"It's not a perfect system, and we'll continue to evaluate ways to improve it."
-- BCS Coordinator Roy Kramer (August, 1998)
True to those words, the BCS formula that Kramer helped create has since been tweaked, re-tweaked, demolished and reconstructed by himself and his peers. The never-ending quest to perfect this imperfect system has defined it almost as much as the controversy it has generated.
The latest incarnation of the BCS formula, however, has some sensible new features that were long overdue.
Finally, there is an accurate representation of the voting for the AP and coaches polls (by utilizing total points instead of the team's ranking). Finally, a statistically sound method is being used to evaluate the computer rankings (dropping the worst and best numbers before averaging). Finally, strength of schedule is not impacting four different elements of the formula that determines which teams play in the BCS title game.
The specific schedule strength component is now gone from the equation, as are the columns for losses and quality wins. And even though the computers are still hanging around, make no mistake: College football's championship selection process is back in the hands of the humans -- for better or worse.
It's interesting that the BCS began in 1998 with the premise that poll voters were too subjective to be given a majority voice in rating the top teams. But just six years later, the same system has now granted almost total autonomy to the polls.
That's close to a 180-degree philosophical turn. How could this have happened in such a short period of time?
Blame The Bowls
For starters, the AP and coaches polls have agreed upon the country's two best teams at the end of every regular season in the BCS era (as they have all but two years since 1980). So, whenever the BCS didn't match the polls with its championship game selection, the standings were met with much skepticism.
Ultimately, though, the key witness against the BCS in the court of public opinion has been the bowls themselves. Three times in the last six years, the top two teams in the polls have not faced each other in the championship game. All three times, both No. 1 and No. 2 were victorious in their separate games, giving tangible support to the argument that the polls were right and the BCS was wrong.
College football history shows that top-ranked teams don't always win their bowl games, which is why New Year's Day used to be so much fun. But within the last six seasons, the higher-ranked teams have won the big bowls and, in the process, transformed the poll voters from too subjective in 1998 to suddenly omnipotent in 2004.
There's no question that if USC had lost to Michigan in the most recent Rose Bowl, we wouldn't have seen wholesale changes to the BCS formula. If Miami had lost to Florida in the Sugar Bowl three years earlier, BCS administrators would not have implemented previous tweaks to the system. But those games ended with the opposite results, and the polls have returned to the spotlight because of it.
Eliminating computers, schedule strength and quality wins may indeed be part of the answer to the problems of the BCS, but putting the sport's championship system back in the hands of these voters is not the solution.
Six years of focusing on the shortcomings of the BCS have apparently made people forget about several age-old problems with the polls. The "media creating the news" dynamic with the AP voters is just the tip of the iceberg.
Another concern is the anonymity of the coaches' poll. Because most of the coaches in the national title hunt are also voters, there are obviously situations where one could benefit by ranking his team higher and a competitor a little lower. And because the coaches' ballots are kept secret -- even from the BCS folks -- there is no way for the public to be sure the poll hasn't been tainted by personal agendas. In fact, there's no way to know for certain that the coaches actually filled out their own ballots.
Aside from possible agendas, there are also inherent biases in the polls. Factors such as marquee names (of teams and players) and timing of losses routinely influence voting patterns, even if totally subconscious on the part of the voter. Also, preseason polls, which are heavily based upon tradition, returning starters and results from the previous year, always give some teams a significant advantage over others before a game is ever played.
But this is not news. The BCS leadership is well aware of these issues with the polls and discussed most of them at length before deciding to hand voters the keys to college football's most expensive car.
"I think we've tried to be up front in saying that it's not a perfect system," said new BCS coordinator Kevin Weiberg last month. "But given the parameters we're working with, we think it's the best system that we have available to us."
There's A Better Way
Maybe it's not. It only seems to be the best system available, because few people recognize the biggest inadequacy of the polls today.
The flaw of the newest BCS formula isn't that it puts the power back into the hands of the people. The flaw is that it puts the power in the hands of the wrong people. Whether anyone wants to acknowledge it or not, today's voters aren't informed enough to truly know which teams are the best, and that's a problem.
You see, the media and coaches polls originated before the era of sports television, so voters back then could only watch teams by being at the games. They evaluated the rest of the bunch by looking at scores and hearing second-hand reports. For the most part, national championships were being decided by people who had never seen the majority of the teams play.
Well, the stage has changed, but the song remains the same.
Everyone is playing on TV now, but the polls still basically operate in the same archaic fashion. On a given Saturday, about 90 percent of the voters are in attendance at a game -- either coaching in it or reporting on it.
Sure, most of the media are able to catch portions of a few other games on TV in the press box, but the focus of their day is still on two teams. Sure, the majority of the voters are looking at box scores, reading the recaps and watching the highlights, but it's just not the same as seeing the game itself.
The top nine teams in last year's final BCS Standings played a combined 112 games during the season; 107 were televised live, and the other five were on tape delay. From my own seat here at ESPN, I watched almost every single one of them.
So, if the technology exists, and someone like me can see all the top teams play every week, then why are the people who hold the ultimate power to judge these teams not watching all the games? That's like your local weatherman looking out the window to decide his forecast rather than checking the Doppler radar.
It's time for college football's championship system to enter the 21st century. Polls are a great tool for generating discussion about college football, but they have outlived their usefulness in the national championship framework.
Players and coaches bust their tails for most of the year, and if they aren't going to be allowed to decide their championship on the field with a playoff, they should at least have their fates in the hands of people who are as informed as they can possibly be.
There are always enough great football minds not working for a school or media outlet to put together an excellent selection committee. Bring them to ESPN, send them tapes of the top teams or whatever. The BCS folks can handle the details.
Let the players play. Let the coaches coach. Let the media write and talk. Just let someone who's watching all the games decide which teams are the best in college football!
Brad Edwards is a college football researcher at ESPN. His Road to the BCS appears weekly during the season.