Wednesday, November 17, 2010

BCS, Please Stop Discriminating Against Smaller Programs by Following my Advice

Hey, Gregg Easterbook, contradict yourself in the space of two paragraphs:

And with TCU and Boise State undefeated but disfavored by the BCS setup, there's huge potential for a BCS title game that will be seen as completely bogus. The BCS insiders will do everything in their power to keep TCU or Boise State out of the title contest, and BCS insiders pull the strings.

Regardless, the obsession with style points and victory margins has gotten out of hand. One-third of the BCS ranking is based on the USA Today poll of coaches -- and most coaches are voting on hype and victory margins, since they can't possibly have time to watch film of any contenders they don't face. Retired college coach R.C. Slocum contends this poll of retired college coaches should replace the USA Today poll in the BCS formula, since retired coaches do have time to take a close look at film of every contending team.
First of all, Easterbrook, you're too smart to ignore the importance of margin-of-victory. If statisticians swear that, empirically speaking, rankings with scoring margin have greater predictive power and Vegas sharps who have skin in the game all base their models off of scoring and yardage stats, then someone who claims to have an appreciation for science and knowledge ought not dismiss scoring margin for purely aesthetic reasons.

Second, I guess it didn't occur to Easterbrook that including scoring margin is a saving grace for mid-majors like Boise State and TCU. Those teams will never be able to stay with major conference teams when computer rankings only take strength of schedule into account because their schedules aren't strong enough. The only way for them to keep pace is for the ranking to account for the fact that the Broncos and Horned Frogs are destroying their inferior opposition. The current set of Sagarin rankings proves my point. Boise State and TCU are #4 and #5 in Sagarin's Predictor, the ranking that accounts for scoring margin, and they are #12 and #6 respectively in his ELO-CHESS, the mathematical abomination that Sagarin produces for the BCS because he is required to do so.

Easterbook has two competing ideologies that are in tension with one another. On the one hand, he is a dime store moralist when it comes to scoring margin, suggesting that Wisconsin shouldn't have kicked extra points in the fourth quarter against Indiana. On the other hand, he is a critic of the BCS and major college football programs. The way for the BCS to open the door to smaller programs would be to allow its computer rankings to account for scoring margin. How does Easterbrook resolve the tension? He is either blissfully unaware of it (and I think he's too smart for that to be the case) or he willfully ignores it.


Jeff said...

Does Sagarin use a cap on margin of victory or just use the actual margin? I recall that back when the BCS included margin of victory in its calculations they counted the difference up to 2 or 3 touchdowns. I think having a cap at 28 points or so is a good way to reflect the differences between a 1 point win and a blow-out but not encourage people to hang 83 on bad teams.

One fact that exacerbates the failure to account for margin of victory is that sometimes top teams annihilate a team so badly that it crushes the morale of the beaten team and they lay an egg the next week or even rest of the season. That appears to have been what happened with Utah and I recall it happening when VT beat Syracuse 63-0 in 1999. The good team then doesn't get full credit for the blow out but suffers the downside of it.

Watch UFC Online said...

I like the approach you took with this topic. It is not every day that you find a subject so to the point and enlightening
This is fascinating, and any cross-correlation adds confidence in any sub-component's reading.
Stream NFL Football