Did BCS Do the Right Thing With Changes For 2001 Season?
by Dave Congrove
No, I am not the self-appointed top expert on college football, nor do I claim to be the foremost mathematician in the world. But I did stay at a Holiday Inn Express last night.
As I am about to begin the ninth season of conducting the Congrove Computer Rankings, I read with great interest the changes implemented in the Bowl Championship Series formula. The bottom line is that the Dunkel and New York Times ratings were dropped, and ratings by Peter Wolfe and Wes Colley were added. The BCS is also establishing a quality win component for teams that defeat a team ranked in the cumulative top 15.
Reportedly, all of the computer rankers were asked to lessen the impact of wide victory margins as there was a widely-held belief that running up the score and embarrassing weaker opponents could enhance a teams ranking. Dunkel has admitted it refused. I can only assume the Times also refused. So, in essence, they dumped themselves.
I am trying to maintain a positive outlook that these changes will be good, but wonder if the way in which they were handled was appropriate. Should the BCS provide directives to the participating rankers on how to conduct their polls? Does doing so improve anything or render them useless by homogenizing the results?
Two years ago, when the BCS increased the number of computer rankings services to eight, the goal seemed to be to invite as many differing opinions as possible and average out the results. They seemed interested in finding the most reliable rankings while keeping an open mind that there are numerous ways to conduct them.
Do they now want everyone to pretty much do the same thing? I hope, upon further review, this proves not to be the case and that the results at the end of the year will provide evidence that these changes were brilliant. Though I am concerned this will not be the case.
Most people will argue that better solutions are derived by inviting unique and original approaches to an idea.
Certainly, the BCS should review and include input from sources who appear to be on the correct paths, not necessarily the same path. At the end of the year, each source should be held accountable for their own results and subject to review based on their own merits. The BCS should not suggest a single thought as to how every ranker should be altered. Any attempt to manipulate them toward a single result weakens the credibility of the participanting rankers and the BCS.
The people who conduct these computer rankers are inherently interested in having their results make sense. Some, like the expelled Dunkel Index, have been around for decades. I may not agree with their formula but I can objectively respect that the unique Dunkel approach has its merits. There is more than one way to skin a cat, so I can fully understand their decision to remain true to their time-tested formula and not honor the BCS request to tinker with it.
I began the Congrove Computer Rankings as a hobby because I believed the polls were too easily subjected to voter bias.
Sportswriters, for instance, generally cover a specific team or conference. I wondered how someone who covers the Big Ten really knows all that much about the Big East or PAC-10.
I also asked myself if I believe that coaches actually sit down right after their own game and research other games. How much time do they truly have to gather and weigh evidence before submitting their findings? Can their ballot truly be impartial?
Those opportunities for gross inaccuracies in the voted polls led me to toil long and hard to develop and fine tune a ranking system that can be both accurate and unbiased. I am comfortable with the result. Can it be improved? No system can be perfect, but it has generally proven by its results to be among the most reliable sources available. I strive to find the ultimate formula that will be guaranteed to reward the most worthy teams (and allow me to take over the world!).
I check and re-check the formula all the time. I think and re-think the possibilities of how it should work all the time. But only twice in nine years have I found a true need to change anything about it. And those changes were minor. The most important factors have been there all along. You must win your games. A tougher schedule provides a better chance to be rewarded. Running up the score is about as meaningless as can be rendered.
I undertake this time-consuming endeavor with great care and concern because I want credible results, regardless of the fact that the BCS does not use my rankings. I imagine my cohorts are just as obsessive.
Naturally, I firmly believe the BCS Standings would be far more accurate if the Congrove Computer Rankings were included. To my knowledge, my system was not reviewed for possible inclusion prior to these latest changes. That does not bother me. There is always next year.
It wont stop me from attempting to create the most reliable poll that I can imagine. Anyone who can imagine a different way to do it should be encouraged to do it themselves.
Whining About Whining About the BCS
Lack of Quality Leadership Opens the BCS To Annual Controversy
Suggested Changes For BCS
BCS Bashing Has Big Bandwagon
Computer Rankings and National Titles