Scoring margin wins championships?

by horn_in_france on Nov 17, 2009 1:08 PM CST Comment 3 comments

Defense wins championships. That's the old adage, provided you put up a few more points than you give up.

I was recently thinking about the prolific 2005 Texas team that set an NCAA record for points scored. We also had a good defense that limited the number of points allowed. Texas was #1 in scoring offense at 50.2 ppg and #8 in scoring defense at 16.4 ppg. I decided to break out some serious math skills (a.k.a. subtraction) and look at scoring margin or margin of victory.

In 2005, Texas was #1 with an average margin of victory of 33.8 ppg. #2 was USC at 26.3 ppg.

The BCS had it right for the final game, and #1 played #2 and #1 won.

I decided to look at seasons since 2004 to look if scoring margin was as clear a signal as it was in 2005. (All stats courtesy of cfbstats.com)

All rank #s below represent scoring margin/margin of victory

2004

BCS title game: #3 USC (25.2) beat #10 OU (18.0)
#1 Louisville (30.1), #2 Utah (25.8)
#11 Texas (17.4)

2006

BCS title game: #12 Florida (16.2) over #4 tOSU (21.8)
#1 Hawaii (22.8), #2 Boise State (22.1), #3 BYU (22.1)
#7 Texas (21.1)

2007

BCS title game: #33 LSU (6.7) over #16 tOSU (13.7)
Florida #1 (30.7)
Texas #6 (23.6)

2008

BCS title game: Florida #4 (23.4) over OU #6 (21.2)
Texas #1 (29.0)
Boise State #2 (26.7), TCU #4 (26.3), #5 Cincinnati (21.5)

Looking back at the last five years of BCS championship games:

2004, 2005 and 2008 all ended with the team with the highest scoring margin as the winner of the BCS title game.

A top scoring margin team from the Big 12, SEC, Pac-10 and Big-10 frequently earns a trip to the BCS title game - exceptions are #1 Texas in 2008 (.0128) and #1 Florida in 2007. Lesser conferences need not apply.

4 of the 10 participants were ranked higher than 6th in scoring margin #10 OU, #12 Florida, #33 LSU, #16 tOSU and 2 of those teams won the game. tOSU lost both of those games.

-2009

It currently looks like the BCS championship game could very well pit two top teams in terms of scoring margin, through last weekend:

#1 Texas              29
#2 Boise State    26.7
#3 TCU                 26.3
#4 Florida             23.4
#5 Cincinnati       21.5
#6 Oklahoma      21.2
#7 Alabama         20

#6 Oklahoma proves that you can win with a lot of "style points" and still not crack the AP top 25.

Texas could finish #1 in scoring margin for the third time in five years, though none of us will care unless we win the game on Jan 7th, 2010

----

Here's a somewhat related post from a few weeks back - The State of... the BCS? A Discussion on Margin of Victory As Part of the Computers.

All comments, FanPosts, and FanShots are the views of the reader-authors who create them.

0 recs | Comment 3 comments

Story-email Email Printer Print

Comments

Some random info from Sagarin's site

   1 Alabama ( 22)| 93.13 3 | 93.69 2
   2 Florida ( 34)| 94.04 1 | 92.50 3
   3 TCU ( 48)| 90.94 4 | 91.37 4
   4 Texas ( 50)| 88.34 5 | 93.98 1

First column is SoS, second column is Elo Chess ( no MoV, uses SoS ), third column is Predictor or MoV ONLY.

If you subtract one predictor from another, you should get the score difference if they were to play each other. Add 3 for home game. So any one of the Top 4 would be favored at home against any of the others.

by notsofst on Nov 17, 2009 2:54 PM CST reply actions 0 recs

This is why the comps SUCK...

They can’t consider MOV. If you are worried about running up the scores…. then cap it by 21 pts.

The point is. OU losing by 1 to Miami is almost as bad as Oregon beating purdue 28-26. But the blind comps can’t recognize the differences. They just see OU with an L and Oregon State with a W.

Sam with nebraska and vtech. Comps can’t see that Nebraska was ahead with 10 seconds left.

Worst example. In Colley Matrix computer… Oregon State got more credit for beating UNLV 23-21 than did Texas for beating UTEP 65-7. Surely we can get that kinda thing fixed!!! Let the comp just see it as a 35-7 game for utep.

Its not JUST about being rewarded for big victories but also for being slightly punished for winning small over weak teams.

by Orangechipper on Nov 19, 2009 11:49 AM CST reply actions 0 recs

MOV must come back to the computer formulas...

I think the regulator on the max is a great idea too. That keeps Stoops in the douche category rather than rewarding him for keeping in his starters and throwing passes when up by 40. It keeps some integrity by not rewarding running up the score, but it also allows the computer to judge performance not just overall records. If you add back in MOV, the better computer algorithms could even take into account the conditions when points were scored. So they could not reward teams running up scores late and not punish teams for giving up garbage time points after the game was well in hand.

by Rickyspub on Nov 19, 2009 12:56 PM CST up reply actions 0 recs