College Football Rankings: A Methodology That Makes Sense

When the University has a break around the holidays I don’t spend all my time thinking about fruit production (shocking, I know). Like many males in this country, during the holiday season I spent a lot of time watching college football. Since I work for a major university, worked previously at another, attended two more, and all of the teams I have had a connection to were playing in a bowl game I spent plenty of time thinking about college football. I think the current playoff system worked out for this year. The committee, I believe, got the four teams right and we are on the cusp of seeing how the whole thing plays out. What I don’t like about the current system (and past systems) is that we have no idea how those final teams were decided upon. There is a lack of transparency. What are the factors being used to determine the final four? Performance statistics? Eye Test? Strength of Schedule? Margin of Victory? It is all so unclear. Being an academic, and as long as the football team is associated with the university name, I believe there should be an academic component to the final rankings as well. The most equitable (and competitive) would be Fall semester grade point average of the football team. Now, I don’t think it should count for a large portion of the final overall ranking system, but it should be a component – after all, aren’t these “student-athletes”?  ESPN has done its own type of work in this area, but they used graduate rate, which is a long term metric.  It includes players who did not even contribute during the current season.

So, over the course of the last couple weeks I developed an easy system that anyone could use to rank college football teams. When I implemented it, the four top teams were the same as those put forth by the committee (although in a different order). How does this work? The rules are as follow:

1. A winning team accumulates the number of wins from the team it defeated. For example, if Team A beats Team B (who has 6 wins on the season), then Team A accumulates 6 points.

2. In the same vein, losses also accumulate, but penalties may occur if: a) a loss on the road has no penalty, b) a loss at a neutral site has a penalty of -0.5 points, and c) a loss at home as a penalty of -1 point. The take-home message is that elite teams should not lose at home. For example if Team A losses to Team C (who has 3 losses on the season) at home, Team A will accumulate -4 points.

3. A team shall not accumulate any points when playing a lower division team, but it will accumulate the losses. I understand there is $$$ involved here and that these games will still get scheduled. But do these games prove anything? In a win, no (higher division schools should have no problem beating lower division schools), but in a loss, yes (in this case the team isn’t good anyway and will not need to worry about being ranked).

4. Points are totaled, and then divided by the number of games played.

5*. The Fall grade point average will be added in this manner: the football team with the top GPA during the fall semester will get 1 point. The lowest will get 0 points. All others will get a percentage of 1 point. For example, if Team A has a Fall GPA of 3.50, the highest of any college football team in its division, it would receive 1 point. If Team Z has a Fall GPA of 2.20 and is the lowest of all teams, it would receive 0 points. The difference between 3.50 and 2.20 is 1.30. The formula would then be [difference in team GPA from lowest school/difference between high and low GPA] = points to be added to ranking. So, if Team G has a GPA of 2.50 this would be 2.50-2.20/1.30 = 0.3/1.3 = 0.23.

*This is a “sticky wicket” in my ranking system. These numbers may be hard to obtain. Until they become widely available via a website or some other method, numbers 1-4 should be used as the criteria for ranking.

Okay, we have the rules established, so how did the teams from 2014 end up? My rankings, using the system above (exclusive of GPA) are listed below. The beauty of this is that anyone can see how teams rank using a little addition and subtraction. Using this system rewards teams for scheduling good teams and penalizes them for scheduling a lot of poor teams and lower division teams.

2014 Final College Football Team Rankings [updated to show final results]

1. The Ohio State University [6.93]

2. Boise State University [6.14]

3. Florida State University [5.89]

4. University of Oregon [5.50]

5. University of Alabama [5.25]

6. Texas Christian University [4.77]

7. Marshall University [4.50]

8. Michigan State University [4.31]

9. Baylor University [3.73]
I only ranked teams with 2 losses or fewer. The Ohio State University and Boise State University were helped by the fact that they did not have a lower division school on their out-of-conference schedule. TOSU had a devastating loss to Virginia Tech University at home, but they were able to compensate for that by playing other decent teams. Boise State had two road losses, both to good teams.  You might notice that if Oregon were to lose the NCG they would drop to #4, below Florida State whom they drilled in the Rose Bowl.  The rankings presented here look over the entire season and not just one game.  Besides, does it really matter who ended up #3 or #4? No, the only one that counts is #1.

Teams in this system would also have an off-field “classroom competition” that would act an in impetus for students to perform better academically. Having this as a component of the ranking would also perhaps subtly encourage coaches to recruit students with better academic backgrounds.