Analyzing the BCS: A Treatise for Reason

Use your ← → (arrow) keys to browse more stories
Analyzing the BCS: A Treatise for Reason
(Photo by Chris Graythen/Getty Images for Under Armour)

It pains me to hear people talk about the Bowl Championship Series. While it may begin (like many sports debates) with good intentions, it always ends up sounding like technology is out to get fans. 

Blame the supercomputer for your team not going to the Sun Microsystems Tangerine Fiesta Bowl; the system must be flawed. 

This misconception is fundamentally untrue. The computer system in the BCS is not to blame—it's the pollsters. 

I know why the BCS creators did it. Statistical inference is not good enough for most people. Most people do not understand the system, so how can they trust it?

But a bunch of journalists blathering their opinion about which giant school is the best—that they understand. 

America the beautiful.

We did not want some end-all computer system telling us who the best is. How are the Mac and PC guys going to know more about football than we do? How can a computer evaluate a team? 

Very easily, in fact. 

Is it not conceivable to think that a computer could pick out the teams who beat the best teams? Or the conferences that have the most teams who beat the best teams? Or the teams who have had the most success against successful teams on the road?

Certainly not. 

The problem is that there are not only two distinctions in Division I college football anymore; there are three: the FCS (Div. I-AA), the BCS conference schools, and everyone else.

No one believes that if Alabama goes undefeated, and Utah does the same, Bama should not get the nod regardless of the fact they played three FCS schools last season.  The “computers,” which rank the schools based on objective results, see the field as one big pool. 

Here is a quick crash course in the computer poll system (for further reading, here is the list of sites of the six different computer systems and their websites):

“Six computer rankings will be used: Jeff Sagarin, Anderson & Hester, Richard Billingsley, Colley Matrix, Kenneth Massey, and Dr. Peter Wolfe. Points will be assigned in inverse order of ranking from 1-25. A team's highest and lowest computer ranking will be discarded in calculating its computer rankings average. The four remaining computer scores will be averaged, and the total will be calculated as a percentage of 100.”—bcsfootball.org

I went to all of the different websites and threw the final rankings into Excel:

 

Sagarin

A & H

Billingsley

Colley

Massey

Dr. Wolfe

1

Utah

Utah

Florida

Florida

Utah

Utah

2

Florida

Florida

USC

Texas

Florida

Florida

3

Texas

Texas

Utah

Utah

OU

Texas

4

OU

OU

Texas

USC

Texas

USC

5

USC

USC

OU

OU

USC

OU

 

Using the same system that the BCS uses, I knocked out the top and bottom scores and ranked all the schools, giving them points based on where they finished in the CPU polls.

 

High

Low

overall score

Average

BCS rank

Utah

1

3

98

24.5

1

Florida

1

2

97

24.25

2

Texas

2

4

91

22.75

3

USC

2

5

86

21.5

T4

OU

3

5

86

21.5

T4

 

Utah finishing first?! Alert Congress! USC and Texas finishing above OU, who made it to the national championship game?    

This does not look at all like the final standings the BCS released after the season.

 

AP

Coaches

Harris

Final BCS

1

Florida

Florida

Florida

OU?

2

Utah

USC

OU

Florida

3

USC

Texas

Texas

Texas

4

Texas

Utah

Alabama

Alabama

5

OU

OU

USC

USC

6

-

-

Penn St.

Utah!!!

7

-

-

Utah!!!

-

 

This is where human error comes in. 

By allowing a weighted scheme involving the human-created polls to skew the results, the BCS is irreparably tainted.

The AP, USA Today (Coaches), and Harris polls are all factored into the equation as well. Teams are given “points” based on how many possible voters vote for them and at what ranking.    

When you allow media members and coaches who can’t possibly see every game to vote on who the best teams are, or have something to gain by voting a certain way or another, you allow a level of bias that throws the system completely off kilter. 

Some examples:

 

The Harris Poll

In the final poll, one Harris poll voter had Rice ranked 18th in the nation.

Ball State had votes placing them as high as 10th, as well as many voters leaving them completely out of the top 25. 

Two voters left BYU out of the top 25, two more Georgia; one other left the Pitt Panthers out of the top 25. 

 

The Coaches Poll

Out of 61 voters, only one did not vote for Florida. That vote went to Utah. I wonder which coach that could have been...

Houston, Connecticut, and Kentucky all received votes to be in the Top 25; I wonder who those coaches could have been...

 

The AP top 25

Rice, Arizona, and Kansas all received votes to be in the top 25. 

Sixteen votes put Utah at No. 1, and one vote went to USC to finish No. 1 in the country; all others went to Florida. 

 

It is time to put an end to this ridiculous system. 

Maybe a playoff system will never work because too many “traditionalists” will shoot it down over the fear that they will lose money (unlike every other competitive sport that has playoffs and makes considerably more money because of them). 

If we are going to make strides in remodeling the BCS, the first step should be to take out the human element and rank the schools using unbiased mathematical inference that does not see school size, tradition, or ridiculous facilities. 

Let the BCS work by getting out of its way.

 

An afterthought: ESPN's final version of the BCS looks different than mine; I am not sure why. I got all my numbers from the individual websites of all six computer-based statistical systems. You are welcome to do the same by following the link I posted. 

Load More Stories

Follow B/R on Facebook

Pac-12 Football

Subscribe Now

We will never share your email address

Thanks for signing up.