Polls say Tigers are better, but are they?| by Carl Dubois on Oct 1, 2009 at 6:50 am |  | |
If I were in charge, you'd have two main reasons to look forward to this weekend. The first, the LSU football game Saturday at Georgia, is already atop your list, so my work there would be done.
Second? The next day, we'd finally get a look at the first college football polls of the season.
No more preseason polls, at least not from the USA Today and Harris voters, who comprise two-thirds of the BCS formula. They wouldn't get to vote until the first Sunday in October.
No more September polls.
Every year, September rankings are largely based on what happened the year before, the number of starters a team has left from the previous year's team, the promise (or lack) of impact freshmen and a hunch or five. Tradition plays a big role.
So, when the No. 9-ranked team loses to an unranked team in Week 2, we call it an "upset." Never mind that by season's end, we just might realize the unranked team was better than the No. 9-ranked team from start to finish.
LSU (4-0) is ranked No. 4. Georgia (3-1) is No. 14 in the USA Today coaches poll, No. 17 in the Harris poll and No. 18 in the AP poll.
If Georgia wins Saturday, will the Bulldogs move ahead of LSU in the rankings? I doubt it. There's too much of a gap. Yet, they'd both be 4-1, with Georgia winning the head-to-head matchup.
Georgia is a slight favorite. Who's to say the Bulldogs shouldn't already be ranked ahead of LSU? Conventional poll thinking says no, but conventional poll thinking isn't always right.
Conventional poll thinking says undefeated teams are better than one-loss teams, which are better than two-loss teams, and so forth. In September, what separates the undefeated teams is perception.
I don't think undefeated teams are automatically better than one-loss teams. I don't think this undefeated LSU team is better than Georgia right now.
I'm on record as being anti-poll in most circumstances, and especially before October. After the 2003 football season, which ended with LSU as the BCS national champion and USC as the AP national champion, I wrote a column saying the AP needed to get out of the championship business.
If the AP didn't disband its poll, I said, it should at least take steps to remove its poll from the BCS formula. The AP and many of its writers and broadcasters were covering news they were helping to create, and that was wrong.
My boss got a call from AP. He told me AP assured him it would continue with business as usual.
But a funny thing happened before the 2005 season: The AP, acting as if it suddenly realized the BCS had been using its poll in the BCS standings formula for years, sent a cease-and-desist letter basically saying, "Whoa. You're using our poll? Shame on you! Stop that now or we'll have to take legal action."
So, the BCS removed the AP poll and replaced it with a new poll, the Harris.
I'm going to be ahead of my time again and tell you eventually wisdom will prevail, and the BCS won't allow the USA Today coaches poll or the Harris poll to cast ballots until October. Sure, there aren't any BCS standings until October, a concession to the waiting I (and others) propose, but those standings are heavily influenced by the September polls, which are heavily influenced by the preseason polls.
Meanwhile, the computers that combine to represent the other one-third component of the BCS formula are clicking and crunching what happened in the first month of the season, so when you slap their output together with early October polls that are skewed for all the wrong reasons, you get bogus results.
Some who vote in preseason polls say they are ranking the teams where they think they should be to start the season. Others say they are predicting where they'll finish the season.
Most in the latter group usually defend their picks by saying things like, "Well, so-and-so has four tough road games, and they're going to lose at least three of them, so I've got them no higher than No. 20."
The problem with that is, if so-and-so doesn't lose that many, the same voters often don't go back and reverse-handicap the way they voted in the preseason. They simply "slot" their votes and bump the team up a notch or two at the time of those "upsets," never accounting for the ball and chain they tied to the team's wings in the preseason poll.
I should point out I know voters who do their due diligence. One, who works for a major newspaper, didn't drop Cal nearly as far this week as other voters did. He says he won't raise Cal as much as others might should Cal rebound with a strong season, because he never punished them severely in the first place.
He's trusting the work he did in the offseason, when he studied Cal (and other teams). He's not allowing himself to be swayed too much by one bad weekend (or, if it happens, one great weekend).
I know polls are supposed to be fun conversation pieces. When they began in the 1930s, they were a way for the AP to generate interest in college football. The polls began roughly 3-4 games into the season, but soon preseason polls became fashionable, and the college football world never looked back.
We need a glance in the rear-view mirror now.
Back then there weren't a handful of preseason magazines hitting newstands in May and June. There wasn't an Internet filled with enough statistical date, feature stories and talking points to keep the average person busy for the entire offseason. There weren't television channels specifically devoted to college sports -- and there weren't shows that talked daily about football during the preseason.
Generating preseason interest in college football? Uh, that job's been redefined, magnified, multiplied and upgraded far beyond anything polls accomplished in the 1930s, '40s, '50s and later.
The reason for their existence 50 years ago is a quaint piece of nostalgia today, not a reality. They're simply not needed, and worse, they are often a misguided influence upon regular-season polls that shape a system that is now Big Business.
Preseason polls have outgrown their usefulness, and college football has outgrown the "Hey, let's get people to pay attention to college football" mentality that led to their creation.
Does LSU owe its No. 4 ranking largely to preseason polls? You betcha. The Tigers are ranked on potential, and with the help of losses by other teams who were also ranked on potential.
Would a loss at Georgia this weekend prompt voters to drop the Tigers in the rankings? Of course.
Would LSU's ranking at that point be any different than it would have been if there were no preseason and September polls? I can't say for sure.
What I can say is a flawed system would be much better, year after year, without any polls until October. One day it will happen. I'm convinced of it.
A reporter asked Georgia coach Mark Richt if it made sense for LSU to be ranked so much higher than Georgia, yet be the underdog for this game.
“Who’s favored? We’re favored?” he said. ”I didn’t know we’re favored. I guess it’s just the home field –- I guess [that] is what they’re thinking. I think they think both teams are good football teams.”
Or, the reporter speculated, given that four top-10 teams lost last week, maybe there’s not much difference between No. 4 and No. 18 anyway?
“This time of year, you don’t know,” Richt said.
I agree. Rankings before October have little credibility, and Richt knows it. Read his other comments in the reporter's blog.
“We’ll find out Saturday," Richt said. "We’ll find out Saturday.”
I agree.
.
Carl Dubois has written or blogged about LSU sports since 1999. He's long been a conscientious objector with respect to voting on polls, and he doesn't much care for non-sports Internet opinion polls either. He will talk more about the LSU-Georgia matchup Friday. You can contact him at carl1061 'at' gmail.com.
|
RSS Feed