TCS Daily


In Defense of the BCS

By Dayn Perry - December 30, 2003 12:00 AM

Sports enthusiasts are well aware of -- and perhaps even indulged in -- the recent sturm und drang surrounding college football's final Bowl Championship Series (BCS) rankings. For the uninitiated, the BCS rankings are an amalgam of traditional team rankings -- voted on by writers and coaches -- computer power rankings, strength-of-schedule measures, etc. The upshot is that the two teams with the best BCS scores will play for the BCS National Championship in the designated bowl each year.

Heretofore, it's worked for the most part, albeit by happy accident in some seasons. In 2003, however, Oklahoma, LSU and Southern Cal all played major-conference competition, and all finished with one loss. How does one distinguish among a one-loss OU, a one-loss LSU and a one-loss USC? Not without lots of kvetching and hand-wringing.

 

Even so, the BCS rankings determined that Oklahoma and LSU were the best college football teams in all the land. Those two will meet on January 4th in New Orleans to determine the national championship. However, it was USC that ended the regular season atop both the Associated Press and ESPN/USA Today final polls. Unsurprisingly, fans and wonks alike took umbrage at the idea of the number-one ranked team in the country being shut out of the national title game. Just as unsurprisingly, they placed the blame on "the computers."

 

There's a strong neo-luddite strain running through many sports fans and most sports writers. This brand of resistance rears its head when there's any sort of reliance on any sort of algorithm to resolve any sort of sporting quandary. Most notably, the computerized component of the BCS formula.

 

A string of predictable columns followed the controversial final BCS rankings. Most of them lamented the fact that the computers have taken over. As Mike Lopresti of Gannett News Service huffed, "computers have logic, but no reason. They cannot sense the moment."

 

What Lopresti and other wailers fail to realize with regard to the BCS computer is that the output is only as good as the input. The BCS did what it was designed to do: determine discrete rankings based on the criteria supplied to it. That it did so inelegantly -- or even incorrectly -- is the fault of human contribution.

 

Since 2002, the BCS computer rankings, which are a mix of separate formulae, have been banned from considering margin of victory in their calculations. Putatively, this was to dissuade teams from running up the score on hopelessly overmatched opponents. In execution, this has robbed the BCS computer of a valuable analytical tool. Great teams, after all, don't just beat their opponents; they often crush them mercilessly. That may be an uncomfortable reality for sportsmanship alarmists, but it's how things are in sports. At the very least, point margin in losses should be a major consideration. As for point margin in victories, there's no reason a "law of diminishing returns" factor can't be incorporated to discourage grievous blowouts while still rewarding thorough beatings. In any event, the omission of margin-of-victory considerations is why USC didn't best LSU in the BCS rankings. This isn't some Kubrickian nightmare realized; this is the human element fouling up the process.

 

The obvious solution would be to correct these shortcomings and allow the computer rankings to make point margins part of the BCS calculus. However, that's not what you'll hear in most circles. Again from Lopresti: "Have a committee -- former coaches, conference commissioners, whomever -- name the two teams ... Would everybody be happy? No. But it would not be indefensible. Like now."

 

Right. Nothing sorts things out and foments change quite like the naming of a committee.

 

The problem isn't the fact that writers and administrators don't get to make bowl assignments by fiat. After all, it's those very ones who've come to the illogical conclusion that late-season games count more than early-season contests. And it's their fealty to pre-season polls that indelibly affects the final ones, with only partial regard for what unfolded on the field. Forgive me if I'm reluctant to grant total control to the talking heads whose views are fueled more by calendar and reputation than performance.

 

Give the computers the tools and resources they need to make the proper decisions and they'll do a better job than the voters ever could. Or just get a damn playoff system. Either way, the answer isn't a heavier dose of the human subjectivity that's held sway over college football for decades.

 

The author is a writer for FoxSports.com and Baseball Prospectus.
Categories:
|

TCS Daily Archives