This is going to be a bit disjointed. I'm just not in the mood to re-litigate a bunch of things after a win where I can sense a loud minority of the fanbase is already sharpening its knives for 2 weeks from now. Wisconsin will answer a lot of questions about this team's ceiling, and trying to read too much into an OT win against a good service academy feels like a recipe for madness.

Worst: Tight Enough to Make Diamonds

They might as well play this clip before every game a service academy plays, right after the national anthem. Think of it as a PSA:

I know I shouldn't have been surprised how this game felt - I sort of predicted this in last week's diary:

I see the Army game being another of those meandering wins, where Michigan is never really in danger of losing but also aren't blowing the doors off due to a combination of missteps and Army's style keeping possessions down. It's not going to be a game for those who have anxiety about these types of things, but considering that encompasses most of the Michigan fanbase it'll likely just mean a lot of people gritting their teeth for 3 hours and then collectively exhaling at the end.

I guess the part I was off about was just how many mistakes Michigan would commit in order to keep it closer than it probably should have. Yes, Shea Patterson and the Michigan's offense seemed to dismiss the very notion of "reading" a defender on the read-option, resulting in Zach Charbonnet rushing 33 times for exactly 100 yards and 3 TDs. On the one hand, I guess it's nice to see the freshman RB Michigan really needed to hit coming through this early; on the other hand, throwing any human being at a defense for 30+ times not named MSU with the apparent instruction of "please save us" is uninspiring work in 2019. But far too often the offense looked out of sorts to a degree I didn't quite expect; not because Army doesn't have a competent defense that can give a team fits, but because they seemed unwilling or unable to take what the defense was giving them. Patterson alternated pinpoint throws into windows to assuming Nick Eubanks for 14-feet tall, and so it was hard to tell how much of that was due to lingering injury issues and how much was it just poor play or miscommunication.

But three first-half turnovers were the big story of the game, as again Michigan couldn't hold onto the ball and allowed the opponent to score off their miscues. What was particularly back-breaking about each of these turnovers was they stopped what were otherwise pretty promising Michigan drives. The first two Patterson fumbles were inside Army's 30-yard line; a score on either of those drives likely changes the complexion of that first half somewhat significantly. And the third fumble put Army in great position to score their only other TD in regulation. Ultimately the turnover margin evened out, with Michigan getting a red-zone pick to halt an Army scoring drive and the final, game-ending fumble in double OT, but Army down 7 or 10 is a far different beast than one up 7 at halftime, and I absolutely believe that the sludgefart ball we saw in the second half was due in large part to the coaching staff being extremely risk adverse. And frankly, if you have that much faith in your freshman RB to hold onto the ball why the hell not.

Adding to Michigan's misery were things like unforced penalties (9 for 58 yards, including a couple of false starts and Khaleke Hudson inexplicably lining up offsides in the first OT to negate what was a good 3rd-down stop), dropped passes, and a couple of questionable calls, the biggest being the negation of Josh Metellus's fumble recovery TD near the end of the 1st quarter. And while I don't have proof of this, my guess Army's pass defense was just grabbing guys because whenever Patterson even threw in the vicinity of a guy downfield he was usually being held or dragged down.

The saving grace for this team was, somewhat surprisingly, the defense, particularly against the run. If I had told you before the game Army would be held to well under 4 yards per carry and under 250 yards of total offense on 66 plays, that feels like a recipe for a comfortable win. Unlike Army's game against Oklahoma last year where the Sooners couldn't get off the field (Army ran 87 plays and picked up 379 yards), Michigan limited Army's long drives (they only had had two 60-yard drives but on the other 9 drives UM limited Army to 113 yards) and rarely gave up the big play. The 3.3 yards per rush by Army was their lowest output since 2015's 2-10 campaign, and the overall rush yardage total of 200 yards was the least since last season's loss to Duke. For a team with some questions up front and not a ton of experience defending this style of offense, it was nice to see Michigan so thoroughly handle the challenge.

So it feels like a cop-out to say "a win's a win", but Army is a fringe top-25 team and Michigan still found a way to win despite numerous issues, some self-inflicted and others less so. Michigan's offense looked absolutely lost at times, but I also don't think Gattis's offense calls for fumbling the ball in ever-more inventive ways, and so a Michigan team that lost 3 fumbles total last year is unlikely to keep up their 5 in 2 games pace, unless Michigan is planning on challenging a national record, which would admittedly be a very Michigan thing to do and still finish with, like 9 wins. And they were down both DPJ and Runyan, and especially with offensive linemen that continuity and experience can't be discounted especially when you're installing a new offense. So maybe I'm just a perpetual optimist, but analogies to Akron in 2013 or the general "welcome to the suck" feel a little excessive.

Worst: Scoreboard Gazing

What drove me crazy this weekend, maybe more than even this game, was how often people would look at, say, Maryland destroying Syracuse or MSU suddenly discovering an offense against Western and perilously over-read into early-season results. Yes, Ohio State beat up on Cincinnati and, because the world isn't fair, Fields will likely be yet another superstar for them. Wisconsin looks to be back to their usual ways, smashing USF and CMU by an aggregate score of 110-0. A week after MSU could barely crack 300 yards against a probably-bad Tulsa team (while giving up only 80 total yards), MSU scored 51 points with 582 yards of total offense, including 251 yards on the ground. My guess is MSU's offense is somewhere in between those two marks, and they'll get a big test next week when Arizona State shows up. Penn State struggled with Buffalo for a half and then turned a couple of breaks into a blowout, and more generally have outscored their two opponents 124-20. As a conference, there are 6 teams averaging 40+ points per game thus far, which ties the Big 12 for the lead in that completely arbitrary category. Do I think that holds true for the rest of the year? Probably not, but a lot of teams are eating their Wheaties thus far into the season.

It's not that I think Michigan has played like the #7 team in the country these past two weeks; the offense has moved in fits and starts, the turnovers have been exceedingly bad especially at the QB spot, and while the defense has played well it's hard to tell if they'll be able to hold up against some of the more dynamic offenses on the schedule. But what seems to be happening is that context is being lost and expectations needlessly rachetted up. Talk of Michigan as a potential playoff contender always felt premature given the uncertainty surrounding how easily the offense would be integrated and the number of departures on defense. I've lamented before that even a very good Michigan offense paired with the past couple of elite defenses probably gets you to that spot, but this year's team just isn't there yet and probably won't ever be. It's not really an excuse as much as recognition of Michigan's ceiling right now; if you want to make that an indictment of the coaches, the program, college football generally that's your prerogative.

But I think it's a fool's errand to judge a team's success by the performance of other teams that have no real connection to them beyond their placement on a future schedule. MSU beating the crap out of WMU looks really impressive, but Michigan also beat the tar out of a similar WMU outfit last year (49-3), and Michigan's offense certainly didn't take off for the rest of the season. Wisconsin is probably a top-10 team right now, but CMU hasn't beaten an FBS program since 2017 and USF hasn't won a game period since the fall of 2018. My point isn't to discredit other teams or their accomplishments, only to point out that staring at a couple of numbers and applying New Math to them doesn't mean it gives you an accurate reality of any team, even moreso if they're used as a barometer for measuring someone else.

Michigan has issues to address, ones that we all sorta knew about as well as some that might come as a surprise. But those problems exist independent of what other teams are doing right now, and especially with a bye week coming up I'm not looking forward to a ton of "OMG, MSU just beat a Herm Edwards-led ASU team, we're screwed!" takes.

Best: The Suite Life of Zach

So it took about a game for Michigan to realize that Zach Charbonnet should be the feature back for this offense, and it's probably a testament to Christian Turner and Tru Wilson that it even took them that long to come to that determination. You look at the raw numbers (33 carries, 100 yards, 3 TDs) and you wouldn't be blamed to wonder what all the fuss was about. It's a pretty inefficient game in terms of average yards, and while he was again solid in pass protection a long run of 12 yards points to a grinding afternoon. And yet, on those 33 carries he somehow only lost 4 yards (and that was on a terribly botched 4th-and-2 attempt to start the 4th quarter) and consistently found a way to slip through tackles and pull a couple of yards out of plays where everyone in the stadium knew he was going to smash into the middle of the line. But after 3 fumbles in the first half it was clear that the gameplan shifted to "hold onto the damn ball and don't turn it over to Army", and it's telling that the coaches thought this true freshman RB was the best option to accomplish that.

I'll be honest - I don't remember the beginning of Mike Hart's career all that well. I remember him competing with guys like David Underwood and Jerome Jackson, and there being some concern to start the year that Michigan really couldn't run the ball effectively, which was jarring after Chris Perry won the Doak Walker award the year before. But after losing to Notre Dame and looking anemic on the ground, Hart became the main back against SDSU and he never relinquished it. Along the way he put up some eye-popping numbers, like a Tom Emanski 200 (200+ yards in back-to-back-to-back games), and he suddenly gave direction to an offense that was also dealing with uncertainty at QB with Henne a freshman starter. He had a calmness, a certainty in how he played that was hard to quantify but clearly put the rest of the team at ease when he had the ball. I know that's a lot of squishy feelingsball stuff, but he felt like a senior starter by game three.

It's still early, and expecting any true freshman RB to put up Hart's first year is sorta crazy, but Charbonnet has a bit of that feeling around him. Michigan clearly became more reliant on running the ball as the game progressed; whether that was because Patterson wasn't healthy enough to run the read option, the staff explicitly decided against it, or some combination of factors I don't know. But they felt putting the ball in his hands was the more prudent option at the time, and that is telling for a team with a lot of talent on it. Hopefully his usage in this offense can evolve to more than what we saw against Army, but thus far he looks absolutely like the top-end RB talent he was touted as coming to UM.

Worst: You Can't Impersonate an Offense

One of my favorite series on Youtube is Wired's "let's talk an expert about something in pop culture" collection of videos, in particular when they talk to a dialect coach named Erik Singer. I've always been intrigued by accents and dialects, and watching him break down how various actors try to capture a particular language or person's manner of speaking is always illuminating.

One of the consistent points he makes throughout the series is that the best performances come when the actor's manner of speaking is natural and integrated, a part of a larger work that becomes, ultimately, seamless. The actor transforms from a person trying to play the role of someone with, say, a French accent to a French-speaking version of that actor simply playing a role. It becomes organic and instinctive, and it's believable because it's part of that actor's DNA. When the accent doesn't become part of the character, it's instead an impediment the actor is fighting against, a wayward creature she's trying to tame.

At times, it feels like Michigan's offense continues to fight the adoption of the system Gattis wants to install. I'm not able to sit here and say that Josh Gattis's offensive philosophy will ultimately be successful at Michigan as presently constituted; I've been wrong numerous times about different offensive systems trying to implant themselves. But at some point you have to give yourself completely over to the process and see what happens, and I'm not sure that's happened at Michigan both this year as well as last year's various wrinkles/flirtations with RPOs and tempo. Rich Rodriguez failed at Michigan for a ton of reasons but one wasn't that he pussy-footed around with running his preferred offensive system; he rolled out a Dual Threet system with two guys who absolutely couldn't run it and took his lumps in the process. To a much lesser extent, Don Brown is running his system at Michigan even if the personnel might not be a great match, and he's make adjustments on the fly to compensate. But he's still aggressive and trusts his corners to play man coverage far more often than a lot of his contemporaries. There has to be buy-in for anything to succeed to its fullest potential.

And yet, in a game where Army was practically begging Michigan to pull the ball when they crashed off the end, Patterson rarely if ever did. Now, I don't know if that's intentional or not; I remember last year's NW game sending people into a tizzy about the running game and then in two weeks he was carving up Wisconsin on the ground. My sense is that there's an intentionality in not running with the ball, either on Patterson's part or (more likely) the coaches'. Again, if Patterson is injured and the staff is concerned about turnovers after 4 fumbles in 2 weeks from their QB, I sorta get that reticence. If they are "saving" plays for bigger opponents, that's insane (and also unlikely). I honestly don't know the reason why, and it's doubly confusing because if it was just an injury issue then you'd assume McCaffrey would have been brought in more frequently even if there are real questions about his passing ability.

Regardless, it still feels like Michigan is acting like they're running this offense instead of actually doing so, and as a result it feels disjointed and uneven. Part of that is undoubtedly because this is Michigan's third-ish offensive coordinator in 3 years, but that's kind of the rub. Michigan's offense wasn't great by any means last year but it certainly wasn't terrible; it finished in the top 30 per S&P+ and that was weighed down a bit by the stinker against Florida. I think the change this offseason was necessary, but it has to be a full-throated commitment.

Worst: Analytical Announcements

I know some people liked Joel Klatt's analysis during and after this game, but I found it incredibly simplistic. It just felt like what happens when someone in the football media learns a new term and then becomes contractually mandated to use it all the time, regardless of context or applicability. Chris Collinsworth calling literally everything an RPO comes to mind, and outside of football Vince McMahon screaming "what a maneuver" is probably the high watermark for redundant phrasing. In this game, Klatt kept saying Army had an "analytical" approach for 4th-down attempts, and as far as I can tell the "analytical" component was "3 or fewer yards to go". That's not really analytics as much as a boolean statement, I guess with the small caveat that they did consider if they were in the shadow of their own endzone. Army is designed to take advantage of short-yardage 4th-down options, and I personally love that they take those risks because it takes advantage of an inefficiency in football and fights against conventions, which inevitably messes up other teams. But to keep treating it as some amazing insight feels a bit specious. It was also undercut by the couple of times he openly wondered why some P5 team didn't run the option, which both ignored the existence of Georgia Tech under Paul Johnson as well as the rather obvious truth that a lot of the "good" parts of the triple option are already found in other offenses while minimizing it's limitations (limited ability to catch up when down, few passing plays that naturally arise from it).

The other key issue I had, and this relates to the earlier discussion about accepting a change in offensive philosophy, was his consistent complaints about Michigan's short-yardage offense. He made a very coherent point that starting Zach Charbonnet 3-4 yards behind the line of scrimmage made it difficult for Michigan to convert. But his solution, namely that Michigan should keep some goal line formations in their offense, doesn't make any sense. College teams have limited time and resources to prepare for games, and so if you are going to install an offense for 95 yards of the field and a separate one for the last 5, you are both wasting valuable time and courting confusion and impracticality for, at best, minimal upside. What I mean is, the offensive system Michigan is trying to run under Gattis has ways to convert on short yardage out of the shotgun; it isn't hard (don't click the link) to look (seriously, don't click it) in recent (do you like pain?) Michigan history (why would you click it?) to see a version of it in action. There were a number of times on short yardage where a pulling Patterson would have waltzed past the line and/or into the endzone had he held onto the ball. You just have to run the play correctly. Similarly, having a "goal line" offense that is demonstrably different than your base just introduces new blocking schemes, new QB-Center handoffs, and a whole mess of other complications for a set of playcalls that tell the defense "we are definitely running right now", which removes even the hint of surprise the formation would normally give you if it was part of your base offense. Yes, maybe in year 2 or 3 of this offense you can tinker with new formations or wrinkles (though still probably not to that degree), but at this point Michigan's struggles converting aren't on some formation weakness nearly as much as it is on the playcalls and their execution.

Best: The Defense, again

This honestly happens all the time with these columns; I write 15 paragraphs about some Very Important Point relating to the offense and then I get to a discussion of the defensive performance and it's very "they did a really good job and deserve a ton of credit." But that's where we are with Don Brown at this point; he's just a pro and got a still-evolving Michigan defense to keep Army under 250 yards of total offense despite 2 OTs. There were a couple of breakdowns (as one would expect) and they did get got by the one play-action pass everyone knew was coming, but otherwise Army struggled to sustain drives of any length. As noted earlier Army had 4 drives of 10+ plays, but the majority of their drives were 6 plays or less. All of their scoring drives came off turnovers, and while Michigan got lucky they threw a pick in the redzone, Michigan also had a clear fumble-and-TD score negated by refs so I consider that sort of evened out. It was still an incredibly impressive performance by a unit that I still don't have a great read on.

In terms of individual performers, I thought Hudson, Hutchinson, Uche, and Paye really played well and consistently limited Army to minimal gains inside. Army is not the type of team to take a ton of TFLs, but their system also requires them to consistently gash an opponent for 3 or 4 yards; a run of 1 or 2 yards isn't a particularly effective run on early downs. By that metric, Michigan did great; about half of Army's first-down and second-down runs were for 2 or fewer yards. For the game Army only had 15 first downs and 12 by rushing, and for a team that typically converts 3rd downs at a better-than-50% clip Michigan holding them to 5/16 is really damn impressive. Hell, including their 3/3 on 4th down they were still below 50% for the game. And Hudson and Glasgow in particular limited the damage Army could inflict when they tried to run outside the tackles; they tamped down some dangerous pitches and outside runs that could have been disastrous.

The defense did make a couple of mistakes, including Hudson's offsides in overtime and Hill being called for holding on 4th down that kept Army's second scoring drive alive. But in a game where the offense kept sputtering along and putting the pressure on the defense to keep Army out of the endzone, I thought they did so admirably. And while it doesn't necessarily mean much, MTSU just put up 600 yards (at 10 yards per play) a week after barely cracking 4 yards per play against Michigan, and most of that coming on their last drive. None of this may be applicable in two weeks againt Wisconsin, but this definitely feels like a defense that is quickly figuring itself out.

Quick Hits

• Michigan's unwillingness to throw downfield surprised me in the moment but upon rewatch it felt likey they did throw the ball when necessary but, probably due to the structure of the game, were still pretty conservative in the end. Also, it felt like 95% of the Air Force pass defense was just holding, as a ton of times Michigan threw deep it was either completed or a PI was called.
• The fake punt definitely felt like an audible on the field and was, honestly, sort of lucky it worked.  Dax Hill had to make the Army guy miss to convert, and I can see that going another way pretty easily against better competition.  I like guys being able to read and react on the field, but that felt more dumb luck than I like.
• The continued absence of John Runyan was felt in this game. With Michigan's QBs not running much, you could tell they wanted to just run plays off one side and couldn't quite figure out one that worked consistently. Last year, Michigan ran behind Runyan when they needed yards and if he's able to return against Wisconsin that will be a major boon for the offense.
• Ronnie Bell had an up-and-down game. He dropped at least one ball and nearly fumbled away a punt return, but he also led the team in catches and converted a number of third-down plays that helped Michigan keep drives going in that second half and OT. My guess is he'll fade a bit once DPJ returns.
• Patterson did struggle at times throwing the ball but I also think people are over-reacting a bit because of that last drive in OT where he went 0-3 and threw two bad balls. His second-down throw to Black should have been caught, and he suffered a couple of other drops in this game to go along with some bad throws. Still, of Michigan's 11 first-downs via pass 5 came on 3rd down throws by Patterson. Even if he was less than 100% to run the ball I can see why the coaches would still want him throwing the ball.

In Two Weeks: Wisconsin

This game has gone from "I think Michigan pulls it off" to "this could get ugly". I am not buying Wisconsin as fully back; Taylor looked good and Jack Coan has already matched his season total of 5 TDs from last year, but as noted earlier they've played two of the worst teams in college football that makes it hard to read too much into the performances. We won't get any new information about them before they play Michigan in two weeks, but I'm fairly confident in saying Michigan can't have a repeat of this game's offensive struggles and hope to keep it close. Wisconsin will be without one of their starting safeties, and Michigan has the receivers to really tax Wisconsin's secondary in a way they haven't all year. The Badgers were a thoroughly mediocre pass defense last season and who knows how they'll handle a team that can throw the ball, though we've yet to see Michigan really take those risks. Gun to my head this feels like a close loss, but also the type of game where a break or two going Michigan's way and they could leave Madison with a huge road win.