Stats Corner: The Idiocy of the Preseason Polls

The AP college poll began in 1936 after a three-way tie for the national champion caused a controversy between the fans of Minnesota, Princeton, and South Methodist (some things change, Princeton national champion in football, while others do not, fans complaining about rankings). In 1950, the AP preseason poll began, ranking Alabama as the number 1 team (certain things do not change) and has continued doing so for the last seven decades. The question proposed today is how effective is the preseason poll and how much of a boost do certain teams and conferences get from the poll.

The first and most important question is “how many times is the preseason number 1 the post season number 1?” The answer is 10: Tennessee in 1951; Michigan State in 1952; Oklahoma in 1967, 1974, 1975 and 1985; Alabama in 1978; Florida State in 1993 and 1999; and USC in 2004. The last two teams are the only ones to go wire to wire number in the polls. We have had 5 different presidents since the last time the preseason poll was correct, not exactly a great record although Alabama was the preseason number 1 this year, so there is still a chance that streak could be broken.

The other problem with the preseason poll is that it creates a disadvantage for certain teams, pick any undefeated G5 program, the only was to move up is for the team in front of you to lose. If they do not lose, you are not moving up. Think MWC 2004 Utah, they started at 20^th, and six weeks later after beating Texas A&M, Arizona, and North Carolina they only moved to 11^th. Then it was 2 weeks at 9, two weeks, at 7, and two weeks at 5 waiting for Virginia, Purdue, Miami, and Wisconsin to lose before they could move up. Even 12-1 Michigan had to wait until the teams in front of them lost before they could claim a place in the college football playoffs. Teams should not be punished because of their rankings which occurred before any games had been played.

We are going to compare the preseason and post bowl AP rankings for 2018 and 2019, we will throw out 2020 because of the pandemic and the chaos it caused games with schedules and cancellations, and the preseason and championship week rankings for 2021.

Dropped out: Wisconsin 8-5, Miami 7-6, Auburn 8-5, Michigan State 7-6, Stanford 9-4, USC 5-7, TCU 7-6 Mississippi State 8-5, Florida State 5-7, Virginia Tech 6-7, Oregon 9-4

Got in: Florida 10-3, Washington State 11-2, Kentucky 10-3, Syracuse 10-3, Texas A&M 9-4, Fresno State 12-2, Army 11-2, Utah State 11-2, Cincinnati 11-2, Iowa 9-4

During the 2018 season 11 teams started in the top 25, but dropped out over the course of the season. Shockingly, all 11 teams came from the Power 5 conferences, with 3 teams having a losing record and another 3 going 7-6. Of the 11 teams who got in 4 came from the G5, including independent Army, which is 36% and 4 more schools could be considered non-powerhouse programs Kentucky, Syracuse, Northwestern, and Iowa. Only 3, 27%, could be considered football programs, Florida, Washington State, and Texas A&M.

Dropped out: Texas A&M 8-5, Washington 8-5, Michigan State 7-6, Iowa State 7-6, Syracuse 5-7, Washington State 6-7, Nebraska 5-7, Stanford 4-8.

Got in: Minnesota 11-2, Baylor, 11-3, Memphis 12-2, Appalachian State 13-1, Navy 11-2, Cincinnati 11-3, Air Force 11-2, Boise State 12-2.

While the 2019 season was, technically, the most accurate with only 8 teams dropping out, once again all 8 teams were from Power 5 conferences and half with losing records. With Texas A&M, Michigan State, Syracuse, and Washington State being on the exiting end this year. It almost seems like the preseason poll is really just a rehash of the post bowl from the previous year. If you are a P5 team, we keep you in until you have a losing record then we have to vote you out. Of the 8 new teams, only two were from P5, meaning 75% of the teams who were added late came from the G5.

Dropped out: Iowa State 7-5, North Carolina 6-6, Wisconsin 8-4, Florida 6-6, Miami 7-5, USC 4-8, LSU 6-6, Indiana 2-10, Penn State 7-5, Washington 4-8, Texas 5-7, Coastal Carolina 10-2, Arizona State 8-4.

Got In: Michigan 12-1, Baylor 11-2, Ole Miss 10-2, Oklahoma State 11-2, Michigan State 10-2, BYU 10-2, Pittsburgh 11-2, NC State 9-3, Wake Forest 10-3, Houston 11-2, Arkansas 8-4, UTSA 12-1, Kentucky 9-3.

Prior to the bowls, more than half, 13 or 52%, of the preseason teams are out of the top 25. That could change slightly, increasing or decreasing, after the bowl games, but the preseason poll was not accurate. There finally was a G5 school dropped out, sorry Coastal Carolina 10-2 is not good enough to stay in the top 25, you are being replaced by 8-4 Arkansas. This year is probably the best example of the stupidity of the preseason polls, 2-10 Indiana at #19, 4-8 fired a coach USC at #17, 4-8 got a new coach Washington at #22, 6-6 got rid of coach who won national championship LSU #18, 6-6 we have the number draft pick and going to win national title North Carolina at #12. On the flip side 12-1 Michigan was not included, or was 11-2 Baylor. For the G5 there is 10-2 BYU, you are still a G5 for the time being, 11-2 Houston, ditto, and 12-1 UTSA. Other than Alabama at #1, the poll looks vastly different now when compared to the preseason.

Final Thoughts

Over the three years there 75 teams ranked in the top 25, 32 of them or 42.7% were not there at the end of the season. Of those 32, 31 or 96.9% were from the Power 5 conferences. Of the 32 teams who replaced them 13 or 40.6% were from the G5 conferences. This shows that the preseason poll is bias towards Power 5 schools, it takes several weeks for the G5 to break into the top 25 and only after they win multiple games and the teams ahead of them lose. The rankings should be done on the field during the games, there really should never be a week one upset because we do not know how good the teams are (looking back is USU’s victory over Washington State really an upset or was there bias against USU and for Washington State based on conferences and how they did last season?). With the transfer portal leading to a larger turnover in roster than ever before, the preseason poll creates a paradigm which is not correct and can take weeks to overcome (I want a time machine to go back 6 months and place $1000 for Michigan and USU to win their conferences). The preseason poll should end and the first poll should not take place until at least week 4. Despite what they say the committee uses the polls as a guideline, with so much riding on rankings in the polls, it is not fair for teams to start so far behind and play catchup during the season.