, and the game ends before they run out of them.
Michigan fucked up and turned the ball over. And the QB was injured, which likely took away the QB keep, caused some errant throws, and led Michigan to not want to expose him to hits from the blitz on deep balls. Yes especially in the 2nd half after it was proven the OL was having trouble with the Army scheme. That equals yesterday.
September 8th, 2019 at 1:43 PM ^
Which is why Patterson never should have played. Michigan still could have run it down their throats but with bigger chunks with McCaffrey as the QB. The threat of someone who will actually run would have put Army into conflict defensively.
September 8th, 2019 at 3:18 PM ^
That's assuming that McCaffrey can bring what Patterson does as a passer. That's far from certain at this point, given that he's attempted 18 career passes.
Patterson was accurate most of the day throwing the ball, including in the first OT when he hit Bell in stride for the first down. He appeared to be cramping up in his legs but it didn't seem to affect his ability to plant and throw on most passes (though possibly on his last one).
September 8th, 2019 at 3:26 PM ^
That would be true unless putting McCaffrey in the game and running would have caused a change in Army's defense, which it very surely would have.
The coaches aren't dumb, and the solutions aren't nearly as obvious as posters on this board make it seem to be.
September 8th, 2019 at 4:16 PM ^
Like I said, it would put Army’s D in conflict. They would have to either put someone on McCaffrey (which they didn’t do with Patterson) which would open up things for Charbonnet. If they refused to cover McCaffrey, then he could run all day. It’s amazing you people think McCaffrey can’t throw the ball. I really don’t know what most of the people look at when they are watching Michigan play.
September 8th, 2019 at 5:00 PM ^
People aren’t saying that McCaffrey can’t throw the ball, they’re saying he’s attempted 18 passes in his career and the coaches know more about the team than random MGoBlog posters do.
September 11th, 2019 at 1:33 PM ^
I have to chuckle at the hate at running.
I have bumped into people in real life who honestly seem like they would be happier if Michigan lost but we played McCaffery and threw alot more.
Going into a shell can be frustrating, no doubt. In the past I got frustrated when Michigan would play games and pass and get incompletions when they had good lines and great backs.
But going dominant with one form of the game or other isn't always wrong.
Having an injured QB and a team that is freaked out by fumbles may be a *great* reason to try to limit your risk and rely on superior talent on the line.
September 8th, 2019 at 12:47 AM ^
Awesome I was just working on something similar but don't have enough points to make a diary-- I'd add to your analysis that Michigan's success rate is MUCH higher on drives where they have fewer first down runs and more passes (either balanced or skewed towards passing). Their success rate was 56% in the first half with 58% passing plays and averaging 6.3 yards/play. In the second half they ran the ball 67% of the time, include 77% on first downs, and and only had a 34% success rate averaging 4 yards/play.
I completely agree that they abandoned their gameplan, which despite the turnovers was working quite well. It seems like Gattis is on a short leash, but whoever (let's be honest, probably Harbaugh) abandoned the plan has to be smarter and not throw out the gameplan because of a few mistakes.
September 8th, 2019 at 1:17 AM ^
Hi - Thanks for doing the success rate %. I don’t know much about the success rate %. I’ve got a question about it though - does the success rate take into account how successful a successful play is, and how bad an unsuccessful is? For example - I would presume a 0 yard run and a strip-sack-fumble are both counted as unsuccessful plays, but one is likely much more detrimental to winning than the other so they count differently? On the other hand, a 7 yard pass to get a first down and a 3 yard pass for a TD is also counted different?
For what it’s worth, we lost the first half 14-7 when we were diverse with the higher success rate, then won the second half 7-0 when we were in the “offensive shell” and the lower success rate, and then won OT 10-7. I guess I kind of want to know if we were still in an offensive shell during OT, but I’ll have to look at the box score myself.
September 8th, 2019 at 1:26 AM ^
Answering my own question -
first OT was run, run, pass (first down), pass (pass interference first down), run (touchdown)
second OT was 3 passes, No first down, and FG.
So OT was in total 3 runs 5 passes.
September 8th, 2019 at 11:08 AM ^
And those passes in the 2OT were not even close. Shea looked baaaaaad there.
September 8th, 2019 at 1:42 PM ^
That's harsh. The first OT pass was just a step beyond Collins's reach - that happens, it was about a 35-yard throw from where he was to the back of the endzone.
The second was very catchable and Black dropped it. I thought Black mistimed his jump there.
The last pass was bad, yes.
September 8th, 2019 at 11:41 AM ^
Success rate doesn't take into account the degree of success. It's just a binary successful/unsuccessful label for each play, and then the fraction of plays that were successful. I'm not aware of a single stat that would factor in all of those variables (turnover versus not turnover, play resulting in a TD versus not, etc.).
In a football game TOs and TDs are relatively rare events so if you're trying to predict something you're probably better off focusing on the more frequent events that precede the really good or really bad events. We know that sacks are more likely on passing downs, so measuring how well a team does staying out of passing downs (which success rate measures) should predict the long-term likelihood of sacks. Same thing with TDs: a team that stays ahead of the chains has an easier time converting yards into first downs, and will ultimately score TDs more regularly.
September 8th, 2019 at 9:38 AM ^
... but whoever (let's be honest, probably Harbaugh) abandoned the plan ...
I agree, it has to be, because this was the exact same shit that insiders said was happening last year. They get behind, the gameplan gets thrown out and things go pear shaped, and everyone can tell. Turns out it wasn't a Pep thing, it's a Harbaugh thing, and that doesn't lend itself well to having a truly elite program.
September 8th, 2019 at 8:16 PM ^
I think the idea that Harbaugh overrides his OCs all the time is a bit overstated; most of the evidence seems to be "the offense plays poorly so it must have been someone forcing bad calls" as opposed to anything tangible. I like the breakdown here because it shows how balanced Michigan's offense really was; even with them running 8 straight times to end the second half they still basically were 50/50 playcalls. And this is against a team in Army that is designed to make you pay for turnovers and Michigan already had 3 of them; I don't see why it would be crazy for a team to adjust mid-game and realize that the only way Army is going to hurt you is if you give them the opportunity to do so, and play accordingly.
Again, this wasn't a good game by any means. But if Patterson isn't going to run the ball and McCaffrey is for some reason not an option at QB, this is the system you probably should run against Army. It sucks to watch, but it also wins you the game.
September 8th, 2019 at 7:16 AM ^
FIX THE DAMN ROADS.
September 8th, 2019 at 7:50 PM ^
Hey, no politics here, this is a sports blog
September 8th, 2019 at 8:29 AM ^
Good write up and break down. Not sure what to make of it or the game.
September 8th, 2019 at 8:40 AM ^
It's so frustrating that we never seem confident enough to let it rip. We are always handcuffed by having to be worried about injuries, concerns we can pass pro, someone's health, or not wanting to put things on film. Will that ever change? What needs to happen to not handcuff the offense? It can't be talent. It can't be experience. It can't be depth. If not now, probably never :/
September 11th, 2019 at 1:37 PM ^
I agree but I suspect that what has to happen is good solid play.
A first half with lots of completions and no turnovers means UM likely isn't going to turtle up the way they did.
If you are the coach, and your recievers are having a hard time catching, and your QB is playing poorly and the team in general is having ball security issues, when it comes down to it would you be willing to say 'More of the same!'
September 8th, 2019 at 9:21 AM ^
Even more in the second half than the play calls was the fact we were snapping the ball under 10 secs most of the time. Is it possible That the coaching team just decided in a one score game the only way we are going to get a big play out of our def is if we do a better job resting them? I can’t find the stat but I know we lost the TOP pretty bad in the first half and came a lot closer to even by winning the second half.
September 8th, 2019 at 9:51 AM ^
The TOP was 31:35 Army/28:25 Michigan. This is insane by Army standards, I'm pretty sure I heard their average TOP last year was something like 38 minutes. This is also one of the things that gives me so much pause about this game - this was not OU/Army where OU tore up the Army D given the chance but just didn't have enough chances. I will say this diary made me feel a bit better as it felt like we had more than 9 possessions in regulation.
September 8th, 2019 at 9:29 AM ^
Thanks for that info.
Almost seems that if they would've had the same old 2TE+FB offense with Charbonnet running it 33x we would have won easily but they decide to spend all summer blowing smoke up our skirts about the new hotness.
I'm embracing the suck now, that's for sure.
September 8th, 2019 at 9:49 AM ^
Harbaugh comes from the Bo Schembechler School of conservativeness. It is what it is, it's what Michigan hired, it's what Michigan wants to be. Give up hope of being an attacking offense its just not going to happen. The program lives in the shadow bo and will continue to do what he did 50 years ago.
September 8th, 2019 at 10:35 AM ^
Insightful
September 8th, 2019 at 7:52 PM ^
Insert hang wringing.
September 8th, 2019 at 10:05 AM ^
Do you think the shift to a largely run oriented Offense was, at least in part, an attempt to use some time to give the Defense some rest? They didn't seem to be able (or want/trust) to substitute liberally on D, and it was a hot day.
September 9th, 2019 at 2:07 PM ^
That's exactly what I was thinking! If they would have had a few quick 3 and outs, the defense would have been gassed!! With Army willing to slog it out, that puts the defense out there for a long time!
September 11th, 2019 at 1:39 PM ^
Stop inserting strategy! We're assigning blame and sharpening pitchforks here!
September 8th, 2019 at 10:37 AM ^
They were scared of Army and decided to play for overtime, thus the switch.
September 8th, 2019 at 1:48 PM ^
Why would we go for two 4th downs in the last 10 minutes of the game if we were playing for overtime?
September 8th, 2019 at 10:37 AM ^
Thanks for doing this. I essentially read an excel column with great interest, may intimate bad things about my mental health.
Doubtful that we will hear more in the early week pressers to decide whether the 2nd half play calls were driven by Shea injury to Harbaugh's conservatism. Maybe both given how badly his throws were sailing...
September 8th, 2019 at 11:48 AM ^
I'm not sure what to make of Shea's 2nd half/OT. Patterson's accuracy and success rate have gone down in the second half of both games so far. If he's hurt, I suppose it's possible it gets more sore as the game goes on? I actually am a Shea supporter but am confused as to why we wouldn't play Dylan more if Shea is hurt and it's affecting his ability to throw and preventing him from keeping the ball on read options.
September 8th, 2019 at 3:15 PM ^
I think Harbaugh fears a QB controversy will ensue. You don’t want the starter looking over his shoulder fearing a short leash. I get that, but if a guy is hurt and that limits the playbook, then controversy or not, get the next guy in. Unless he’s not nearly ready to run the offense, but we’ve never heard that from the coaches.
September 8th, 2019 at 5:57 PM ^
I completely agree with this. What made our run game dangerous at times last year was the threat of Shea pulling and taking off. Even when fully healthy, Dylan is better at this than Shea. If Patterson was injured and they were protecting him by running a lot, this would have been the time to bring Dylan in for a series or two and go with a run-heavy read set. It's not like he's incapable of throwing the ball when called for.
To be clear, I am not calling for Patterson to be benched. I think he won the starting job for a reason. I just think that in the second half one or two series of McCaffrey might have produced a desperately needed score.
Coming out of halftime I texted my brother (sitting in a different section) that I'd like to see us start with McCaffrey for 1 series and no matter what happened with it Patterson gets at least the next two series.
For whatever reason Patterson didn't pull and run when it was clearly there. He had success with this read last year so it would seem to me that it was a coach decision to have him handoff every single time. If the QB isn't going to read and potentially pull and run then this formation doesn't work very well. Might as well just run power like last year if they aren't going to take advantage of the read option. Maybe it was execution and poor decision-making on Patterson? But after what he did last year with this, I am guessing it was a coaching decision.
September 8th, 2019 at 11:06 AM ^
Thanks for doing this SRJK.
Yeah, we don't know. Good analysis. If I had to guess, what caused the change was the health of both QBs. Possibly something pre-existing that was aggravated during the game. There were some bad misses in those first couple of drives.
But that's just a guess.
September 8th, 2019 at 11:42 AM ^
They may have been “quite diverse” in the early going, but the calls still lacked creativity. I cannot stand the stupid clap snap count. Not only is it easy to time on defense, but not having another snap count is awful coaching. I keep waiting to see something new and creative, but it never comes. Harbaugh clearly took the play calling duties away from Gattis after the second fumble and you could tell based on how mind numbingly dull the play calls were. I wonder what Jim thinks when he sees teams like OU, Clemson, LSU, or Texas... I long for the day when we bring in an offensive savant and can actually look like an offense out of the 21st century.
September 8th, 2019 at 2:19 PM ^
"I long for the day when we bring in an offensive savant and can actually look like an offense out of the 21st century."
Short memory, as we tried this in 2009.
If the memory of trying to run a spread offense with Threet and Sheridan platooning at QB has left you, then I am envious.
I'm also not sure it's possible in this era to be substantially, consistently better than 9-3 without pervasively cheating.
September 8th, 2019 at 1:05 PM ^
Nice write up. I disagree with your second half analysis a little though and your conclusion they turtled after the 2nd fumble/strip sack.
Rewatched the 2nd half twice on the condensed game video. The second drive (drive 7) was more run then pass but that was because they ran more when they got close to the goal line. They moved down the field that drive primarily with passes. The runs on the final drive in regulation (drive 9) were likely because someone (probably Harbaugh) decided to strategically to run the clock down to put Army’s offense out of position to run their offense. Klatt suggested this during the broadcast. Disagree with the philosophy if you like (I do) but it doesn’t seem like a reversion to the offense we saw last year. They still ran the ball using concepts we’ve seen this season and if Patterson kept on some of those runs or (earlier runs) they’re likely more open.
I think the issues with the game are this:
The team had 9 possessions 3 of which ended in turnovers. Of the remaining 6 they scored on 3 and got to Army’s 19 on a 3rd. The limited number of possessions really affected how the game and offense looked. The other issue is the QBs didn’t keep ever. That’s either a strategic decision (arguably a poor one) to hold things back or Patterson is injured. My guess is it’s 80% the latter. They threw the ball and used speed in space concepts throughout both halves until about the 5 min mark in the 4th. Problem was drives ended in turnovers and also turnover on downs due to a failed short yardage play.
Thats what I hated most about this game: we still don’t know if the offense is bad or is good and just had a bunch of unforced errors. Last year after Wisconsin I knew they were a good team. They could lose and we may not have known if they were elite but we knew they were good. Right now I don't know and nobody will know till we play Wisconsin. I would bet that offense looks more functional than what we’ve seen though. 2 weeks with a bye is a lot of time to figure shit out and get healthy.
September 9th, 2019 at 2:19 PM ^
We still don't know whether the offense s bad or not but then again, neither does Wisconsin. By all accounts going into this season we knew that Wisconsin in Camp Randall was going to be a tough game. If you have 2 games that you are projected to win handily, why show Wisconsin your entire playbook. Let them see just the basics, before we play them. As you said though, we will know a lot more after this next game.
September 8th, 2019 at 1:06 PM ^
Great job with this breakdown. 100% agree with your last two paragraphs. Think you hit the nail on the head. Whether it was Shea's injury or Gattis/Harbaugh shifting the game plan something was definitely amiss with the offense. Let's hope the bye week they get this settled.
September 8th, 2019 at 1:50 PM ^
I had this very thought during the game. Whether Shea's injured, coaches lost confidence in his ball security, or simply didn't want to put too much stuff on film, the second half playcalling contrast was apparent. So much so that I wondered aloud if Harbaugh took over the reins to establish some run game dominance.
That, and whether Shea was doing his best John O'Korn impression (sorry John O'Korn).
September 8th, 2019 at 2:06 PM ^
Through the first overtime, Patterson was 19-26 for 207 yards and no interceptions. (We also drew three or four pass interference penalties.) . The 0-3 in the second OT drags down his completion percentage and YPA numbers significantly. Up to then he was having a solid day passing, the fumbles aside. But those two fumbles loomed large in a game in which we had just nine possessions in regulation, and they seemed to frighten our playcalling into a shell.
September 8th, 2019 at 2:15 PM ^
The tale of two halves wasn't just the run/pass proportion flipping toward heavily run, it was *how* they ran.
In the first, nearly every run had a read component (even though Patterson gave every time). They all but scraped that in the second and the defense got to tee off on the RB much more. Many plays didn't even have jet/orbital motion like Gattis likes to keep the ends from crashing down as aggressively.
September 8th, 2019 at 3:24 PM ^
Thanks for the analysis and write-up. We complain (with some justification) about the more conservative approach later in the game. We would also complain (with some justification) if the O had continued to use a more wide-open approach and continued to turn the ball over. Seems like a pick-your-poison situation. I thought that on passing plays in the second half, Patterson tended to do what O'Korn had done--look for the primary receiver, and then bail out if that receiver was covered--but I don't know for sure, because on TV, I couldn't see all of the routes at the same time. Better execution will be important, going forward. Especially ball security, of course.
September 8th, 2019 at 5:13 PM ^
I agree that the passing heavy offense got shelved after the turnovers and Michigan decided to play TOP game in the 2nd half focused on ball security. The reasons have already been covered by others, so I won't harp on that any further.
What really frustrated me in this game was that they went into that shell without adjusting some of their other decision making. At least one of the last two drives in regulation stalled out well within field goal range and I think taking the points there wins it in regulation since possessions were becoming scarce on the account of both teams running the same "hang onto the ball forever" strategy. If you're playing 1950s football - play 1950s football dammit!
I also agree with some posters suggesting that a run first, run second offense makes more sense with McCaffery behind center. From these eyes Shea has a better command of the passing offense but for various reasons they're not willing to let him make the keeper reads. A shotgun running team needs the QB to be a threat to read and react - whether than be utilizing jet action, RPO, or running keeps. The combination of play calling (mostly standard Zone Read with no RPO action or jet over it), and the lack of options let Army tee-off on Charbonnet - who is excellent when he doesn't have 9 defenders all locked on to him.
Like many others I suspect that the TOP offense and keeping Shea in the game in the third quarter were both Harbaugh decisions. The combined consequences of those decisions (along with not taking a field goal) were agonizing to watch.
September 8th, 2019 at 6:33 PM ^
If Shea is hurt, put in McCaffrey. If Shea has lost his edge, put in McCaffrey.
Unless of course D-Caff is so far behind Shea that things would be orders of magnitude worse than now. That may be true, but it's not what we've been led to believe (which of course means nothing, but still....)
Remember that JH only makes a change at QB when he absolutely has to....I think it will take at least one loss for him to do this, and that's assuming DM is ready. Only when he has no other recourse will he change.
All I know for sure is that we won't win anything significant with the gameplan we saw on Saturday. We don't have an all-world defense this season. Thanks for the work SJRK.
Comments