By the Numbers: Comparing Harbaugh’s First 5 Seasons

Last January, Phil and I discussed a comparison of Jim Harbaugh’s 4-year coaching record to the last 4 years of Lloyd Carr’s coaching career (Harbaugh and Carr each had an advantage in 7 metrics, with 3 ties).  My general feeling was that Harbaugh had returned Michigan to the level at which Lloyd had left it. In my eyes, that made the 2019 season pivotal for the Harbaugh era.  Phil and I wondered whether the program would continue to rise, and I was certainly optimistic.

Now, in February of 2020, we have more information to process, and we can complete a 5-year analysis.  After Michigan completed a frustrating 9-4 season in 2019, and the 2003 Big Ten Championship season (10-3) gets added to the Carr side of the ledger, the comparison swings in Lloyd Carr’s favor.   Carr has an advantage on 7 metrics, Harbaugh in 4, with 2 ties. The advanced metrics were ignored for lack of 2003 & 2004 data, but Lloyd’s 3-year average would have been 3 more wins for him. This seemed like bad news for a Harbaugh optimist like myself.  So, I wanted to extend the comparison to see how the Harbaugh era, as it stands right now, compares to some other notable college football coaches.

METHOD

In addition to Lloyd Carr, I chose these coaches to compare to Harbaugh: Jim Tressel, Urban Meyer, Mark Dantonio, Nick Saban, Dabo Swinney, Brian Kelly, James Franklin, & Chris Petersen.  From the Carr comparison, my curiosity immediately jumped to Jim Tressel’s final five seasons at Ohio State.  After completing that comparison, I wanted a better apples-to-apples comparison so I wanted to use the first five seasons (of their current tenure) for the rest of the coaches I selected.  In addition to comparing the coaches’ records through their first five seasons, I also wanted to be able to use consistent metrics. The advanced statistical rankings didn’t exist back into the early 2000s, and recruiting has also changed significantly since then.  Each metric was compared head-to-head with Harbaugh’s results to create an Advantage %. Records and stats came from Sports-Reference.com.  Recruiting rankings came from 247sports.comF+, FEI, & S&P+ rankings come from FootballOutsiders.com.   Here are the 17 metrics that I compared:

  • Final Rankings in F+, FEI, & SP+  systems (SP+ now on ESPN in 2019)
  • Overall W-L record
  • Offensive Pass Yds / Game & Rush Yds / Game
  • Points Scored / Game
  • Defensive Pass Yds / Game & Rush Yds / Game allowed
  • Defense Points Allowed / Game
  • Big Ten W-L record (if applicable)
  • W-L Record vs. OSU (or record vs. Michigan for OSU coaches)
  • W-L Record vs. MSU (record vs. Michigan for Dantonio)
  • W-L Record vs. Notre Dame (record vs. Michigan for Kelly)
  • W-L Record in Bowl Games
  • Number of NFL Draftees
  • Recruiting Class Rank (cycle following the season)

RESULTS

ANALYSIS

The summary table above lists the coaches from least favorable comparison for Jim Harbaugh (Jim Tressel – final 5 seasons) to most favorable (James Franklin) sorting by ascending Harbaugh Advantage Percentage.  Using this method, Jim Harbaugh falls between Lloyd Carr (final 5 seasons) and Chris Petersen’s 5-year tenure at Washington. 

If we only look at the initial 5-year start for each coach (removing Tressel & Carr), it’s clear that Nick Saban and Urban Meyer are in a league of their own.  Urb won a national title during the first playoff season in year 3, and made a return to the playoff in year 5. Saban won his 2nd national title at Alabama in year five, and went back-to-back in year six.  

Once you move past those two historically successful coaches, there is a very tight grouping between Jim Harbaugh and the other 5 coaches I selected.  Each coach averaged a 9-4 record through his first 5 seasons. I was surprised that using my Advantage Percentage, Harbaugh compares favorably to all 5 of those coaches, despite Brian Kelly reaching the BCS Championship game in year 3.  I also included each coach’s Year 6 record for a possible glimpse into the future. Year 6 results range from unacceptable (7-6 for Dantonio) to achievable (11-2 in 2019 for Franklin).  

One other interesting note on Big Ten record: If Harbaugh had managed one more B10 victory, his average record would have rounded up to 7-2, and it would have been +1 over Carr.  Had he suffered one more loss, the average record would have rounded to 6-3 and been -1 to Carr. The program has clearly returned to the Carr level. I think we’re all very anxious to take the next step forward.