Hi all. First and foremost, I hope everyone is staying safe and doing their best to get through these difficult times.

Since I’ve had a lot more time on my hands these days (and have been running out of shows to watch), I began compiling some data together of the Michigan-Ohio State rivalry. Given the one-sided nature of the rivalry, the notion that Ohio State is operating at an historically great level, and the criticism that Jim Harbaugh has received, I wanted to take a holistic approach in examining Harbaugh’s tenure relative to the history of the series. Is it true that Ohio State is indeed performing at the peak of the program’s history? Is the criticism of Harbaugh warranted (or at least blown out of proportion)? What other insights can we gather?

To do this, I wanted to go beyond arbitrary measures of success, W/L records, head-to-head results, etc. Rather, I thought it would be more insightful to visualize how good Michigan and Ohio State teams were over time using an advanced rating system such as Bill Connelly’s SP+ or ESPN’s FPI. While I found the aforementioned ratings as well as others not to be easily accessible for use, I found sports-reference.com’s Simple Rating System (SRS) data to be very well consolidated and downloadable. The definition of SRS: “A rating that takes into account average point differential and strength of schedule. The rating is denominated in points above/below average, where zero is average, although it should be noted that margin of victory has been assigned a lower bound of 7 points and an upper bound of 24 points.” For example, Harbaugh’s 2019 squad with an SRS rating of 15.5 is 15.5 points better than the average NCAA team.

To start, I plotted both team’s SRS ratings over time beginning in 1953 when the Western Conference officially became the Big Ten Conference, while also including 5-year moving averages. In addition, I layered over box charts for each of Michigan’s head coaches since Bo Schembechler to help visualize how strong the competition was during a coach’s respective tenure. (As a side note, I’m using a liberal definition for “box chart”, and have provided keys for each of the visualizations. If you have a problem with it, well, you can get off my lawn. The true application of a box chart can be found here.)

Since inheriting a Michigan program that within just a seven-year span fielded its five worst teams since 1967, Harbaugh has yielded a five-year stretch with teams that are roughly in line with Michigan’s long-term historical average. At the conclusion of the 2019 season, Harbaugh sustained a five-year moving average SRS rating of 15.9, which is the highest the program has achieved since 1997 (16.6 rating). In other words, the program is in its best spot in over two decades. Since the beginning of the Lloyd Carr Era, Harbaugh has fielded Michigan’s 2nd (2016 team; 17.6 SRS rating), 5th (2018; 16.5), 6th (2015; 16.3), and 10th (2019; 15.5) best teams over a total of 25 seasons. Since the beginning of the Gary Moeller Era, those respective teams stack up as the 5th, 10th, 11th, and 15th best over a total of 30 seasons.

Not bad, but how does this compare to Ohio State? This is where the box charts come into use.

At the conclusion of the 2019 season, Ohio State boasted a five-year moving average SRS rating of 21.3 (best since 1977 when the five-year average was 23.2), with the 2019 edition of Ohio State (27.4) being the best team in the program’s history since 1973 (29.7). Only Bo Schembechler has faced off against a better Ohio State team. Furthermore, the 17.7 lower bound of the Harbaugh Era (i.e. the worst Ohio State team he’s played) as well as the 21.3 middle bound (i.e. average Ohio State SRS rating he’s played) are significantly higher than that of all his predecessors.

Visualized a little differently, I plotted out the SRS differential of the two teams (+/-) where the blue markers indicate a season in which Michigan fielded a better team than Ohio State, while the scarlet markers indicate the opposite. I also layered in Michigan’s SRS ratings over time as well as additional box charts for each coach’s tenure.

Not including the Rich Rod Era, Jim Harbaugh has dealt with the largest SRS +/- deficit in the history of the series. What’s interesting is that despite Harbaugh assembling significantly better teams on average than Hoke, the gap between Michigan and Ohio State hasn’t closed at all (though Hoke’s numbers are buoyed by Luke Fickell’s 2011 edition of Ohio State). Harbaugh joins Rich Rod as the only two coaches to not have a single team rated higher than Ohio State during their respective tenures. Furthermore, Ohio State’s stretch of dominance in fielding better teams for eight straight years between 2012 and 2019 is only superseded by Michigan’s 1985-1994 string when it fielded 10 straight better teams (7-2-1 record during that span).

Taken a step further, I filtered the Ohio State teams into three categories of elite (SRS rating greater than 21), good/very good (SRS rating between 21 and 14), and below average (at least by Ohio State standards, with an SRS rating below 14). I understand this is a very arbitrary categorization, but teams with >21 ratings are relatively scarce and the majority of teams hover in the 14-21 range. It’s also helpful to use round football numbers with a touchdown multiple of 7.

As a percentage of Ohio State teams that each coach has faced, Harbaugh has gone up against the highest share of elite teams of any coach. Furthermore, 100% of the Ohio State teams Harbaugh has faced have fallen into the elite or good/very good categories, with Bo Schembechler having to compete with the next highest share of elite to good/very good teams at 71.4%.

So this is all to say that no matter how you slice and dice the numbers, Harbaugh has had to go up against the strongest Ohio State regime of any coach and it’s not particularly close. To further justify this you can go to 247’s team recruiting rankings, or you can stare directly into the sun if you wish. So maybe we can all take into context Harbaugh’s 0-5 record versus Ohio State and come to grips that unfortunately the man many coined as the savior of Michigan Football hasn’t been able to conquer the Death Star that is a blueblood program at its historical peak. Well, yes and no…let’s look at more data.

Since Harbaugh’s hire, if you’ve only watched the Pete Finebaum Show, or have only read the MGoBoard for that matter, you’d be lead to believe he’s failed miserably. Haters will say it’s fake, but in reality Michigan has been a Top 10 program during Harbaugh’s tenure, ranking 6th in five-year average SRS rating. (As a side note, comparable advanced rating systems yield similar results. For instance, Michigan ranks 7th over the same span per Bill Connelly’s SP+.) Ohio State is 3rd. Big Ten rivals Wisconsin and Penn State are just behind at 7th and 8th, and Notre Dame checks in at 10th. What really stands out in the chart is how far above Alabama, Clemson, and Ohio State are above the rest of the pack. The difference between Ohio State at 3rd and Oklahoma at 4th is 4.2 (!), while the difference between Oklahoma at 4th and Notre Dame at 10th is only 2.6 (!).

Next, I plotted each team’s SRS rating over time.

Michigan has been remarkably consistent and among the nation’s best during Harbaugh’s coaching campaign. This chart is a little cluttered, so it’s not immediately obvious…but every single program has had at least one team with a higher rating than even Harbaugh’s best team (2016; 17.6 rating); including Ohio State (5), Wisconsin (2), Penn State (1), and Notre Dame (1). Furthermore, if we keep the same definition of elite as before (i.e. SRS rating of >21), every program has add least one elite team except for Oklahoma, Michigan, and Notre Dame. Harbaugh has yet to break the glass ceiling.

The programs of Alabama, Clemson, and Ohio State are in a league of their own with a five-year average SRS rating of elite status. And to reiterate, the difference between Ohio State at 3rd and Oklahoma at 4th is 4.2, while the difference between Oklahoma at 4th and Notre Dame at 10th is only 2.6.

So I find this both reassuring and depressing. On one side, Jim Harbaugh has Michigan among the best across the college football landscape. But on the other side, Michigan is a clear tier below the nation’s elite – a group that includes archrival Ohio State – and Harbaugh has yet to field his own exceptional team. I should be content, or even happy, that the program is in such a great position. Yet I’m not (nor is the rest of the fan base) because of the all too familiar anguish that is the last Saturday of November.

Let’s plot that anguish.

To do so, I want to compare how Michigan was expected to perform in The Game versus how the team actually performed. Given SRS is scored as how many points above/below average a team is, I can project an expected point differential for a game by subtracting Ohio State’s SRS rating from Michigan’s SRS rating and making an adjustment for home field advantage. I can then measure how far a team deviated from expectations by subtracting the expected point differential from the actual point differential.

E[Point Differential] = (Michigan SRS) – (Ohio State SRS) +/- (3 point home field advantage)

Error of E[Point Differential] = (Actual Point Differential) – (E[Point Differential])

It’s not a perfect science, but if I do a regression analysis there’s a pretty strong correlation between expected and actual point differentials. A home field advantage of 3 points resulted in the tightest correlation. Also note that the lines here are being calculated retroactively, and by that I mean that the actual results of The Game and any postseason games would be factored into each team’s SRS rating (and thus the expected point differential calculation). I ain’t a betting man, so I don’t really know off the top of my head all of the lines that were set for the recent Michigan-Ohio State games. But as an example, the E[Point Differential] calculations of Ohio State -8.9 in 2019 and Ohio State -4.3 in 2016, I believe, are roughly in line with where bets were taking place at the time.

The following chart depicts the results from calculating the Error of E[Point Differential], with the blue markers indicating games in which Michigan exceeded expectations and with the scarlet markers indicating games when Michigan underachieved. I layered over additional box charts for each coach’s respective tenure.

Jim Harbaugh immediately stands out. The three worst beat downs, relatively speaking, since the beginning of the Bo Schembechler Era have come in the last five years with Jim Harbaugh at the helm. Three separate times Harbaugh has underperformed to the tune of roughly three touchdowns. And while almost every predecessor got handed a relative beating of two touchdowns, Harbaugh has averaged an underperformance of two touchdowns.

The only time Harbaugh has overachieved was the infamous 2016 game.

The only time the rivalry has seen beat downs of similar magnitude were 1976 and 1993 when Michigan was on the right side of history.

Harbaugh’s resume when it comes to The Game is historically terrible. So do you believe in regression to the mean (or in other words, that the Error of E[Point Differential] should result in amounts closer to 0 in the future)? Or is there something systematic in Harbaugh’s approach that causes such significant underperformance in The Game?

To summarize:

• The data makes a strong case that Ohio State is indeed at the peak of the program’s history. In 2019, Ohio State had its best team since 1973 and sustained its best five-year moving average SRS rating since 1977. In the last five years, Ohio State as well as Alabama and Clemson have been in a tier above the rest of college football.
• Jim Harbaugh seems to be in the right place at the wrong time. Michigan is in the best spot it has been in a long time, but Ohio State has operated at a level that no Michigan coach has experienced before.
• There’s not much doubt that Michigan under Harbaugh has been one of the nation’s best programs. But his lack of success (to put it lightly) in The Game has distorted perception. He’s also yet to have broken the glass ceiling and have a truly great team, and you can chalk that up to not having an elite QB or DTs, Ohio State simply having our number, the football gods hating us, or whatever you want.

Looking towards the future:

• Simply put, Michigan has not yet been able to escape purgatory. Just on the cusp of glory only to be subdued by our archrival. As fans we’re only left to question. What if past events transpired differently? Does unearthing the foundation that Harbaugh has set get us to the promised land, or does it just kick the can farther down the Road to Success?
• There are reasons to have faith that Michigan can turn the corner, reasons that are both within the program’s control and out of the program’s control (to an extent).
  • Identity. I was rereading umgoblue11’s diary post Myths and Half-Truths: An In-Depth Look at Michigan’s Recruiting Process from someone in the Industry, and one of the primary themes was Michigan’s lack of identity. When Harbaugh first got to Michigan, he built the team in his image, with a fullback and tight end. It was Enthusiasm Unknown to Mankind. It was [Try Hard and You Will Stay and] Those Who Stay Will Be Champions. It was Jordan brand. It was show business with Signing with the Stars. It was spotlight with celebrities like Michael Jordan and Tom Brady being honorary captains. It was calling out the jive turkeys on Twitter. It was in your face and in your head all the time and take no prisoners. What is it now? Or what should/will it be? Hopefully Michigan can find something.
  • QB Room and Modernization of the Offense. Shea Patterson was expected to be great, but he was only good. Still, Michigan’s QB room is as healthy as it has been in a long time. QB depth was the reason Brian thought 2019 could have been The Year. At Big Ten Media Days last year, Harbaugh went as far as saying about Shea and Dylan, “…I see them both playing. Where it stands now, I see maybe redefining what a starter is.” We can all speculate how much golf was actually played that summer or what would have happened if Dylan had not suffered a terrible concussion in the Wisconsin game, but alas. The jury is still out when it comes to Dylan McCaffery, Joe Milton, Cade McNamara, and (in the distant future) JJ McCarthy. The probability that a QB from that group turns out to be very good to elite is much higher than Harbaugh’s past years. And that comes in turn with the modernization of the offense. It probably shouldn’t have taken Year 5 of the Harbaugh Era to switch to a modern offense given the current landscape of college football, but the important thing is that it’s in process and Gattis’ first season showed a lot of positives after a rough start. Can’t wait to see what Year 2 of Speed in Space brings.
  • Gaining Competitive Advantage. People won’t forget how Harbaugh pushed the envelope in gaining a competitive advantage in any way that he could. Satellite camps, spring practice abroad, hires such as Chris Partridge, etc. Harbaugh has to find other ways to gain that competitive advantage, and maybe that is being the catalyst for progression in student athlete welfare and NIL rights.
  • Playoff Expansion and Conference Realignment. These could go a long way in increasing Michigan’s margin for error when attempting to make the playoffs, and being represented in the postseason as one of college football’s best X teams with a chance to win it all could only help the program. My money is on things changing at some point.

If you read this entire diary post, thank you. I hope this helped fill the sports void in your life at least a little bit.

Go Blue!