Three years ago I called for the removal of college football preseason polls, at least for those polls that factored into the BCS standings. My reasoning was that voters are loath to reconsider their preseason votes, making it very difficult for teams to move up in the polls without the teams in front of them losing. Again, that was three years ago, when I had little influence as a sportswriter. Now, rich and famous, I can facilitate real change. And I’ve decided I want to take it a step further and eliminate polls until the regular season is complete. When the playoff arrives in two years, the powers-that-be shouldn’t force—or even allow—the selection committee to release polls throughout the season.
For ranking purposes, each season should be judged in a vacuum. Media outlets that want to release preseason polls are free to consider how teams played the previous year—they should, in fact, if they want the most accurate rankings—but polls that help determine which teams play for the national championship should not. Why? Because coaches—whose votes count towards the BCS—are stubborn and biased. Don’t believe me? Then surely you will trust SCIENTISTS when they say that coaches overrate their own teams, teams from their conferences, and teams they played against.
We saw the conference favoritism last season. In the debate over No. 2 between Alabama and Oklahoma State, the SEC coaches backed ’Bama and the Big 12 coaches chose the Cowboys. This year, like all others, voters are stubborn and let their preseason expectations dictate how they vote throughout the season. Coaches rarely move a team ahead of another until the higher-ranked team loses.
An analysis of this week’s Coaches’ Poll shows that most teams, based on their preseason ranking, remain ahead of all teams that were ranked lower in the preseason and have the same number of losses. For the handful of teams that aren’t—like Clemson, for example, which was ranked higher than Florida in the preseason but is now behind the Gators despite both teams having one loss—it is almost always the case that they lost before the team that leap-frogged them. It is another example of stubbornness—coaches stick with their guns until a team loses, then keep their new rankings until the new “better” team loses.
This season, the polls are obviously critical to the national title hopes of the three remaining unbeatens: Oregon, Kansas State, and Notre Dame, which are ranked 1-3 in that order in the Coaches’ Poll. Oregon began the year ranked fifth, Kansas State was 21, and Notre Dame was 24. At no point has there been any change in that order. If all teams win out, it is reasonable to expect they will remain in that order, with Notre Dame left out of the national championship game.
I’m not saying this is unfair. Nor am I saying that Notre Dame would be any more deserving than Oregon or Kansas State. But the voters are unlikely to do any maneuvering at this point; instead of assessing the season at the end, they will stick with what they thought at the start. In every season since the BCS reconfigured its format in 2004, the teams ranked 1-2 in the final pre-bowl polls were also 1-2 in the BCS standings. (This was the case in all but two seasons since the BCS started in 1998.) The computers don’t mean much of anything.
Recalling that scientific data from before, Pac-12 coaches (there are six who vote) are going to support Oregon and Big 12 coaches (four) will get behind Kansas State. As my writer friend David Vranicar tweeted after the latest Coaches’ Poll was released, “Notre Dame could finally get burned for not having a conference.” Currently, the gap between Oregon and Kansas State is about as half as small as the gap between K-State and Notre Dame.
It’s very unlikely, but Irish fans looking for hope outside of an on-field upset can refer to 2006. That’s when one-loss Florida jumped one-loss Michigan in the final pre-bowl poll even though Michigan did not lose that week. What makes that situation unique is that Michigan had already lost to the No. 1 team, Ohio State, and had not played in two weeks. Florida won the SEC Championship and the voters bumped the Gators to No. 2. Notre Dame won’t have that luxury. Obviously none of the teams in question have played each other. And Notre Dame’s season ends on Nov. 24. Assuming Oregon wins out, it will play the following Friday in the Pac-12 Championship; Kansas State plays on Dec. 1.
Of course, a voter flip-flop would be much more of an upset than one of these teams losing.
7 thoughts on “Should College Football Get Rid of In-Season Polls?”
Good stuff as always. A few nits:
1. I’m not a huge fan of pre-season and in-season polls prior to week 5/6, but how else would you determine that one team beat another really good team?
2. An ND conference affiliation would help with voting, but ND in a conference would surely have a far worse strength of schedule
3. The likelihood of all 3 teams remaining unbeaten is under 10% going in to this weekend; not a lot to worry about.
Correction – Odds that all 3 remain unbeaten is 14%. Exciting games to come!
Look at this guy with his fancy “percentages”! But you’re probably right — I still think one or more of the unbeatens will lose. Of course if there is only one unbeaten that will start a whole new debate, but at least a team that won all of its games wouldn’t be left out. As for the polls, media outlets would still release them, so the casual fan would still be able to get a sense of which teams are “good.” Thanks for the comment, Beach!
If all 3 unbeatens lose, we could potentially see an Alabama/Georgia vs. Florida NC game. Whoa, more SEC shoved down our throats!
I don’t want to think about such a doomsday scenario right now.
Unlike Jimbo Fisher, I like the BCS computer rankings. Florida State’s 17th computer ranking should correct itself if they beat Florida (ranked 3rd in the computers) but it is their own fault for scheduling Murray St and Savannah St. I realize that the computer formulas have some bias built into them but that bias can’t be changed mid-season by media talking heads, campaigning coaches, conference affiliations, over-hyped offenses (Oregon), or over-hyped defenses (Alabama).
I like human rankings in the off-season because they set the table for fun debates and exciting early season match ups.
Lastly, I believe that the new BCS playoff selection committee will lean on computer rankings to make decisions similar to how the RPI is used in selecting the men’s basketball tournament. And, that the committee’s use of the computers will be greater than the current 1/3 value in the BCS rankings. Although, they’ll probably never admit that.
All good points though I actually think (or maybe just hope) the committee would be at least as open as the basketball tourney selection committee is when it comes to their criteria for choosing teams.