For my first piece on 18Stripes, I had to think about how I could contribute. I know less about football than Burgz or Larz (and others), and very little of recruiting compared to Brendan and Tyler (or anyone). However, I know a fair bit about scientific research – and I know a lot about hating other teams. My particular prompting for this piece was that the team I hate most has changed over time. From 2002-2010, the team I most hated was USC. From 2011 to 2014, it was Michigan. From 2015 until the present, it’s been Stanford. I was surprised by this, and I’m wondering how much of my own craziness can be explained by science. It turns out that the answer is: “A fair bit!” Some people have studied this a lot.
My goals here are to:
- Summarize the theory and evidence of sports team rivalry, as laid out by researchers who specialize in this subject
- Test how this explains my personal rivalry rankings
- Have some fun and hear your stories of loathing other teams
Outline of the Science:
Let me start with an important disclaimer: When I say ‘science’, that term should be treated loosely here. I’m sure the authors are doing their best, but at the end of the day, when people talk about ‘science’, they’re not talking about some objective force; they’re discussing currently accepted theories that might be based on flawed data (or flawed interpretations of that data). This is true even in the hard sciences (physics) and doubly true of any social science (like economics, psychology, or sociology). The next time you read a headline that says ‘Science says X’, realize that article has been written (or at least edited) by someone who hasn’t spent a lot of time on ‘science’.
The work I’ll discuss today comes from by B. David Tyler and Joe Cobbs. They’ve made a career studying sports rivalry and there are two papers in particular I’ll be summarizing: one is a discussion of the factors that determine rivalry and another is a careful look at measuring rivalry.
Let’s start with the factors that define rivalry. In their paper “Rival Conceptions of Rivalry: Why Some Competition Means More than Others” [download here], Tyler and Cobbs sent out written surveys that were filled out by 37 fans. The researchers (and probably some underpaid graduate students) then verbally coded different aspects of the responses. So, for example (from the paper): “The teams play in the same division and meet many times each year” would be coded as ‘frequency of competition’. Then, once they’d coded the qualitative text into a binary variable (e.g. frequency of competition – yes or no) they could tally all the different factors that defined rivalry. They initially coded 25 different variables, but then reduced / collapsed the coding to 11 different variables.
If you’re like me, you’re alarmed that there’s only 37 data points (read: 37 people completed the survey / wrote essays about what rivalry meant to them). I think the good news is that the researcher findings match what you might intuitively respect. Even better, they ran another survey with an N of 472 fans to see which of these 11 variables people reported as most important to feeling a rivalry towards a team. Once they had reported importance scores for each of these dimensions from the fans, they used a technique called Principal Component Analysis to determine the number of factors that best explained rivalry. They found three main factors explain 60% of the variation in report rivalry scores; based on the relationship between these factors and the 11 different recorded variables, they decided to call the three core factors: Conflict, Peer and Bias.
The list that follows is verbatim from Tyler’s and Cobb’s website; it shows that each of the 11 reported contributors is most correlated with only one major factor; the only exception was competition for personnel, which was almost equally correlated across Conflict, Peer, and Bias. After each of the 11 elements, in brackets, I record the average score (on a scale of 1-6, 6 being highest importance) that the 472 fans assigned to its importance in defining their reported rivalry.
- Conflict – There are five observable factors that define the first principal component:
- Defining moment [4.72]: Specific incident, positive or negative, between the competitors. For USC, this is the Bush Push. For Stanford, this is the glorious 2012 stand. For Michigan, you can pick. I’ll choose #RememberTheSix.
- Frequency of competition [5.12]: Recurring competition between the opponents. USC, Stanford, and Michigan all check this box – recently, at least.
- Historical parity [4.7]: Comparable success over a long period (greater than 10 years)
- Recent parity [4.72]: Comparable success within the last 10 years – As we’ll see, this is true of USC, Michigan, and Stanford
- Star factor [4.7]: Extraordinary individuals (performers, personalities, or legacies) – All three schools have produced notable talents
- Peer – There are three factors that define the second principal component:
- Geography [4.61]: Teams are located close to each other – this is more true of Michigan than USC or Stanford
- Cultural similarity [4.12]: Shared values between the teams/institutions/cities – here, I would say Stanford and Michigan are more similar to us than USC, but reasonable people might disagree. Despite their recent hilarious scandals, USC is higher ranked than Michigan in US News and World Report’s ranking of undergraduate universities.
- Competition for personnel [4.38]: Competition for recruits, coaches, players.
- Bias – There are three factors that define the third principal component of rivalry:
- Cultural difference [3.51]: Disparate values between the teams/institutions/cities
- Relative dominance [4.52]: One team aspires to overcome the historical success or dominance of the other team
- Unfairness [3.37]: Perceived preferential treatment toward one team by league or competition authorities (e.g., governing bodies, referees). I mean, we complain about this, but I feel this is much more true of Ohio State (never investigated seriously for NCAA violations under Tressel and Meyer) than it is of any of USC, Michigan, or Stanford.
Conflict, Peer, and Bias are not equally important. In using Principal Component Analysis (PCA), the authors found that Conflict and Peer explained more about rivalry than Bias. Intuitively, this matches with the higher average score the fans assigned to each component of Conflict and Peer.
This research helps explain the shift over time in the degree to which I hate different teams. As I mentioned above: from 2002-2010, I hated USC the most. From 2011 to 2014: Michigan. From 2015 on: Stanford. Let’s see how the factors of rivalry for each of these three teams have changed over time.
Let’s start with USC. They claim to be an academic peer (cultural similarity) but most patently are not. While not a nearby school (geography), they have historical parity in terms of football performance and they meet the “frequency of competition” criteria. The Bush-Push is perhaps the worst defining moment of my college football experience (although it’s the poor spot on the previous fumble out of bounds that really upsets me).
What could have happened since 2009 that made me hate them less?
Ah, I see. They went from beating us eight years in a row to starting to lose to us more often than not. But that’s only one factor. Maybe something else happened to their program?
I’m gratified to see that their win percentage (recent parity) has fallen from an astounding .900 (Pete Carroll cheated, but was also a good coach!) to a more reasonable level. Let’s hope the decline continues. We should note that our own five year average will take some time to improve (thanks, 2016!), and in general, this is a pretty slowly changing measure. That makes USC’s decline all the more notable. Eric’s comments about how USC is basically decades of mediocrity sandwiching Pete Caroll seems… empirically descriptive. I would also note that there’s been some competition for personnel, but the diverging academic requirements (at least, for athletes) of the two schools have seemingly reduced this over time. Please correct me if I’m wrong, recruitniks. All of these factors contribute to a decline in the relative sense of rivalry.
We see Michigan was beginning to improve as USC’s performance (both in general and against ND) declined. The above chart of win percentage makes it hard to see – for which I apologize – but from 2010 – 2014, Michigan’s five year win pct against ND was 80%, 80%, 60%, 80%, and 60%. We would lose to Michigan in bizarre and horrible ways (defining moments), and in 2009 at the Big House we were on the wrong end of two of the worst referee / review calls I’ve seen watching ND (unfairness). Michigan is just down the road (geography) and another historical peer – #1 and #2 in win percentages.
But then 2014 happened. 31-0 (37-0!) happened. The failure of Brady Hoke made it clear that Michigan was stuck in neutral, and their five year win percentage began to fall, and with the failures of Hoke’s latter seasons, the perception of recent parity disappeared. Most importantly, we suddenly stopped playing Michigan every year – the frequency of the competition disappeared. As of this year, they have beaten us zero times in the last five years… and we’ve only played them twice.
The rivalry I felt toward Michigan has faded. Sure, I’m sure I experience the same sense of schadenfreude about their losses that most of us do – and more so for Michigan than for programs we have played less frequently in recent times, like Florida State. But I don’t compare Notre Dame to Michigan anywhere near as much as I used to, and when we stop playing them (I hope forever, because the game really brings us zero benefits as a team) after 2019, I expect the difference to be even more noticeable.
[Side note: I think the rivalry with Michigan for me was much more about dealing with Michigan fans. Two of my best friends are UM fans and incredibly classy about Michigan’s wins and losses. But the rest of the Michigan fans I know are among the least gracious fans on the planet. So it was highly important to me that ND win those games. Losses to USC never brought the day to day misery that losses to Michigan did.]
Now we get to Stanford, who I’ll say that no Notre Dame fan ever hated until very recently. I didn’t even think about them at all until 2010, when I found myself thinking: “Darn, these guys look good!” By 2011, I was experiencing frustration about our losses to them. Stanford is very well explained by Tyler and Cobbs’s research. We weren’t even playing them until 2007 (this is why their five year win pct against ND is missing for the first several years of the chart above), so until recently there has been zero competition. Now they’ve got a five year win percentage against us of .6 (which will stay that way even if we win again this year) and their 5 year winning percentage has been above ours since 2010. We are most definitely competing with them for personnel, in large part because they are the one D1 football program where the academic standards (at least for entry) are higher than Notre Dame. For me, this constitutes cultural similarity – at least relative to other D1 schools.
I’ve been on Stanford death watch for five years. Credit to David Shaw: I’ve been wrong every year. But this year I’m not alone. I’m hoping that the forecasts of Bill Connelly, ESPN, sundry college football gamblers, and our own Brendan Reilly come to pass and that Stanford has another year of under-performance. They were supposed to have a gangbusters offense in 2018 and they did not; this year the veteran offensive line (and almost all their skill positions except QB and one WR) are gone. Their recruiting has slowly declined in recent years and they’ve had a lot of coaching attrition.
Please, God. Please.
I’d like to finish with one small summary of another paper: “All Rivals Are Not Equal”. In this paper, Tyler and Cobbs point out three more properties of rivalry:
- Teams can have more than one rival
- Rivalry is continuous in scale – there are different degrees of rivalry
- Rivalry need not be reciprocal – one team can see another as it’s rival without the feeling being returned
If you play around on the Know Rivalry Website, you’ll see that the 8th most unbalanced (not reciprocated) rivalry is the “Little Brother” MSU, who’s fans scored UM as a 67 out of a scale of 1-100 on rivalry points. UM fans scored MSU as only a 17.
The other one that jumped out at me was the #1 most unbalanced rivalry: Boston College fans scored ND as 72 out of 100. ND fans evaluated BC as a… 2. That’s right: 2 out of 100. Given when I started being an ND fan, I am delighted to say that thanks to Brian Kelly, my concern about BC has reverted to the population norm.
If you’ve made it this far in the article – thank you! I’d love to see your comments and hear your own thoughts and stories on rivalry. Looking forward to spending more time with you as we get nearer to fall.