Sources: Sports-Reference, ESPN, digital.sandiego.edu, Pro-Football-Reference

I know, I know. In football, yards are sexy, but at the end of the day, points are the only thing that matters. So let me assure you that Part II of this analysis will focus on where the Harbaughffense gets its points. This first part isn’t meant to do much more than answer “how does Harbaugh move the ball?”

In the time that Jim Harbaugh has been at Michigan, if I could characterize how I’ve perceived his approach to offense, the single descriptor would be “situational.” Some teams, like Ohio State under Ryan Day, seem to approach every single down as if it’s 1st and 10, no matter the score, no matter the situation; they’ve had a ton of success doing so, but sometimes that approach works against you. Witness their failed 4th down conversion attempt in The Game 2022, where they attempted a medium-long pass when any kind of a running play probably would have succeeded.

Harbaugh, on the other hand, seems to assign value to any given down & distance, and moves the ball accordingly. 3rd and 1 being much higher value than 1st and 10, etc. He also doesn’t hesitate to almost completely shut down the offense if he deems that the game is functionally over: en route to a 78-0 win over Rutgers, there were 3 passes attempted in the 2nd half. That might be a bad example, because that game featured a grand total of 8 completions, was never competitive, and we tied a rushing record. We also know that he views QB “success” vs. “failure” on a drive-by-drive basis; one of his favorite statistics is reportedly Points per Drive.

So…uh…we won’t be controlling for any of that here. I don’t think that Harbaugh values yards (as their own end goal) pretty much at all; the only exception I can think of is when he made sure that Rudock got a 3,000-yard season.

Who are the horses?

Let’s see who Harbaugh was working with any given year.

Bold = played in the NFL (or recently drafted and pending playing time)…but it doesn’t discriminate between “had a cup of coffee” vs. “superstar”. Some NFLers are more equal than other NFLers. Everything from 2011-2014 is bold…obviously.

Ok, so what jumps out at me first is that Jim Harbaugh, even though he is known for his RB and TE utilization, has had some TALENT at the WR position.

The second thing I see is that Harbaugh wasn’t originally a FB-driven coach, but I guess Owen Marecic will do that to a guy. Harbaugh wasn’t as keen on FB involvement during his NFL time, but immediately went back to it when he came to Michigan; it’s tapered off a bit since 2020, but the conversion of Kalel Mullings from LB to de-facto-FB-even-though-listed-as-RB makes it pretty clear that Harbaugh always wants a mooseback.

To be completely fair, Hassan Haskins was his own mooseback, and I think Harbaugh was fine with Blake Corum being his own mooseback until the injury.

Yardage

First, the easy one: Total Yards.

Last two years we’ve been moving that football. Let’s have some context. Yards per Reception, Yards per Rush, and Yards per Play*

* Gotta define “play” here. Because this analysis isn’t looking at passing attempts, only completed passes, “Yards per Play” really means something closer to “weighted average of rushing yards per rushing play and receiving yards per completed pass”. Sorry.

Ok, this is why Harbaugh doesn’t prioritize yards as an end in themselves. Look at 2020 and tell me if you think our offense was working correctly. Look at 2018 and 2019 and tell me if it looks like we were finishing drives. Look at 2017 yards per reception and tell me that John O’Korn was every bit as effective as Jake Rudock in getting our WRs the ball.

Also: holy s*** Andrew Luck!

Let’s talk rushing yards:

Ok, the number one question I have, even before we take a look at a macro level view of where rushing yards come from, is “I NEED TO KNOW ABOUT QB RUSHING YARDS!”

I’m writing this before looking at the stats, but dudes and dudettes I’m pretty sure I had the impression that Harbaugh likes a QB that can run. But since coming to Michigan…what’s the phrase that Brian and Seth use?...”slider set all the way to ‘give’?” It feels like rushing yards from the QB position have been pretty minimal. Let’s take a look.

This is skewed by sacks – dammit NCAA stats – but duuuuuuude. We haven’t just been imagining things: QB rushing yards at Michigan have been off what we might consider “normal Harbaugh trend.” He even had Alex Smith netting about 150 rushing yards per year in San Francisco before Kaep took the starting job.

I know that this take is more feelingsball than anything, but the charts we’ve looked at so far only reinforce my belief that 2018 could have been a special year if it hadn’t been for Pep “we will only take a deep shot once per quarter” Hamilton. I’ve got a note to see how this bears out when we look at scoring in Harbaughffense Part II: Field Goal Boogaloo.

One question answered. Let’s see how this is incorporated into total rushing yards every year.

I love statistics. “What percentage of rushing yards came from the RBs in 2007?”, you ask. “More than 100%”, I answer.

The thing that stands out to me is that Harbaugh’s use of a bona fide FB as a ball-carrying mooseback is a Michigan thing. Owen Maricic wasn’t getting a ton of yards. Bruce Miller didn’t get a ton of yards, and was listed as a RB even when he was. But in 2015, my god Houma and Kerridge were chain-moving parts of the offense. In 2016 we can see that Khalid Hill didn’t get a lot of rushing yards, but I bet that Part II of this analysis (i.e., where do points come from?) will show that he may have set a record for the number of 1-yard touchdowns.

I’ll also admit that I was expecting that a lower ratio of RB yards to total rushing yards would be indicative of a healthier offense. But once again, yards by themselves are not telling the whole story here: the 2016 and 2020 Michigan offenses and the 2009 Stanford offense aren’t comparable. One thing that we CAN conclude is that needing your RB to make up for net negative rushing yards from the QB position…is bad. Get your hot takes here.

Receiving

Once again I’m writing this first paragraph before I look at the stats, but I’m expecting a solid mixture here between WR yards and TE yards. We’ve had big-play TEs since the day Harbaugh arrived at Michigan, and we’ve all lamented the lack of a “jump ball to Braylon/Hemingway” kind of attitude.

I’m also expecting a somewhat infuriating lack of RB receiving yards. You know how sometimes a team will have a guy who’s listed at RB, but he’s really a WR who lines up in the backfield? Christian McCaffrey is probably the best current example. Marshall Faulk was amazing at this (and everything else). LaDamian Tomlinson as well. Alvin Kamara. The Patriots won a Super Bowl with James White. Urban Meyer had an infuriating run of Spread H-backs. I want one. I want many. I want justice for Chris Evans and his un-defendable wheel routes.

Actually, let’s look at that first. RB receiving yards and the percentage of RB receiving yards compared to total receiving yards.

So I was a bit off base here! Turns out that the Harbaughffense gets receiving yards from the RBs at a pretty steady clip…uh…except one time Honestly, though, this chart probably says more about the QB than the RBs…anyone remember 2019 when Michigan had 3 NFL WRs to feed, but also Shea Patterson would completely ignore the side of the field where the RB had leaked out the backfield and was blitheringly wide open, and stare down his covered #1 option until he took a sack?

It also paints a picture of Cade McNamara as a check-down wizard who took free yards whenever they were available…wait a sec, let’s check the stats on this:

Dude Chris Evans in 2017 was lethal. No wonder MGoBlog fixated on him getting more touches.

I’m honestly not sure what to take from the last couple charts. This might be making the case for RBs to become a central part of the receiving corps; if you have a QB who can deliver the short ball at a near-100% clip…like…say… Tom Brady, there’s nothing wrong with 8-10 “free” yards. Then, when you have a back like Donovan Edwards, or Chris Evans, or Saquon Barkley, you lean on them.

Oh, and Saquon Barkley’s receiving stats at Penn State from 2015-2017: Yards per Reception 8.1, 14.4!!!!, and 11.7. Total receiving yards 161, 402, 632!! You want a Moneyball stat? My money is on “RB Receiving Yards” until further notice.

Let’s look at overall mix of receiving yards by position, and what % goes to the WRs.

Now THIS is interesting. Not only has the importance of the FB in the passing game kinda gone away after Khalid Hill left (kinda...keep in mind that from 2020-2021 Haskins was his own fullback), but we have wild swings in TE utilization. Not what I expected. Jake Butt in 2015 and Vernon Davis in 2013 really carried the load in terms of receiving. The 2018 version of Zach Gentry was also big in the passing game.

It’s also interesting to note that Michigan was trying to feed its trio of NFL receivers in 2019, the highest share of WR receiving yards that Harbaugh has ever had. Or maybe Josh Gattis just hates TEs.

What does it all mean???

I think this backs up what the eye test has told us for the past couple years: Harbaugh has finally succeeded in recreating (the methods of) his Stanford success. Looking at the players he’s been working with, it’s really no surprise that “NFL talent across the board” = successful offense, but you still have to deploy that talent effectively, and I think that’s where the last couple years have been the standouts.

If I was a gambling man, I’d bet $1 that Cade McNamara isn’t going to play in the NFL. Erick All has a chance if he’s truly truly healthy and if Iowa decides to play offense this year. All of which makes 2021 even more impressive, because there wasn’t NFL talent across the board, but we got the yardage results (and W-L results) as if there was. The 2022 team was stacked, and it shows. Rushing average (and total!) was the best that Harbaugh has had in his career, solidly beating the previous best: the year before. Last two years have been special.

But yards still don’t tell the full story. There are just too many variables to make it a useful stat. I don’t like offensive statistics which can be skewed downwards if your defense and special teams are good. “Good between the 20s” is another thing that we’ve gotten frustrated with over the years – our offense can rack up 80 yards and then pay it off with 3 (or zero!) points. Opponents playing bend-but-don’t-break defense against us can pad yardage stats, big plays distort averages, and, of course, Michigan has a fair share of blowouts every year where entire halves of football are more about ending the game sooner as opposed to actually attempting to move the ball or score.

And so, finally, I’d like to submit for your approval, “The Michigan Difference: doing an extraordinary amount of research and work, only to conclude that it isn’t terribly useful.”