Monday, November 3, 2008

Week 11: On Polls

With Election Day arriving tomorrow, I figured it would be appropriate to do a little commentary on the plethora of polls that determine college football's champion. Or, to be more accurate, a scathing commentary on why in the world computer ranking systems are used in determining who gets to play for the national championship.

I know why computer ranking systems came into the equation. Back in the stone ages when the BCS was evolving out of the primordial soup of the bowl alliance and bowl coalition, a select few computer ranking systems were lumped in with the human polls in the BCS formula, largely because the human polls were regarded as too incomplete. It was thought, and rightly so at the time, that human polls contained inherent bias - no one could possibly watch every game and many teams, specifically west coast teams who routinely played late at night, went unnoticed and underappreciated. And it was thought that general regional bias existed - people voted for names they knew and teams from their region. Traditional powerhouses were given preference over upstarts and the major conferences ruled the roost. Computer polls, the argument went, contained no bias; they simply spit out the rankings from a cold and unemotional formula.

Nowadays, the power program bias and regional bias still exist to some degree, but the argument that human voters can't submit an objective ballot is moot.  In the internet and 24-hour sports news era, a voter can read a quick synopsis of each game within a few minutes of its conclusion, they can TiVo or DVR games and fast-forward through, and can watch relevant highlights on College Football Final and espn.com. More than ever, voters can make an accurate assessment of who is better than whom.

Computers, on the other hand, cannot. Football is more art than science - it's an emotional game that ebbs and flows in momentum, and contains lots of those little things a computer can't see - a lucky bounce here, a questionable call there, a foot that lands millimeters out of bounds rather than in - plays that don't show up on stat sheets but have major impact on the outcome of winners, losers, and the final score. A cold mathematical formula can't possibly take these things into account. People can. Having computers as part of the determination on how teams stack up against one another is a bit like adding a computer formula as 1/3 of the electoral college. It's ludicrous and shouldn't be. Let people who can watch a game with their own 2 eyes assess the rank of each team, not soulless machines programmed by math wizards in an ivory tower somewhere! (Note: not a knock on the men who devise these intricate formulas - they're much more brainy than I. I just need to validate my use of the word "scathing")

My solution to minimizing bias and eliminating computers from the rankings? Combine the three major polls (AP, Coaches, Harris), add some more knowledgeable, independent observers to the lot, and let that one poll determine the BCS rankings. Eliminate preseason polls and begin the rankings after everyone's played 3 games. Make each person's ballot public (how hard is it to publish it on the web each week?), and hold the voter accountable for their votes. If clear bias or a record of inconsistent or questionable voting appears, put them before a vote of their peers to determine whether or not they deserve to keep their vote. Neat, tidy, simple, and fair.

On to the picks...

1. Penn State at Iowa
The Nittany Lions more than any other team have a target on their back from the computers. Their schedule won't possibly be able to stand up to that of a Big 12 or SEC team, so the threat of them being leapfrogged by a one-loss team is very real. You'd think they'd begin to feel the pressure of not only having to win, but win convincingly enough to impress both the pollsters and computers. But my hunch is that this team is level-headed enough, and well-coached enough, that those things won't cause them to crack under the pressure. Nor will it influence their game plan. I had this Iowa game circled about a month and a half ago as the danger game on Penn State's schedule. The Hawkeyes run the ball well and play good defense, and those things make you dangerous. If they can hold onto the football and keep the clock running, they can make things dicey for the Lions. But I think their offense is too one-dimensional to really dent Penn State's defense. Penn State won't walk into Kinnick Stadium and blow out the Hawkeyes, but they'll be plenty strong enough to win in classic Big Ten, grind it out fashion.
Penn State 28, Iowa 10

2. Oklahoma State at Texas Tech
After winning college football's Game of the Year, Texas Tech has to be due for a letdown, right? They face a still-underrated Oklahoma State team that just might be the best in the Big 12. They've got the most balance of all the contenders, with a 5th ranked rushing attack nationally to compliment Zac Robinson, Dez Bryant and a lethal passing game. And their defense is the most solid of all the Big 12 powers. They nearly took out Texas a week before the Red Raiders did, in Austin to boot. Mike Leach will have perhaps the most difficult coaching job of his career this week, getting his boys focused to play a very good team after last week's hoopla and the coming of the national spotlight to Lubbock. His unconventional ways might serve him well here. But I think the Cowboys get the W in another Big 12 thriller.
Oklahoma State 34, Texas Tech 28

3. Alabama at LSU
LSU will be a trendy upset pick, but I don't think the Tigers are very good. Florida and Georgia both gashed them, and while Alabama won't put up 50, they'll be able to move the ball. And the Crimson Tide defense will give Andrew Hatch or Jarrett Lee fits. Alabama survives another week.
Alabama 30, LSU 17

4. California at USC
If this were a road game for the Trojans, I'd be wary of picking them. But they're rolling right now, especially at the Coliseum, and Cal won't prove to be much more than a bump in the road.
USC 41, California 14

5. Ohio State at Northwestern
Northwestern doesn't come into games against Ohio State ranked very often, so the small yet hardy group of fans in Evanston will be ready. But Ohio State will bounce back and get a win here. I like Terrelle Pryor's maturity and competitiveness, and I think he'll show that he's got a good head on his shoulders by being able to put the heartbreaking loss to Penn State behind him.
Ohio State 27, Northwestern 13

6. TCU at Utah
I was called out this week for neglecting the Mountain West. I'm not hatin', there just hasn't been a worthy and intriguing game for awhile! Until now. One of those crazy Thursday nighters that are so hard to predict (and so hard to watch - it's televised on the "CBS College Sports Network". Who even has that channel?). Utah appears to be limping into this game, but that could just be the manner in which they win: grinding, tough wins that don't look impressive on paper. Meanwhile, TCU looks to be getting better each week, especially since their thrashing of BYU a few Thursdays back. I'll give them the nod on the road - they'll win it with defense.
TCU 21, Utah 17

7. Georgia at Kentucky
It'll be hard for Georgia to pick themselves up and be ready to play for this one. Their slim national title hopes and their SEC title hopes burned to the ground last week, so they won't have much to play for. Remember, this is the preseason #1 we're talking about - they were national title or bust. Will they continue to crumble? Or will banged up Kentucky roll over and let the Bulldogs by. Kentucky's 2 SEC wins were both by one point, against Mississippi State and Arkansas, so that tells me they've not got a whole lot to work with. Georgia steadies itself, at least for the time being.
Georgia 26, Kentucky 14

8. Georgia Tech at North Carolina
Time for the weekly rapid-fire, no thought ACC picks. The true definition of crapshoot. I've actually done pretty well in ACC games since my rant a few weeks back - I think my formula of "you're ranked, you'll lose 4 times out of 5" is working. But what happens if both are ranked? Can they both lose?
Georgia Tech 24, North Carolina 23

9. Clemson at Florida State
Call it the 1/2 Bowden Bowl.
Florida State 31, Clemson 21

10. Maryland at Virginia Tech
Maryland's already lost this season after being newly ranked, so the precedent is set.
Virginia Tech 20, Maryland 13

11. Cincinnati at West Virginia
West Virginia is coming on strong, but they still look to be a team that would get crushed by an elite opponent. Good thing no such team exists in the Big East. Whoever wins that conference, along with the ACC winner, may just collide in the worst matchup in the history of the BCS. Good thing for Miami that they get to host the BCS Championship game along with the Orange Bowl this year. And while we're here, kudos to Marshwan Gilyard for his response to the kid he steamrolled in the stands last week, and kudos to the little guy who took a shot and got back up. He's got some slot receiver in him.
West Virginia 34, Cincinnati 27

12. Louisville at Pittsburgh
Pitt plays to the level of it's competition, and Louisville just lost to Syracuse. So expect an ugly game.
Pittsburgh 31, Louisville 26

13. Notre Dame at Boston College
Tough road trip for Notre Dame after a difficult-to-swallow 4OT loss at home to Pitt. Boston College loves playing spoiler to Notre Dame, and they'd love to complete the severing of any mention of "Notre Dame" and "BCS" in the same sentence. Unless that sentence would read, "Notre Dame is yet again not going to play in a BCS bowl."
Boston College 25, Notre Dame 23

14. Stanford at Oregon
These two are tied for 4th in the Pac Ten. 
That's about all the commentary I can muster for this one.
Oregon 42, Stanford 27

15. Oklahoma at Texas A&M
I feel obliged to pick either this game or Florida-Vanderbilt, not because I expect it to be a tough pick, but because it's hard skipping over both of those national title contenders. I'm choosing this game because it'll be closer. Texas A&M is improving while Vandy is declining. And the Aggies have more history. Oklahoma should show up and continue the status quo in this one, which lately has meant scoring half a hundred by intermission and flipping on cruise control in the second half. A&M might make it a little more interesting than that - they could hold the Sooners under 40 before halftime. Moral victories like that are worth celebrating in Aggieland, where they've not had much to cheer about lately.
Oklahoma 52, Texas A&M 31


No comments: