On Nov. 18, 2000, Drew Henson scored on a fourth-and-goal naked bootleg to salt away a 38-26 victory for Michigan at Ohio Stadium. The win dropped John Cooper to 2-10-1 against Michigan. Not coincidentally, it was also Cooper’s last game at the Horseshoe.
History doesn’t back up Ohio State having a curse on Michigan
Recently, the Buckeyes have beaten both the Wolverines and expectations. But the Wolverines don’t just turn into mush when they see scarlet and gray.
Rick Osentoski-USA TODAY Sports
Bill Clinton was still president, and his successor would not be known for 24 days. Charlie’s Angels was the No. 1 movie at the box office. “Independent Women, Part 1” by Destiny’s Child was the number No. 1 single, having just displaced Creed’s “With Arms Wide Open.”
Michigan has not won in Columbus since. It’s only won twice in Ann Arbor.
Now, Michigan has another chance. At stake is the division title, probably a conference title (either team will be a significant favorite over Northwestern in Indianapolis), and a good shot at a Playoff berth. This is in addition to the usual bragging rights in play when the Wolverines and Buckeyes meet in one of the sport’s most famous rivalries.
Michigan is a 3.5-point favorite, an extreme rarity in this series and especially in Columbus, where they’ve only been favored once (in 2004) on an eight-game slide. S&P+ says the Wolverines are 7.1 points better on a neutral field, so about 4 in Columbus.
This raises the question: Has Ohio State won 15 of 17 against Michigan because the Buckeyes have had better teams, or because of some combination of the Buckeyes playing above their level in the big game and the Wolverines playing below theirs?
This is a relevant question going into the 2018 meeting. If the former is the case, then Michigan should have little to fear, because finally, the Wolverines have the better team.
If the latter is the case, then Ohio State should have confidence that the teams’ established levels through eleven games will not present a handicap.
To answer this question, we can look at the teams’ S&P+ ratings for the period starting in 2001, when Ohio State’s current run commenced.
S&P+ tells us what the difference between two teams should be on a neutral field. No computer system’s always right, but this one routinely beats Vegas spreads. With a three-point home-field adjustment, we can see how reality’s meshed with expectations:
- 2017: OSU beat S&P+ projection by 1.1
- 2016: OSU State beat it by 2.2
- 2015: OSU State beat it by 29.1
- 2014: Michigan beat it by 12.4
- 2013: Michigan beat it by 7.5
- 2012: OSU beat it by 0.7
- 2011: OSU beat it by 9.8
- 2010: OSU beat it by 9.8
- 2009: OSU beat it by 0.2
- 2008: OSU beat it by 12
- 2007: OSU beat it by 5.8
- 2006: OSU beat it by 0.3
- 2005: Michigan beat it by 0.6
- 2004: OSU beat it by 13
- 2003: Michigan beat it by 7.4
- 2002: Michigan beat it by 1.1
- 2001: OSU beat it by 15.2
Ohio State has done better than the computers say it should’ve against Michigan this century. By how much? That depends.
On the one hand, Ohio State’s outperformed S&P+ by an average of 4.1 points per game over the last 17 years, including at least slightly in 12 of 17.
On the other hand, the average is dragged in Ohio State’s direction by one particular game: 2015’s, in which the teams should have been even (Ohio State was 2.9 points better per S&P and the game was played in Ann Arbor) and the Buckeyes won 42-13.
Ohio State’s over-performance in 2015 is almost double any other delta between the expected and actual results. If we look at the median instead of the mean, the Ohio State’s over-performance is 1.1 points per game, a negligible amount.
In seven of 17 games, the margin’s been within 2.2 points of what S&P+ would project. Seven times, someone’s outpaced that by more than one score. Six of those times, the team outperforming expectations has been Ohio State.
Ohio State shouldn’t (and likely doesn’t) expect Michigan to play badly just because this is Michigan going up against Ohio State.
With the standard caveat that this is college football and the sample sizes will always be too small to establish statistical reliability, these numbers don’t prove Ohio State is likely to exceed its usual performance level or that Michigan is likely to lag behind its own.
The median is probably a better measure than the average with a once-a-year sample over 17 seasons, especially with one huge outlier. And that number is small. Ohio State performing one point better than expected this weekend would end with Michigan celebrating its first win in Columbus in 18 years.
In three years of Jim Harbaugh-Urban Meyer matchups, Ohio State’s done better than the S&P+ number each time. But the only major example of that was 2015. The three other times Meyer’s OSU has played Michigan and outpaced the projection, it was by 2.2 points or less each time. In pure performance terms, Meyer’s showings haven’t been that surprising.
If you wanted to find famous periods in this series’ history that suggested Michigan was over-performing or things were dead even, you could. UM’s 58-49-6 all-time mark is largely the product of dominating in the early 20th century.
But if we just look at the Ten Year War between Bo Schembechler and Woody Hayes, Michigan was 5-4-1 both in actual record and performance vs. S&P+ projections. (You can find historic S&P+ numbers here.) The Wolverines beat the computers by 2.5 points per game from 1969-78, a solid showing by the machines.
Maybe the most indicative game in that span? The teams’ memorable 10-10 tie in 1973. In that game, S&P+ would have predicted ... a tie.
See More:
