Monday, November 15, 2010

NCAA College Football BCS Standings Week 12 - ESPN

NCAA College Football BCS Standings Week 12
Here are the biggest differences between the computers and the humans:

The computers don't like either Wisconsin (5 human, 12 computers) or Ohio State (7 human, 13 computers).

On the other side of the coin, the computers really like Oklahoma State. They rank it #6, compared to an average of #11 on the two human polls. In fact, four of the six computer services rate Oklahoma State above Boise State.

Missouri is another computer favorite (16 human, 11 computers)

I thought you guys might be interested in an e-mail I wrote to my son today. He's a huge Wisconsin fan, and wondered about the possibility that they could win the national championship. The answer is that they cannot win the BCS standings under any circumstances, but it is conceivable that they could finish #2 behind TCU, if every other big team loses. In that case they would play for the championship, and could win that game!

Here's the letter:

There are six computer rankings. The BCS system throws out the best and worst for each team and averages the other four to get the "computer score." The computer score is calculated as follows: 25 points for a first, 24 for a second, etc. Thus, the highest possible score for the four rankings is 100/100, which is expressed as 1.000. To determine Wisconsin's score, for example, throw out the 9 and one of the 14s, leaving 11-11-11-14 as the four that count. The easiest way to do the math is this - add the four together and subtract from 104. That leaves 57, expressed as .570

The score for the human polls is actual points divided by possible points. If you get every possible #1 vote in the Harris poll, you would have 2850 points (25 points for each #1, 114 voters). It is scored 25 points for a first, 24 for a second, etc. Wisconsin has 2268 points, so they are awarded a score of .7958. Every possible #1 vote in the USA Today poll would result in 1475 points (25 times 59 voters). Wisconsin has 1197, so their score is .8115

The final score is just a straight average of the three components:

  • .5700 computer
  • .7958 Harris
  • .8117 USA Today
Weighted score: .7258.

The teams are then re-ranked based upon the weighted score.

Wisconsin is really hurt by the computer third of the rankings. Their .57 compares to .89 for LSU. (LSU is rated 3-4-4-4 by the computers.) It's almost impossible to make up that big a gap. To make up a .32 deficit, they would have to score .16 higher in each of the human polls, meaning they would need .96 and .97 in the human polls to overcome LSU's edge in the computer rankings. That's better than Auburn's actual! In other words, they would have to be #1 in the human polls to beat LSU overall. #2 would not be good enough! (And that's just to beat LSU - not even considering the four undefeated teams and Stanford.)

Looking at the bigger picture, Wisconsin cannot finish #1 and would need a miracle to finish #2.

Here's why they can't finish #1:

If you look at the situation realistically, Wisconsin has no chance of any kind to pass TCU, no matter what happens. TCU's season is pretty much finished. They will finish undefeated. They have only one game left, and that is a laugher against a 1-9 team, New Mexico, which lost 72-0 to Oregon. Therefore, Wisconsin needs everyone else to lose, in which case they could finish #2 and play TCU for the championship (and they could win that game!)

Can they finish #2? It's unlikely, but theoretically possible, considering the following:

Both Auburn and LSU have a tough game left. Auburn plays Alabama and LSU plays Arkansas. (LSU's other game is against Mississippi. Auburn has no other game. The other week is an open date.) Boise State probably has the best chance to lose, since they have three more games, including one ranked team. Oregon has two games left, one of them against a ranked team. Stanford, which is also rated above Wisconsin in the overall BCS standings, has two games left against unrated but fairly tough opponents.

No comments:

Post a Comment