(09-06-2016 09:14 AM)quo vadis Wrote:  The SEC remains in the #1 position for the fourth straight week. The SEC picked up a win thanks to Ole Miss beating Memphis, this raises their win rate marginally to .579.

The ACC remains #2 despite suffering Syracuse's loss to Notre Dame. The ACC does fall below .500, but their win rate of .477 remains marginally ahead of the #3 B1G.

The #3 B1G, #4 PAC, and #5 Big 12 remain where they were last week thanks to no qualifying results.

Rankings for October 2nd, 2016

Code:

CONFERENCE               WINS                     LOSSES           SEASON RECORD    


1) SEC                     MISS                 --------                   11-8 (.579)
Previous rank...1,1,1,3

2) ACC                     --------                  SYR                   10.5-11.5 (.477)
2,4,2,2    

3) B1G                      ---------                  ---------           10.5-12 (.467)
3,2,4,1  

4) PAC                    ---------                     ----------          6.5-11 (.371)
4,3,5

5) Big 12                  ---------                   --------            4-13.5 (.228)            
5,5,5,4

& 2 wins for super win (see method)

# 1.5 wins for quality win (see method)

* 1.5 losses for bad loss (see method)

@ 2 losses for an awful loss (see method)


METHOD

1) Only OOC games count.

2) Notre Dame is always regarded as a P5 opponent, including for ACC teams. BYU and Army are regarded as G5 opponents.

3) All references to "rankings" are to at the time the game was played.

4) How losses are assessed:

a) A conference is docked 1 loss for losing to any P5 opponent.

b) A conference is docked 1 loss for losing to any G5 opponent that is ranked in the Coaches or AP poll, including "others receiving votes".

c) A conference is docked 1.5 losses for losing to any unranked G5 opponent. This is a penalty for a "bad loss".

d) A conference is docked 2 losses for losing to any FCS opponent, a penalty for an "awful loss".


5) How wins are earned:

a) A conference is credited with 1 win for any win against any unranked P5 opponent.

b) A conference gets credit for 1 win against a G5 opponent if the G5 opponent is ranked in the "others receiving votes" of the AP or Coaches polls.

c) A conference earns 1.5 wins for beating a ranked team (G5 or P5, in the AP or Coaches top 25). This is a bonus for a "quality win".

d) A conference earns 2 wins if the win is against any team (G5 or P5) ranked in the top 10 in either poll. This is a bonus for a "super" win.

6) Wins against FCS teams or unranked G5 earn no credit.

7) For bowl season, the following modifications apply:

- Rankings of bowl game opponents are based on CFP rankings only. For bowl games pitting P5 vs G5, all win and loss credit is earned according to rules 4b-c and 5c-d and 6 above, with CFP substituted for AP/Coaches. So for example, beating a CFP unranked G5 team still earns no credit for the P5 conference, but losing to an unranked G5 in a bowl game does mean it suffers the "bad" loss penalty.

Quo, I'm trying to duplicate your results. I started with the ACC. I have them with 10.5 wins, as you do, but I show 12 losses to your 11.5.

1 loss games (7):

Syracuse...USF and Notre Dame
Pitt.....Oklahoma State
Virginia....Oregon
Duke.... Northwestern
UNC.... Georgia
Va Tech...Tennessee

1.5 loss games (2):

NC State...East Carolina
Virginia....UConn

2 loss games:

Virginia... Richmond

Where am I going wrong?

(10-07-2016 01:19 PM)ken d Wrote:  

(09-06-2016 09:14 AM)quo vadis Wrote:  The SEC remains in the #1 position for the fourth straight week. The SEC picked up a win thanks to Ole Miss beating Memphis, this raises their win rate marginally to .579.

The ACC remains #2 despite suffering Syracuse's loss to Notre Dame. The ACC does fall below .500, but their win rate of .477 remains marginally ahead of the #3 B1G.

The #3 B1G, #4 PAC, and #5 Big 12 remain where they were last week thanks to no qualifying results.

Rankings for October 2nd, 2016

Code:

CONFERENCE               WINS                     LOSSES           SEASON RECORD    


1) SEC                     MISS                 --------                   11-8 (.579)
Previous rank...1,1,1,3

2) ACC                     --------                  SYR                   10.5-11.5 (.477)
2,4,2,2    

3) B1G                      ---------                  ---------           10.5-12 (.467)
3,2,4,1  

4) PAC                    ---------                     ----------          6.5-11 (.371)
4,3,5

5) Big 12                  ---------                   --------            4-13.5 (.228)            
5,5,5,4

& 2 wins for super win (see method)

# 1.5 wins for quality win (see method)

* 1.5 losses for bad loss (see method)

@ 2 losses for an awful loss (see method)


METHOD

1) Only OOC games count.

2) Notre Dame is always regarded as a P5 opponent, including for ACC teams. BYU and Army are regarded as G5 opponents.

3) All references to "rankings" are to at the time the game was played.

4) How losses are assessed:

a) A conference is docked 1 loss for losing to any P5 opponent.

b) A conference is docked 1 loss for losing to any G5 opponent that is ranked in the Coaches or AP poll, including "others receiving votes".

c) A conference is docked 1.5 losses for losing to any unranked G5 opponent. This is a penalty for a "bad loss".

d) A conference is docked 2 losses for losing to any FCS opponent, a penalty for an "awful loss".


5) How wins are earned:

a) A conference is credited with 1 win for any win against any unranked P5 opponent.

b) A conference gets credit for 1 win against a G5 opponent if the G5 opponent is ranked in the "others receiving votes" of the AP or Coaches polls.

c) A conference earns 1.5 wins for beating a ranked team (G5 or P5, in the AP or Coaches top 25). This is a bonus for a "quality win".

d) A conference earns 2 wins if the win is against any team (G5 or P5) ranked in the top 10 in either poll. This is a bonus for a "super" win.

6) Wins against FCS teams or unranked G5 earn no credit.

7) For bowl season, the following modifications apply:

- Rankings of bowl game opponents are based on CFP rankings only. For bowl games pitting P5 vs G5, all win and loss credit is earned according to rules 4b-c and 5c-d and 6 above, with CFP substituted for AP/Coaches. So for example, beating a CFP unranked G5 team still earns no credit for the P5 conference, but losing to an unranked G5 in a bowl game does mean it suffers the "bad" loss penalty.

Quo, I'm trying to duplicate your results. I started with the ACC. I have them with 10.5 wins, as you do, but I show 12 losses to your 11.5.

1 loss games (7):

Syracuse...USF and Notre Dame
Pitt.....Oklahoma State
Virginia....Oregon
Duke.... Northwestern
UNC.... Georgia
Va Tech...Tennessee

1.5 loss games (2):

NC State...East Carolina
Virginia....UConn

2 loss games:

Virginia... Richmond

Where am I going wrong?

You're not - I am. I mistakenly counted ECU as a 1-loss game rather than a 1.5 loss as they actually were.

Thanks for the correction.

Curious where the AAC would rank? How far behind B12?

Give SEC 100 negative points for Florida LSU screwup. Other conferences found ways to avoid this by moving games to today to Sunday or later this season.
Cheers!

(10-07-2016 10:07 PM)tigerjamesc Wrote:  Curious where the AAC would rank? How far behind B12?

Quo is the only fan of an AAC team who likes the P5, G5 divide and won't show that the AAC is right there behind Big 12.
Cheers!

(10-07-2016 10:18 PM)billybobby777 Wrote:  Give SEC 100 negative points for Florida LSU screwup. Other conferences found ways to avoid this by moving games to today to Sunday or later this season.
Cheers!

It disgusts me to see the overreaction beforehand to events that could happen. Be prepared by all means, but the hysteria over everything that naturally has been occurring from the dawn of time is typical nanny state crap at its most absurd. I went through Camille, Hugo, and Eloise. The barrier islands and Gulf Stream are why hurricanes seldom come in under the Carolinas. I think Brunswick has only suffered 1 major storm in over 100 years. Savannah gets surge but seldom a landfall. Jacksonville which I think Dora hit is seldom a target either.

The really stupid part of all of this cancellation was the fact that even by projections the storm was due to be well past Gainesville (which is inland an hour from Jacksonville) by game time. The game could easily have been played tomorrow and what's more if the storm had caused catastrophic damage with a rare landfall that crossed Gainesville then this afternoon would have been ample time in which to call it off. The State of Florida could have just requested that R.V.'s and travelers stay inland until the storm passed.

What are the people of this country going to hear from our governmental officials if we ever face a real catastrophe? We wait to late to help on Katrina and overreact for one that misses so far. The cancellations in the Carolinas were much more merited given weather patterns the proximity of the Gulf Stream (it's over 27 miles off the coast of Jacksonville), and the history of direct hits.

(10-07-2016 10:07 PM)tigerjamesc Wrote:  Curious where the AAC would rank? How far behind B12?

Not behind. If the AAC were included, they would be 7-14.5, a .325 win %, in fifth place between the PAC and the Big 12.

Contrary to what "billybob" claims, the AAC isn't included only because it isn't an all-conferences comparison, it is a P5 conference comparison.

(10-07-2016 11:14 PM)quo vadis Wrote:  

(10-07-2016 10:07 PM)tigerjamesc Wrote:  Curious where the AAC would rank? How far behind B12?

Not behind. If the AAC were included, they would be 7-14.5, a .325 win %, in fourth place between the PAC and the Big 12.

Contrary to what "billybob" claims, the AAC isn't included only because it isn't an all-conferences comparison, it is a P5 conference comparison.

(10-08-2016 08:50 AM)tigerjamesc Wrote:  

(10-07-2016 11:14 PM)quo vadis Wrote:  

(10-07-2016 10:07 PM)tigerjamesc Wrote:  Curious where the AAC would rank? How far behind B12?

Not behind. If the AAC were included, they would be 7-14.5, a .325 win %, in fourth place between the PAC and the Big 12.

Contrary to what "billybob" claims, the AAC isn't included only because it isn't an all-conferences comparison, it is a P5 conference comparison.

A happy Memphis fan who sees a ranking that eliminates the AAC on purpose? The AAC ranks 5th. It's ahead of the Big 12. His rankings ignore that. And you are not only ok with that, but saluting Quo? I've see everything now.
Cheers!

(10-08-2016 09:02 AM)billybobby777 Wrote:  

(10-08-2016 08:50 AM)tigerjamesc Wrote:  

(10-07-2016 11:14 PM)quo vadis Wrote:  

(10-07-2016 10:07 PM)tigerjamesc Wrote:  Curious where the AAC would rank? How far behind B12?

Not behind. If the AAC were included, they would be 7-14.5, a .325 win %, in fourth place between the PAC and the Big 12.

Contrary to what "billybob" claims, the AAC isn't included only because it isn't an all-conferences comparison, it is a P5 conference comparison.

A happy Memphis fan who sees a ranking that eliminates the AAC on purpose? The AAC ranks 5th. It's ahead of the Big 12. His rankings ignore that. And you are not only ok with that, but saluting Quo? I've see everything now.
Cheers!

What part of "it's a P5 conference comparison hasn't sunk through"? This isn't something anti-AAC. I've been doing this since 2009, and back then it was the "BCS" conference comparison, involving just the six BCS-AQ conferences, meaning it did include the forerunner of the AAC, the Big East. Because that's what I'm interested in comparing.

Do you just enjoy being weird?

One obvious flaw with this method would be if the confs don't have equal number of opportunities for super wins and quality wins.

Plus the criteria for super and quality aren't being retroactively updated every week, as we find out that some teams were nowhere near as good as they started out (Iowa, for example).

(10-08-2016 09:13 AM)MplsBison Wrote:  One obvious flaw with this method would be if the confs don't have equal number of opportunities for super wins and quality wins.

Plus the criteria for super and quality aren't being retroactively updated every week, as we find out that some teams were nowhere near as good as they started out (Iowa, for example).

The first problem is a general problem with trying to rank college football teams using any methods - the schedules played even within a conference are usually significantly different, and between conferences it's often the case that two teams have no common opponents at all.

The second is possibly a problem, because teams often are overrated, especially early on.

But also maybe not. For example, the Big 12 got a "quality win" when Texas beat Notre Dame when the Irish were ranked #12. But if we applied standards retroactively, it would just be an "ordinary" win. But ... is that really accurate? Maybe, when Texas and ND played ND really was a top-15 quality team, but losing to Texas burst their bubble emotionally, which then made them a worse team going forward. In that case, the Big 12 does deserve credit for beating a top-15 team and it would be inaccurate to retrospectively downgrade that win.

Point is, just because a team was ranked #10 when they are 3-0 but unranked later when they are 6-4 doesn't necessarily mean they were "overrated" when they were 3-0, that they really weren't a #10 team when they were 3-0. They might have been at the time, but then got worse and no longer are. Teams can both get better and get worse over the course of a year, but the retroactive approach assumes that a team was always as good as it is right now, whenever "right now" is.

So I am comfortable with my "at the time the game was played" approach.

quo,

Yes. You make a great argument. In fact, it's actually the same that I would make myself if I were in your position.

So as long as it stays consistent, I would put that as more of a secondary issue.


The main problem comes in the unequal opportunities. Because, as you pointed out, CFB just doesn't have the set, standard across the board schedule of the NFL.

But I do think there is something a bit more clever that could be done in your algorithm, to address this ...

(10-08-2016 09:47 AM)MplsBison Wrote:  quo,

Yes. You make a great argument. In fact, it's actually the same that I would make myself if I were in your position.

So as long as it stays consistent, I would put that as more of a secondary issue.


The main problem comes in the unequal opportunities. Because, as you pointed out, CFB just doesn't have the set, standard across the board schedule of the NFL.

But I do think there is something a bit more clever that could be done in your algorithm, to address this ...

No question. I definitely sacrifice some accuracy for parsimony, i.e., the ease and convenience of the calculations.

It wouldn't be a difficult update, at all.

(10-08-2016 09:59 AM)MplsBison Wrote:  It wouldn't be a difficult update, at all.

Well, what's your idea?

Every analysis of college football must be flawed by the very nature of the system in which it operates. I would suggest that if somebody would like to change the criteria or methodology, they are free to offer their alternative analysis.

The fact remains that, if there were one "correct" way to do this, we would all have the definitive answer to the question "who is better". But then what would we talk about here? And what would the talking heads on ESPN and Fox talk about?

Embrace the imperfections! And don't shoot the messengers. Keep up the good work, quo!

First of all -- your system is symmetric with respect to wins and losses. So you're doing more work than you need to, in order to arrive at the same conclusion. All you need to do is accumulate the win points for each conference, using the same criteria, to get the same ranking.

But my idea is then to simply normalize each win criteria by dividing the win points of that criteria by the number of opportunities the conference had to earn points in that criteria.


The rationale would be: if conf A wins 2 out of 4 super wins, while conf B wins 1 out of 1 super wins ... it could be argued that conf B could've won 4 out of 4 super wins had it been given the same opportunity as conf A. And so, we shouldn't penalize conf B simply for lack of opportunity.


Could you make a counter-argument/counter-example to that rationale? Certainly. So I'll understand if you don't adopt that modification.

(10-08-2016 09:05 AM)quo vadis Wrote:  

(10-08-2016 09:02 AM)billybobby777 Wrote:  

(10-08-2016 08:50 AM)tigerjamesc Wrote:  

(10-07-2016 11:14 PM)quo vadis Wrote:  

(10-07-2016 10:07 PM)tigerjamesc Wrote:  Curious where the AAC would rank? How far behind B12?

Not behind. If the AAC were included, they would be 7-14.5, a .325 win %, in fourth place between the PAC and the Big 12.

Contrary to what "billybob" claims, the AAC isn't included only because it isn't an all-conferences comparison, it is a P5 conference comparison.

A happy Memphis fan who sees a ranking that eliminates the AAC on purpose? The AAC ranks 5th. It's ahead of the Big 12. His rankings ignore that. And you are not only ok with that, but saluting Quo? I've see everything now.
Cheers!

What part of "it's a P5 conference comparison hasn't sunk through"? This isn't something anti-AAC. I've been doing this since 2009, and back then it was the "BCS" conference comparison, involving just the six BCS-AQ conferences, meaning it did include the forerunner of the AAC, the Big East. Because that's what I'm interested in comparing.

Do you just enjoy being weird?

Yes I do enjoy being weird.
In 2009, 2010 and 2011 when you examined the 6 BCS conferences you had a major flaw due to ignoring several ranked schools, several top 10 and top 5 schools,---Boise, TCU, BYU, Utah, Houston--and the MWC and CUSA were better than the Big East and ACC in 2009 and 2010. The MWC was better than 3 of the 6 BCS conferences in 2009. That's the flaw in your comparisons. This year you'd better hope the AAC doesn't finish higher than the Big 12 or it's another major flaw IMHO.
Cheers!

(10-08-2016 02:39 PM)billybobby777 Wrote:  

(10-08-2016 09:05 AM)quo vadis Wrote:  

(10-08-2016 09:02 AM)billybobby777 Wrote:  

(10-08-2016 08:50 AM)tigerjamesc Wrote:  

(10-07-2016 11:14 PM)quo vadis Wrote:  Not behind. If the AAC were included, they would be 7-14.5, a .325 win %, in fourth place between the PAC and the Big 12.

Contrary to what "billybob" claims, the AAC isn't included only because it isn't an all-conferences comparison, it is a P5 conference comparison.

A happy Memphis fan who sees a ranking that eliminates the AAC on purpose? The AAC ranks 5th. It's ahead of the Big 12. His rankings ignore that. And you are not only ok with that, but saluting Quo? I've see everything now.
Cheers!

What part of "it's a P5 conference comparison hasn't sunk through"? This isn't something anti-AAC. I've been doing this since 2009, and back then it was the "BCS" conference comparison, involving just the six BCS-AQ conferences, meaning it did include the forerunner of the AAC, the Big East. Because that's what I'm interested in comparing.

Do you just enjoy being weird?

Yes I do enjoy being weird.
In 2009, 2010 and 2011 when you examined the 6 BCS conferences you had a major flaw due to ignoring several ranked schools, several top 10 and top 5 schools,---Boise, TCU, BYU, Utah, Houston--and the MWC and CUSA were better than the Big East and ACC in 2009 and 2010. The MWC was better than 3 of the 6 BCS conferences in 2009. That's the flaw in your comparisons. This year you'd better hope the AAC doesn't finish higher than the Big 12 or it's another major flaw IMHO.
Cheers!

FWIW, for 2010, Sagarin had the #6 Big East clearly ahead of the #7 MWC and WAY ahead of CUSA, who was ranked #9 among all conferences.

And in 2009, Sagarin had the Big East at #2, ahead of all the other BCS-AQ conferences except the SEC, and WAY ahead of #7 MWC and #9 CUSA.

So as usual, you are off your billybob here.

But enjoy keeping track of all my flaws.