Wolfe, Colley, Sagarin, Billingsley, Massey, Anderson, and Hester is not a law firm. They are the seven gentlemen who provide computer analyses to the Bowl Championship Series. Just how relevant are they?
Google is the best invention since Ben Franklin organized the public library concept.
Only a nerd would say that.
I am a nerd. I’ve been a nerd my entire life, although I have sacrificed time, dignity, and money running away from my nerdism. More about that later.
In the seventh grade, I built a small analog computer. I also won the Golden Horseshoe, one of West Virginia’s highest honors the state can bestow on a junior high student. My eighth grade was highlighted by my award-winning science project on the analysis of highway bridge design.
However, nerdism can only take you so far. By ninth grade, no one liked me.
I had to think of something quickly. I was not a particularly handsome kid. I didn’t have a personality. My music talents were non-existent. I couldn’t dance. I had never played sports. Therefore, convoluted logic led me to believe that sports seemed to be the obvious path to being cool.
My hand-eye coordination left a lot to be desired. That ruled out everything that had to do with a ball. There was but one route: football.
Think about it: A) with the headgear and all the padding, football can make you rather anonymous, B) you’re even more anonymous if your position is on the line, and C) with my aforementioned dropsies, the line it was.
I’m 5'10" and played at 160. I didn’t like to hit. What the hell was I doing out there, one may ask?
I answered during my last year of high school ball by gaining fifteen pounds of muscle and digging deep to find my sociopathic side.
Despite the sick assassin way I used to quietly delight in pancaking unsuspecting maniacal linebackers, no colleges called.
Now that that’s over, I’m back to being a nerd, with my BS in mechanical engineering magna nerd laude and my engineering job.
Continue to read after this note if you wish to learn more about the statistical methods the computer folks use. Otherwise, just accept the following to be true:
Computers are necessary for the Bowl Championship Series in its current form. They provide the objective analysis that cannot be found with coaches and writers.
If the Rose Bowl accepted money they want, which is probably in the amount it took to beat the Soviets and win the Space Race, thereby allowing a playoff to be instituted, it’s my opinion that computer analysis would be absolutely essential.
However, the BCS is flawed and the chances of a playoff are those of the earth’s velocity of rotation accelerating. Then Peter Wolfe, Wesley Colley, Jeff Sagarin, Richard Billingsley, Kenneth Massey, Jeff Anderson, and Chris Hester would deserve the same respect as the coaches and the writers and award their own mythical yet valid national titles
Ranking college football teams is an enormous statistical problem, a moving target of almost infinite variables that takes a computer to get its hands around. Unfortunately, the computer can do only what a human tells it to do. Hence, the term “garbage in, garbage out.”
To understand a computer system’s rankings, you must first have a feel for the data being entered. For example, the Bowl Championship Series requires that all computer ranking services used for the one-third total input consider only win/loss and completely disregard the score margin.
Other data may be considered. Ken Massey takes the game’s venue into account, whereas Wesley Colley discounts the home/neutral/away factor, since no mathematical method of measuring has satisfied his criteria. Interestingly, Mr. Massey works the date of the game into his system as well.
All six services factor in the strength of schedule, but Richard Billingsley also statistically looks into strength of conference schedules, in particular the strength of the out-of-conference slate and the conference’s OOC won/loss record.
Similar to the BCS itself, Jeff Anderson and Chris Hester do not publish any rankings, so there is no need for corrections to completely subjective preseason ratings.
And, finally, if I’ve kept you this long, Peter Wolfe, Ken Massey, and Jeff Sagarin delve into advanced statistics, advanced like your math professors love to love. Mr. Wolfe and Mr. Massey use a maximum likelihood estimate.
This means (with help from Wikipedia) you can measure a sample population of heights of your acquaintances, for example, and see how closely your distribution approaches that of the complete population of heights of everyone in the country. That is just about as well as I can explain it without using words like “asymptotics.”
In addition, you can go into the Mariana Trench of pure mathematics by adjusting your population with regards to the complete population, using a correction based on Bayesian probability.
That gave the nerd in me a headache, but let me put myself out of my misery and use one sentence to describe the main idea of most anything Bayesian.
Here goes: the standard Bayesian correction (again using Wikipedia) considers probability of an uncertain order, like ranking college football teams, and adjusts it based on the new data, like the information of the week’s games the computer people get early Sunday morning.
That wasn’t so bad.
Just remember: As the BCS is presently structured, computers are necessary, adding objectivity to our human foibles.
And, that’s less than 1,000 words.