Winning isn’t everything for the Top 25, is it?
Part 1 of 6: The Top 25 Unplugged
It is in our human nature as football fans to be emotional and biased. We also tend to gravitate to information that is easily understood and meets our expectations. As Americans, little is more revered than a winner in sports and winning is placed above all other factors in determining the best teams or players. Is it not inconceivable that the humans that vote and the humans that maintain their computer rankings skew their judgement in favor of games won and lost?
I have maintained for many years that the BCS rankings and ultimately the entire ranking system are based on two things. Total wins and the perception of the power ranking of each conference by the voters and the non-transparent computer systems that provide their rankings. SOS (Strength-of-Schedule)...pfff whatever. Boise State and a handful of non-AQs have proven that individual SOS is meaningless. To challenge any question, if you have the data, then test it. I did and here are the results.
The 1st Challenge Question:
Is it really a ‘Winning is Everything’ system?
I was asked last year to provide a ‘top 25′ from the CFBMatrix for a national publication comparison of all the pre-season predictions. I had never done one before as I viewed it as a tool only to create arguments and intrigue early in the college football season. The only important ranking is the top 2 for the National Championship Game and the rest is for marketing of the conference, possibly recruiting and fan bragging rights.
So I ranked them by what I saw was the best overall teams, coaches and SOS. I did not include Boise State’s and TCU’s as they played a combined 1 AQ team and had schedules that would project most AQ schools finishing with 10 or more wins. What I quickly learned, once published, is that if you do not go mainstream with whatever the media and fans want in a top 25, you are wrong and disregarded.
Scoreboard Rules and Herd Behavior
Most people do not like to be excluded or put down in their way of thinking, especially if it is outside the box. It is no different for coaches or SIDs (Sports Information Director) doing poll voting or math wiz’s and their adjustable ranking models. They want to fit in and cause no controversy. For football rankings, people understand two notions. The first is whatever the masses say in our democratic society, that is the best and right path. If ESPN says an undefeated Hawaii team or a one loss TCU team that plays nobody is a top 15 team, then it is a top 15 team. The second idea of sports in general is the scoreboard. There will never be a good SOS comparison system and most of us cannot sit down and watch film to dissect individual teams. If team A has 11 wins and team B has 10, then by golly team A must be better. And that is exactly how the BCS ranking system has worked.
Average Wins/Loses Per Ranking
This is part is really simple. The 12 game schedules started in 2006. Going back to that point gives me 150 total top 25 rankings over 6 years. I took all the win/loss totals of each team for every BCS season end ranking spot and averaged them out. If you drew a chart of what you would expect if you believed the system was a winning is king I would would just like the one below. The blue and red lines are the actual average wins and loses per each team ranking spot over the last 6 years. The black lines are the actual trend lines. Folks, it doesn’t get much more clear than this chart.
The chart to the right gives you a breakdown that shows you the average numbers of wins and loses for each team at the corresponding year end BCS national ranking since 2006. In nearly every case the lower the rank, the less average wins and more average loses. You want your team to get a NCG berth then win at least 11 games. You want to see a top 10 ranking, then win 10 games or more. You want a top 25 rank just win more than 9 games and be in the right conference. 8 wins usually isn’t going to get you into the top 25 if you are in the wrong conference or a non-AQ team.
There is no doubt that the BCS composite rankings favor win totals first and foremost. The polls and the computers. Period. End of the chapter but not the story. Not by a long shot. With it clear that the BCS ranking system goes by total wins it begs the next questions to be answered:
Chapter 2: Which is more total win biased The AP Poll or BCS? or Have we improved or regressed in the Top 25?
- Dave Bartoo, The CFBMatrix & your Anti-Homer Follow @cfbmatrix