Look, before we get started on this post, I want to say two things: (1) I have never had so much apathy for college football in my life. I don’t know what happened to me. Someone, please, shake me from this slumber! (2) Really, the rankings should only be for the top five teams. Who cares about the rest? Great, you are #18 or you made the Cotton bowl at the end of the year, so what…
Joke of Rankings
I say this every season so take it for what it’s worth: I think preseason polls in CFB are a joke and don’t believe there should be any polls at all until the first week of October. After a month of games, injuries, suspensions, etc. true contenders & pretenders come into focus.
— Todd Blackledge (@Todd_Blackledge) September 5, 2018
Quick thought on Week 2 of the CFB season…The teams that won last week are likely not as good as we think, and the teams that lost are likely not as bad as we think…The beauty of Football lies in the day to day grind pursuing improvement
— Joel Klatt (@joelklatt) September 8, 2018
Todd and Joel have a much larger platform than me, and get paid to talk college football, but this is something that I have been saying for literally all of my life. Glad to see someone heard me. I usually start most conversation with this line, “See the rankings for college football? Such garbage. Right? What was your name again?” Works. Every. Time.
I remember all to well in 2015 when my beloved Utes were ranked in the top 25. We had just come off a 9-4 season in 2014, opened with a win at Michigan and posterized Oregon in week four. Then came the brutal wake up call with USC. The #3 ranking the Utes had meant nothing and we fell alllll the way down back to the Vegas bowl at the end of the year.
Do they matter? When do they matter? Should I care?
If we are talking about your grades, then: yes; from the start; and yes, you should care. School is important and especially for college football players…..But as we are talking about college football rankings, the answers are a little different. It depends on: what school it is (shockingly, tradition matters!), the week (I think we can all agree a ranking in week 13 is more valuable than week four), and position of the rank (is there really a massive difference between 25 and 18?).
It is hard to nail down exactly when rankings matter. Let me clarify, as that last sentence is misleading. (maybe I should delete it? Nah. Lets goo!) What I mean is, as a fan, when should YOU care about the ranking; like, how confident should you be in that ranking. And if that is still not clear, let me give you an example (its similar to that story about my Utes).
Let’s say that you are rooting for your alma mater (which is latin for “place of debt”) and the preseason rankings come out and lo and behold, your boys are ranked #8! Does that actually move the needle? Does that get you pumped? Do you put any merit into that ranking? You can feel however you want to about it, who am I to tell you how to feel, but what I can tell you is that the fellas that put the rankings together don’t really know…They definitely don’t for the majority of teams (outside the top five) in the preseason.
Let’s do some analysis
I looked at the 2015, 2016, and 2017 seasons. This is not a massive sample size, but you what do you care? Oh, the numbers may be very different if I included more seasons. Below I give a few takes or analysis I did.
In this first one I asked a simple question: How accurate are the rankings week over week? Specifically on the top 10 and five, cause again, who really gives a rip about any other ranking? To do this I found the frequency of teams that were ranked in the top 10 and top five week over week throughout the season.
In the below tables you will see how many times a team was in the given ranking bracket (top 10 and top five) from 2015 -2017. I limited the tables to 10 and five teams respectively. This ranking is inclusive of every published AP poll, which is preseason to week 15.
- Alabama is of course at the top of the list for both. 89.13% of their top 10 appearance are top five rankings. Clemson closely matches that rate with 88.37%.
- Not on the table, but interesting to note, is that Auburn has been in the top five just once. Whereas Louisville has been in the top five six times (thanks Lamar).
- There is an obviously decrease in accuracy with expanded teams. With notable drop offs from the top two to three teams. And it is especially hard to get it dead on when you contract the bucket of rankings. There are about 13 teams per year that appear in the top five list.
|Top 10 Ranking||Total Appearances||% of Times in Top 10|
|Top 5||Total Appearances||
% of Times in Top 5
The top 25 list averages about 46 teams appearing each season, with some obvious frequent faces. Alabama, Clemson, and Ohio State have always been ranked in the top 25. Oklahoma wasn’t ranked in the top 25 just once over the past three years. On the flip side there are 16 teams that appear on the rankings five times or less. Most notably, Texas (5), Miami (4), Kansas State (3), BYU (2), and Texas Tech (1).
Here are the average accuracy rates per ranking bracket over the past three years combined: Top 25 – 38.46% ; Top 10 – 26.97% ; Top 5 – 20.75%. Makes sense, it is much harder to really nail down the very best teams every week, right from the start. However, there are teams that live up to the billing, like Clemson and Alabama in 2016, who were ranked in the top five every single week (Alabama did it again in 2017).
If you were to look at the ranking path per team week over week, the chart would be an absolute cluster. Rather than doing that, let’s look at the path of where preseason favorites, meaning the top five, started and ended and then look at the path of the those who finished in the top five.
Looking at the charts below, you can see points of separation, or where there is a divergence or significant movement up or down. Take 2015 for example, everything is relatively smooth until week nine, the first week in November. In 2016, its week 10, which happened to be the first week in November. And then 2017, week 10, the first week in November.
I find it interesting to note that not one team that ended in the top five started outside of the top 25 teams. There are definite movements around week five to six (early October) and week 10 to eleven (early November).
Sure, there are limitations to this, but it is interesting to visually see the dramatic shifts and journeys of teams reaching a final ranking. No Cinderellas here.
Up next is correlation. The idea here was to correlate the top 25 team’s final ranking to each previous week. I was hoping to see for when the correlation significant and it stuck. See you could have a really high correlation in week one, but then the next week it completely drops off. Admittedly the correlation isn’t perfect, as I only had up to ranking #25. Meaning, that if a team was outside the top 25 their ranking would be blank. Also, in 2017 the rankings were dead on for the top five teams, but they were just in the opposite order. However, the correlation shows as zero.
Looking at the charts you can see quickly that week 10, early November, is when the rankings start to see the most significant correlation.
The AP poll has a long history, it dates back to the early 1930s, but honestly, the rankings mean little to nothing until well late in the season. Its just fun to create lists and have debates, but as well to market an otherwise unremarkable matchup. Would anyone watch Virginia Tech if they weren’t ranked? But all the sudden you rank them 22 and it will bring eyes to the game.
At the same rate, some… very few…teams warrant the hype. Take Alabama for instance. They’ve been rated top overall 106 times, the most of any team. And it helps, to be the top overall week over week carries a weight, a brand, continues a tradition. With all lists, its hard to quit talking about it.
——-// Update 9.14 //——-
Adding this seven minute segment from Joel Klatt. Incompetence is listed a few times…
"The voters in that poll REFUSE to move teams based on the games that are actually being played."@joelklatt says college football should get rid of the AP Poll. pic.twitter.com/EyCJHiX1c9
— FOX College Football (@CFBONFOX) September 13, 2018