Bowl Challenge Cup Standings

by utesfan100 on Jan 8, 2010 9:26 AM EST Comment 16 comments

These are the final Bowl Challenge Cup Standings.

The MWC earned the 2009-2010 Bowl Challenge Cup and teh best record among all conferences.

Conference	Wins	Losses	PCT	ROS
MWC	Wyoming BYU Utah Air Force	TCU	0.800	0.714
Independents	Navy		1.000	0.667
Big East	Rutgers Pittsburgh South Florida Connecticut	West Virginia Cincinnati	0.667	0.625
SEC	Georgia Auburn Florida Mississippi Arkansas Alabama	Kentucky Tennessee LSU South Carolina	0.600	0.583
Big 10	Wisconsin Penn St. Ohio St. Iowa	Minnesota Northwestern Michigan St.	0.571	0.555
Big 12	Nebraska Oklahoma Iowa St. Texas Tech	Texas A&M Missouri Oklahoma St. Texas	0.500	0.500
WAC	Idaho Boise St.	Fresno St. Nevada	0.500	0.500
Sun Belt	Middle Tenn. St.	Troy	0.500	0.500
ACC	Clemson Virginia Tech FSU	UNC BC Miami(FL) Georgia Tech	0.429	0.444
C-USA	SMU Marshall	UCF S. Miss Houston East Carolina	0.333	0.375
PAC 10	USC UCLA	Oregon St. California Arizona Stanford Oregon	0.288	0.333
MAC	C. Michigan	Ohio Temple Bowling Green N. Illinois	0.200	0.286

Sort Order:

ROS
Most Bowls
Alphabetical

What is ROS next?

Using PCT in with a finite sample size can give misleading results, especially when the sample size is small. If a conference wins all their bowl games they will have PCT of 1.000, which if treated like a a probability would indicate that they would have certainty of winning the next one. This is clearly an error, and the reason that the official Challenge Cup Standings through out conferences with less than three games.

In the 1700's Pierre-Simon Laplace derived a formula for determining the best estimate of the probability of a pass/fail trial given a finite number of runs of this trial, called the Rule of Succession.

Wins and losses of football games can be treated as a pass/fail trial. In this context the formula becomes:

(Wins + 1) / (Total Games + 2)

For large numbers of games this approaches the PCT but remains a touch closer to 0.500 in all cases. It even allows an estimate of 0.500 if no games have been played.

This formula allows conferences with widely varying number of bowls to be compared more fairly. With this formula a 1-0 conference would be identical to a 3-1 conference and 2-0 equals 5-1.

In general, a record of W and L equals a record of N*(W+1)-1 and N*(L+1)-1 or any N.

1 recs | Comment 16 comments |

Story-email Email Printer Print

Comments

Please but that in none Mathmatician

I have no idea what any of that means

Players who should be in the Hall of Fame
DIck Lebeau, Pat TIllman, Dwight White, Donnie Shell, L.C. Greenwood, Ray Guy, Steve Tasker, Greg Llyod, Andy Russel, Chris Carter and Kevin Greene
2009 Pittsburgh Steelers the Perefect Storm of Suck
Chris Henry 1983-2009 RIP

by WVPiratesfan on Dec 21, 2009 7:54 PM EST reply actions

ROS in non-mathematician.

Like PCT the conference with the higher number is the better conference.

Unlike PCT it is fixed to account for the fact that the number of games played impacts how good our estimate is.

BCS Evolution -- Punctuating the Equilibrium - twitter

by utesfan100 on Dec 21, 2009 10:26 PM EST up reply actions

I understand now

by WVPiratesfan on Dec 22, 2009 12:02 AM EST up reply actions

Either way MWC looking good

Go Falcons and Go Horned Frogs

Mountain West Connection

by Jeremy Mauss on Dec 25, 2009 3:57 PM EST reply actions

4-1 usually will win the Bowl Challenge Cup.

BCS Evolution -- Punctuating the Equilibrium - twitter

by utesfan100 on Dec 25, 2009 5:00 PM EST up reply actions

Good work on the ROS

I have never heard about that descriptive statistic before, but it provides for a much better comparison.

by Rocket Ship Science on Dec 28, 2009 1:09 PM EST reply actions

Great use of stat, but maybe..

That’s a cool addition to account for the number of bowls a conference plays in (even though a lot of that is tied up in money). Seems like a final correction to be made after all the games had been played would be to factor in the competition the conference faced. This could be something as simple as multiplying a conference’s ROS by the average ROS of the conferences from which opponents came from. Such a correction would allow for playing (and even losing to) a high ROS conference opponent to potentially be more helpful than playing against a low ROS conference opponent.

Reason I say this is my own conference. I love the Big East and happy that we lately have had a respectable bowl win pct, but our competition outside of the top 2 bowls is not always as strong as other conferences.

by tdwv on Dec 30, 2009 4:27 PM EST reply actions

I have ethical opjections to that.

I don’t like to use measures that are beyond the control of those being judged unless it is directly related to the task at hand.

When conferences begin being able to select the quality of their bowl appearances I might add that wrinkle.

Besides, the Colley Matrix does exactly what you are talking about, granted Wes Colley does not perform a bowl only run of his method.

BCS Evolution -- Punctuating the Equilibrium - twitter

by utesfan100 on Dec 30, 2009 4:38 PM EST up reply actions

I thought that Navy was the only indie that made a bowl, so wouldn't the ROS for the indies be 1.000

Players who should be in the Hall of Fame: DIck Lebeau, Pat TIllman, Dwight White, Donnie Shell, L.C. Greenwood, Ray Guy, Steve Tasker, Greg Llyod, Andy Russel, Chris Carter, Kevin Greene and Jerry Kramer
"If you give Arians a fullback, he won’t use one. Instead, he insists on using Matt Spaeth, who probably doesn’t cast a shadow because it would require blocking sunlight." Cliffharrisisstillapunk with some very true words

by WVPiratesfan on Jan 1, 2010 3:20 PM EST reply actions

The PCT is 1.000

With 1 win in 1 game the ROS is:

(1+1) / (1+2) = 2/3 ~ 0.667

BCS Evolution -- Punctuating the Equilibrium - twitter

by utesfan100 on Jan 1, 2010 4:39 PM EST up reply actions

That Bowl Challenge Cup is going to look mighty nice on the desk of the MWC commissioner when Automatic Qualifier discussion rolls round in a few years.

by HawkeyedFrog on Jan 4, 2010 7:46 AM EST reply actions