Here is the Johnson Index rankings after week 1:
| Rank | Difference | Team | Win | Loss | Score |
|---|---|---|---|---|---|
| 1 | - | Miami | 1 | 0 | 1000 |
| 2 | - | LSU | 1 | 0 | 969 |
| 3 | - | South Carolina | 1 | 0 | 968 |
| 4 | - | Texas A&M | 1 | 0 | 967 |
| 5 | - | Florida | 1 | 0 | 963 |
| 6 | - | SMU | 1 | 0 | 960 |
| 7 | - | USC | 2 | 0 | 959 |
| 8 | - | UCLA | 1 | 0 | 951 |
| 9 | - | Oklahoma | 1 | 0 | 947 |
| 10 | - | Ole Miss | 1 | 0 | 945 |
| 11 | - | Penn State | 1 | 0 | 943 |
| 12 | - | Mississippi State | 1 | 0 | 934 |
| 12 | - | Iowa | 1 | 0 | 934 |
| 14 | - | Notre Dame | 1 | 0 | 932 |
| 15 | - | Ohio State | 1 | 0 | 926 |
| 15 | - | Virginia | 1 | 0 | 926 |
| 15 | - | Pitt | 1 | 0 | 926 |
| 18 | - | Nebraska | 1 | 0 | 925 |
| 19 | - | Cincinnati | 1 | 0 | 919 |
| 20 | - | Indiana | 1 | 0 | 917 |
| 21 | - | Texas | 1 | 0 | 914 |
| 22 | - | Illinois | 1 | 0 | 912 |
| 23 | - | Oregon | 1 | 0 | 904 |
| 24 | - | Houston | 1 | 0 | 898 |
| 25 | - | West Virginia | 1 | 0 | 897 |
Conference breakdown:
- 9 B1G
- 8 SEC
- 4 ACC
- 3 Big 12
- 1 Independent
Another season of the best sport in the world kicked off this week, and so follows another season of the Johnson Index. LSU destroyed Clemson, big schools beat up on teams that nobody could place on a map, and Michigan was gifted a win and proceeded to storm the field as 28 point favorites.
College football is so back.
But this blog isn't meant to be about the games as much. There are plenty of places for people to go for that analysis. I like these posts to cover the art of algorithmically ranking teams and the struggles faced.
And unfortunately it's week 1 for everyone, including software engineers ranking college football teams. I didn't do my due diligence with testing the Index before games kicked off. During the offseason, it looks like the website I used to get my stats added some bot protection. This means that in the morning when the script runs and tries to get the results of all of yesterday's games, the website blocked me and there was nothing to build the rankings on.
So we had to scramble. Claude and I worked tirelessly through the night to switch the backend data supplier so that there'd be limited delay for the loyal followers of the Index. The poll is now backed by ESPN statistics, and I would like to welcome them to the team. This is fitting, as I plan to one day sell my algorithm to ESPN for $1M. No more, no less.
The first couple weeks of the Index are my favorite. It's where you get takes that annoy people the most. Fans of teams like Ohio State (14) or Georgia (T-86) might think that they are underranked. I could argue the opposite, that the AP Poll weighs preseason rankings way too heavily. Every year there's a team that starts preseason ranked high and it takes them way too long for them to drop out even though they haven't earned anything. Last year's preseason AP Poll had Penn State #2 and Clemson #4 before their seasons imploded. Notre Dame started out 0-2 yet the AP Poll still had them ranked in week 3. Poll inertia (where you can only drop a team a certain number of spots week to week) never made sense to me. It is especially egregious when the team is only ranked highly because of a preseason poll that was made based on vibes.
I also notably don't rank wins against FCS teams. This is just a decision I made a few years back because FCS teams aren't in the rankings at all, so beating them 73-0 doesn't tell me anything. I'm already getting some choice feedback on this from fans of teams that destroyed an FCS school this week (not naming names). Speaking of FCS, I would like to welcome to the FBS North Dakota State and Sacramento State. I have a list of FBS teams in the codebase that I need to manually update every new season when teams are promoted. Kind of a fun way of adding them to the Index.
So that's mostly it. Don't look too deeply into the rankings this early, there's bound to be a lot of shakeup in the next couple weeks. There's a great slate of games coming up next week, enjoy it! We only get so many college football saturdays each year.