A Better Football Poll?

Posted on October 16, 2007 by


I posted the following at The Sports Economist. In addition to re-posting this column here, I have also added below the first Wages of Wins College Football Ranking.  Hope you enjoy.

The BCS has finally made its appearance and we now “know” that Ohio State is the number one team in the land.

Okay, we really don’t “know”. The BCS combines two things people don’t like: Computer polls and Human polls. Hostility towards computers picking teams is pretty widespread. And it turns out, some people aren’t too thrilled with the human polls either. Here is what Todd McShay of ESPN The Magazine (October 22, p. 98) had to say about the coaches and media who vote in the polls:

“… voters are right now irresponsible. Coaches can’t pay enough attention and end up having grad assistants vote. Media members have to focus on the teams they cover. It’s just insane that the NCAA doesn’t have a paid group of former coaches and players whose full-time job is to watch games, then rank the teams.”

What McShay is asking for sounds quite similar to the Harris Interactive Poll. The big difference is that it doesn’t appear that the NCAA is paying for the Harris poll. So would the world be a better place if the NCAA paid for a poll?

Let’s think about this in terms of costs and benefits. McShay wants the NCAA to hire a group of former coaches and players to watch college football games. Let’s imagine the NCAA pays each coach or player $10,000 per year to do this job (which is a very low wage for a full-time occupation). If 100 coaches and players are hired, the NCAA poll is going to cost $1,000,000. That seems like a pretty substantial cost.

Now what are the benefits? Presumably we are now going to have a poll that is more “legitimate.” Experts are now going to focus their attention on watching all the games. And consequently, we are now going to “know” who the best team is in the land.

But is this really true? Here are just a few of the problems with this suggestion.

1. Even if these people are paid to watch football all Saturday, they cannot possibly watch all the teams. So every voter is still going to have gaps in their knowledge.

2. And even if you could watch all the games (which you can’t), the sample of games is too small to determine the best team. For example, who is the better team, South Florida or Ohio State? Voters now say Ohio State. Computers say South Florida. Who’s right? Right now Ohio State has played seven games while South Florida has played six. But these games were against completely different opponents. So not only is each sample too small to tell us much, we don’t even have two samples that are comparable.

3. Beyond these technical problems is an economic issue. Costs for a firm are justified because they lead, either directly or indirectly, to a revenue stream. But how does having McShay’s poll enhance the revenues of the NCAA? Will more people go to the games because they believe in McShay’s poll? Will this lead to more television viewers? One could even argue that having a poll that is less controversial would actually reduce the attention paid to the NCAA, and therefore reduce revenues.

In the end, I think it would be “insane” for the NCAA to follow McShay’s suggestion. In fact, the decision to let people provide free polls to the NCAA seems perfectly rational and reasonable. The current arrangement increases attention paid to the sport and hence boost revenues. And with a cost of zero, every member of the NCAA has to love these polls.

The Wages of Wins College Football Ranking

Explaining why the NCAA shouldn’t pay for a poll doesn’t help people who want “a better poll.”  To help these people, I have decided to use my vast analytical skills to create the first Wages of Wins College Football Ranking.  Many minutes of hard work went into this ranking and I think it will settle once and for all the identity of the top teams in college football. Let’s start with number one.

1. Colorado State Rams

I know Colorado State is the obvious choice for number one, but let me explain a bit of the reasoning behind this choice for the few people who “don’t get it.” 

To date the Rams are the best team I have ever seen in person.  Okay, I haven’t seen the Rams play in person since the 1990s.  But I can’t imagine they look too much different.  In fact, I saw them on TV earlier in the year and they looked pretty similar to the Rams I saw ten years ago.  At least, the uniforms were very familiar.

Now some might note that the Rams have yet to win a football game in 2007.  And some people out there might think a failure to win any games means CSU can’t be the number one team in the land.  But let me cite the wisdom of William Ford.  Ford has owned the Detroit Lions since the 1960s, so he has more than four decades of experience in football.  When asked about the fans reactions to the Lions inability to win games a few years ago, Ford replied (and I don’t have a link to this, but I am pretty sure this is an accurate quote): “I can understand why fans are frustrated.  Fans tend to focus on wins and losses.”

What does this statement mean?  We can’t evaluate a team in terms of wins and losses.  So people who think the Rams winless record tells us they can’t be number one are focusing on the wrong issue.

What should we focus on? I often hear commentators in football describe a great player as follows:  “What’s great about this guy is that he’s a football player.  That’s all you need to know. This guy is just a football player.”  See, that’s what you want on your team.  Football players.  Now, as noted, I have seen the Rams on TV this year.  And I can assure you, the Colorado State team is comprised of football players.  All the players I saw had helmets and were wearing pads.  And whenever I see someone in pads and a helmet I always think “that’s a football player.” The Rams have an entire team of such players, so they must be the best. 

2. Michigan Wolverines

Michigan is currently 5-2, having lost to Appalachian State and Oregon.  But those were the first two games of the year.  Both teams played a spread offense, which I don’t think should count.  Real teams don’t play gimmicky offenses. I would add that I think Michigan thought these first two games were part of its exhibition season.  So I am not sure Michigan played its best players or that its players even tried that hard.

To understand why Michigan is ranked so high you need to remember that sports is all about “what have you done for me lately.” That’s the key. And lately Michigan is perfect.  Michigan has won five consecutive games.  One of these victories was eve over Notre Dame. 

To see why this is impressive, consider the next two questions:

Who has won the most games in college football history? Michigan

Who has won the second most games in college football history? Notre Dame

So there you go.  Michigan beat the second best team in the entire history of the game. And that means Michigan must be pretty good (not Colorado State good, but still, right up there). 

3. Nebraska Cornhuskers

This past week Nebraska got beat by Oklahoma State.  In fact, they got slaughtered.  But I think OSU was playing Barry Sanders, and that would be unfair.  Plus, losing to OSU was just a ploy.  We all know that Nebraska was trying to get Tom Osborne back into the program.  And to do this, the team needed to lose big to a team that had no chance to beat the Cornhuskers.

Of course, the plan worked to perfection.  After Nebraska lost, the Athletic Director (Steve Pederson) that fired Frank Solich (for winning just nine games in a season) was also fired.   This led Osborne to take over the entire athletic department. 

Given this plan, we need to throw out all the games the Cornhuskers failed to win this year.  And when we do that, Nebraska has a perfect record in 2007. 

Well, I could go on.  But I think those are the only teams I need to rank.  Some might note that Stacey, Marty, and I went to school at Colorado State.  And people might also note that I was born in Detroit, Michigan and my family moved to Lincoln, Nebraska when I was eleven. Although all that is true, Colorado State, Michigan, and Nebraska are not on top because these are the teams I follow. They are on top because the numbers and logic clearly lead to that conclusion.

Hopefully this little exercise has cleared up any confusion people have about who the best teams are in college football.  I could repeat this feature at the WoW Journal each week, but I can tell you that after several years the top three teams in my poll have never changed.  The logic and numbers change, but the results are constant.  And that’s why my college football model is the best.  The past perfectly predict the future.  Wow, that’s a great model.

– DJ

Posted in: College Sports