SI: Measuring Each Power 5 School’s Conference Value
While looking for something else, I came across this article on Sports Illustrated. It basically ranks P5 schools based on who would be most desirable for conference expansion.
I'm not really all that interested in conference expansion, but what I liked about the article is how the ranking includes a mix of football and non-football factors to rank each team's desirability. Many here were talking about the importance of academics and/or non-revenue sports in the multiple threads on expansion and future B1G members, so I thought this would be of interest to the board.
The different components of the SI aggregate ranking were:
- "Football ranking: This is a five-year average of the Sagarin ratings from 2017 to ’21"
- "Academic ranking: This is simply the most recent U.S. News & World Report’s national universities rankings, released in 2021"
- "All-sports ranking: This is the Learfield Directors’ Cup Division I standings for the 2021–22 academic year"
- "Football attendance: This was an average of home-game attendance (via NCAA data) from 2017 to ’21, tossing out ’20 since that season was heavily impacted by the COVID-19 pandemic"
- "Broadcast viewership: This was the total number of football games that drew one million or more viewers: from 2017 to ’21, also tossing out ’20"
Based on this index, the top-10 most desirable teams (with the individual component scores and overall score) were:
Other B1G teams of interest in the top-25: PSU (11), MSU (20), Iowa (21).
In terms of overall conferences, as expected the SEC and B1G have the highest average scores for all teams (25.1 and 25.8, respectively) with a pretty large drop for the ACC (39.6), (Pac-12 41.4), Big 12 (49.3).
Additional Analysis
As a quick side project, I thought it would be interesting to take these data and see how heavily the football and non-football components weigh on each team's overall ranking. Who are the well-rounded teams, who are the football factories, and which schools are, well, not so good at stuff? Just kidding. For this purpose, I created a separate measure for each team's average score for the football factors (football ranking, football attendance, broadcast viewership) and non-football factors (academic ranking and all-sports rankings). I then plotted each school along these two dimensions.
Keep in mind that lower scores on each dimension are better. The quadrants can be thought of like this:
- Lower-left quadrant: balanced in terms of football and non-football factors (You want to be here).
- Top-left quadrant: score is heavily influenced by non-football factors.
- Bottom-right quadrant: football factories.
- Top-right quadrant: Cave! Hic Dragones (You don't want to be there).
The first graph plots the top-25 in the SI desirability rankings. Teams are color-coded by conference (I included ND in the ACC).
Michigan, Notre Dame, and UT stand out as pretty balanced teams. OSU and UGA's scores are more driven by football, but still within the respectable range on the non-football dimensions. This visualization also shows that USC will be a pretty good grab for the B1G. As it stands, the B1G has 4/8 teams (with the addition of USC) in this quadrant. If we could pick off ND, then we'd have 5/8. Not too shabby.
Moving away from the most desirable quadrant. Stanford's ranking, as you might have guessed, is pretty heavily driven by non-football factors, as is the other newest member of the B1G (UCLA). Also no surprise that Bama is pretty much the football factoryist of the football factories. Miami and Tenessee? Haha.
The next graph includes all P5 teams (not just the top-25). Sorry about the small label sizes and I hope they are readable. I had to reduce them to avoid overlap and a really messy appearance.
I'll let you draw your own conclusions from these graphs. But I'll just say: holy shit, the Big 12 is screwed when UT and OU leave. All teams will be either in the lower-right quadrant (football factories) or top-right (don't go there!). Picking up BYU, UCF, Cinci, and Houston are just going to add to the overall weakness of the conference.
That's all I've got. Any surprises here? Any teams that you'd definitely like to grab if the B1G were to expand? Do these results change your minds about any teams?
this shows that washington would be more desirable than oregon in the push for 20... i would have though reversed. but i guess the non-FBS has other stuff besides basketball
In the SI analysis, "Viewership" is only 20%, and the ranking metric for that category (# of games w/ over 1M viewers) seems flawed--Oklahoma is not the most valuable TV brand in college football. I would guess that viewership and academics have a much higher weight to the Big Ten.
This is my biggest gripe with Forde’s method as well. Using only 4 years of data is a pretty small sample, and the 1M cutoff is too low to identify TV value. Using a 4M cut-off, the top 5 in order is:
1 Alabama
2 Ohio State
3 Michigan
4 Georgia
5 Oklahoma
which seems more accurate to me.
If the question is what schools are best for a conference to add, TV viewership is, and should be the biggest factor. Bringing in a school that dilutes the current or future payout to existing schools is not a win, period.
Washington has some solid academics. That would make them a pretty good fit.
Big Ten, please make this happen.
And they bring the Seattle market
A real hotbed of CFB viewership...
Viewership doesn’t really matter much. It’s not like people in NYC watch Rutgers or college football in general.
The appeal to large markets, such as Seattle, is that the B1G is able to include their games in cable packages in those areas. If you have a cable package, you know damn well there are channels in that package that you never watch. Yet, you’re paying for them. So whether people in Seattle watch or not, they’re paying for it if they have cable.
I’d actually prefer Washington over Oregon for the Big Ten, though it’s likely if we take one we’re also taking the other.
No way ND is higher than Michigan in academics. Looks like some massaging of data goin on here.
They might have a narrow focus on something at the undergrad level.
I always roll my eyes at this as well. I did check out the US News & World Report rankings and they do have ND at #19 somehow with Michigan at #23. Funny thing is Michigan basically beats ND across the board at graduate major rankings.
Yeah, it's huge public uni with wealth of graduate programs v. tiny private with rep for excellence in undergraduate ed. You look at the annual London Times ranking of the world's best schools and M has--when I've looked--been around number 15, ND nowhere to be found. But. .. carry on, SI, carry on.
As a graduate, a huge point of pride for me is having gone to a public university. I know that Michigan's endowment is obscene and none of these issues are black and white, but let's just say that I'd rather have gone to the school that Iggy Pop and Arthur Miller attended than the one where Amy Coney Barrett went. :)
And the one with the original Big House v. Half Pint House.
I'm 0-4 at Michigan games @ ND, so think I'll stop going now. But my first was in 2004. There was an awful PI call against us in the endzone. I started bitching about it. Then some ND fuck a few rows ahead of me turned around and said "go read the rule book, oh wait, you went to a PUBLIC school". I can't believe I kept my cool.
Can you reformat that to have the x and y scales actually match? it makes it look like UM is clearly superior - and we are - but by these metrics, that's not true.
I saw this as well and after the chatter last week about us being in all the most watched games - i was surprised that we were 5th in viewership
Thanks!
Thanks!
Do you mean setting the max of the x and y axes to be the same value (like in the first graph) or changing the location of the quadrants? The first change wouldn't do anything to the visuals. I suppose it might have made sense to set the quadrants for the second graph along the 50th percentile for each dimension (or similar), but I wanted to keep it consistent with the first graph. Again, though, that wouldn't have any effect on Michigan's position on the graph.
The viewership data are 5-year averages, which might have influenced the rating there.
I just realized that I changed the axis titles in the second graph. (kicking myself) They are all measuring the same thing, just titled slightly differently ("score" and "rating" are the same thing).
Maybe its a browser setting that is adjusting the picture somehow, though i doubt it...
What i see is that the distance between the data scale points on the x axis are wider than vertically on the y access. So as a result, the vertical distance looks mroe scrunched than it should.
look at the space between 5 and 10 on each. if its the same for you, then its my screen, but if not.. that's what im referring to...
It essentially diminishes the value of y because it scrunches similar datapoints together moreso than x values.
Ok, now I see. There is a bigger space between 5 and 10... on the x-axis than the y-axis. You have really good eyes!
I don't know how to change that. My stats program (Stata) sets that automatically. I've been meaning to teach myself Python to up my data visualization game...
same length blue line...
that's what she said.
Outstanding post!
I would love to see this as the BIG TEN CONFERNCE for football.
Gracias! This really makes it look like picking up USC was a great move. Based on these data, grabbing ND and Washington would really make it a powerful conference across the board.
Great post!
Oklahoma being the #1 in viewership doesn't seem right to me. Can anyone explain this to me? The other in the top 9 seem fine, but I don't see how Oklahoma is above M, ND, Bama, and Georgia without a title game and with an Alamo Bowl appearance. Unless viewership is only based on the 2020 season?
Thanks!
That is odd. They actually excluded the 2020 season so that can't be the reason. I don't know how to explain that one.
Their metric says "Broadcast viewership: This was the total number of football games that drew one million or more viewers: from 2017 to ’21" - I'm going to guess this means OU's numbers are goosed a bit by walking all over the Big 12 for a few years. That's a lot of Big 12 championship games and high profile bowl games that count as 1 million viewer games for OU but that's a flawed methodology to include those games - people are tuning just as much because it's a CFP semifinal as they are to see OU. Show me the total viewership in the regular season and I'll bet my arm Michigan beats OU comfortably
Miami is the 35 year old pregnant stripper of college football.
The first item in their criteria is much more dynamic than the last 4. Michigan could climb into the #1 spot fairly easily with another good season or two.
I thought this, too. I feel like they need some sort of combination of the last 5 years and the last 25 years.
Honestly for realignment and conference dollars, I'd advocate the last 50 years are far more consequential than the last 5. The fact that Iowa State had a few good seasons changes nothing about the perception of their program on the whole
For some reason, your post made me think of Lisa Simpson...
"As intelligence goes up, happiness goes down. See, I made a graph. I make lots of graphs." ~Lisa Simpson
This makes FSU more appealing. I hate the race to 20, but FSU, UNC, Stanford and Washington seems like a good pull.
With a few good years of football, FSU could make a move into the bottom-left quadrant. As it stands, they seem like a pretty good addition. I don't think Florida wants another in-state team in the SEC, so they could be an easy draw for the B1G.
Surprises? Why yes. I am surprised Michigan is not #1, and we are behind Ohio St, of all damn schools.
Look at it this way. They have to win a whole lot against UM to beat us. All we would need is one more win and we will be over them. That is the power of being good at many things.
If anything I think this confirms that we are near the end of conference expansion in the current climate. There's just not that many desirable programs left that are not already part of the B1G or SEC.
After Notre Dame, the next highest ranked team is Washington, who is mostly there by viture of being mediocre across the board.
^^ this
I really don't understand how M is so low on their football score. 16th seems really low. I don't see how teams like Iowa State, Texas, A&M, Okie State and even Penn state are ahead of us in the period of 2017-2021.
well you could go back and look at the rankings for starters.
But that is a poor metric. Michigan is much bigger of a football brand to be ranked 16.
I didn't make up the rules, it's based on actual football results/rankings over that time period.
It's a terrible, arbitrary metric they picked. Anyone thinking any of those schools would get picked before Michigan is huffing glue and SI is apparently one of those people
Is Sagarin that bad of a rating system that they have a Texas team that went 37-25 over that time frame ranked four spots ahead of a 41-18 Michigan team in that same time frame? I guess having a worse record in a much worse conference = better to Sagarin. Maybe bonus points for losing to Maryland multiple times? Time to switch to a ranking system that isn't complete horseshit. Hard to take any of this seriously when you see that kind of stuff.
Great post! The biggest surprise to me was how valuable Wisconsin seems to be. They always strike me as kind of a mid-tier program. Maybe not though.
Yeah, that was a bit of a surprise to me as well. Florida also caught my eye. I knew they were decent academically, but their academic ranking (only three spots behind UM) and performance in all sports caught me by surprise.
This is awesome. Thanks for this OP!
It makes it pretty clear what is worth it and what is not.
Grab ND & either FSU or Washington. I'd lean Washington because I think they are better positioned to excel in football with an influx of money and it keeps the Big Ten vs. SEC mostly a north vs. south battle. It makes for a ridiculous wonky conference for scheduling, but whatever at this point.
Thanks! I'd rather see Washington than FSU as well. Great football atmosphere and I do like the North-South distinction. One advantage to FSU is that it could help with recruiting in Florida. Maybe?
I would choose FSU many times over because of the recruiting angle.
I would have loved to see Texas...over anyone else-along with USC would have been the real coup.
Lots of math in this blog post.
But is has all those pretty color-coded pictures. Even a koala could understand.