Hello There, Guest! (LoginRegister)

Post Reply 
ACC eliminating divisions 2023
Author Message
Crayton Offline
1st String
*

Posts: 1,786
Joined: Feb 2019
Reputation: 104
I Root For: Florida
Location:
Post: #181
RE: ACC eliminating divisions 2023
(05-16-2022 01:42 PM)Hokie Mark Wrote:  
(05-16-2022 11:44 AM)Crayton Wrote:  
(05-15-2022 06:32 PM)Hokie Mark Wrote:  
(05-15-2022 07:49 AM)Crayton Wrote:  At some point this week I may do a sim on divisionless to see how often non-head-to-head tie-breakers need to be employed. I imagine for divisions it is ~1%, but for divisionless I have no idea. 20%?

put me down for no more than 5% to 10% tiebreakers without divisions.
oh, and bump up tiebreakers WITH divisions to at least 5%, too.

Results are in:
Divisionless ("pods" North/NC/South): 34.46%
Divisionless (Wahoowa's rivals): 32.59%
Divisionless (but using current schedule): 30.85%
Divisions (current schedule): 3.29%

5% was a better guess than 1%. But with no divisions (under a variety of rivalry structures), every 3rd year will require the ACC to use a non-head-to-head tie-breaker.

Other metrics to track? Frequency of rematches? Average CFP rank of champion?

Which teams in particular did you show ending up tied? I just can't imagine that many ties among teams which wouldn't play each other in the regular season.

The factor you may be missing is 3-way ties. Unless 1 team sweeps the other 2 you have to use more esoteric tie-breakers. 3-way ties at 7-1 were slightly more likely that 3-way ties at 6-2 (for the #2 spot).

I used a mix of current and historic strength. Clemson-Pitt is probably the most common. Because the NC teams largely play each other, one of them can sometimes luck their way to 7-1. I thought to track which pairs were most commonly tied for #2 without head-to-head, but that may need to be done a different day.
05-16-2022 04:27 PM
Find all posts by this user Quote this message in a reply
MWC Tex Offline
Heisman
*

Posts: 7,777
Joined: Aug 2012
Reputation: 179
I Root For: MW
Location: TX
Post: #182
RE: ACC eliminating divisions 2023
(05-16-2022 11:44 AM)Crayton Wrote:  
(05-15-2022 06:32 PM)Hokie Mark Wrote:  
(05-15-2022 07:49 AM)Crayton Wrote:  At some point this week I may do a sim on divisionless to see how often non-head-to-head tie-breakers need to be employed. I imagine for divisions it is ~1%, but for divisionless I have no idea. 20%?

put me down for no more than 5% to 10% tiebreakers without divisions.
oh, and bump up tiebreakers WITH divisions to at least 5%, too.

Results are in:
Divisionless ("pods" North/NC/South): 34.46%
Divisionless (Wahoowa's rivals): 32.59%
Divisionless (but using current schedule): 30.85%
Divisions (current schedule): 3.29%

5% was a better guess than 1%. But with no divisions (under a variety of rivalry structures), every 3rd year will require the ACC to use a non-head-to-head tie-breaker.

Other metrics to track? Frequency of rematches? Average CFP rank of champion?

That is a pretty high percentage as I had suspected.
05-16-2022 04:34 PM
Find all posts by this user Quote this message in a reply
esayem Offline
Hark The Sound!
*

Posts: 9,858
Joined: Feb 2007
Reputation: 427
I Root For: Statefan
Location: Tobacco Road
Post: #183
RE: ACC eliminating divisions 2023
(05-16-2022 03:29 PM)PeteTheChop Wrote:  
(05-16-2022 02:50 PM)Hokie Mark Wrote:  I don't really care lol, it was all based on responding to a smarmy fantasy post by Pete. You think UNC Asheville is joining the ACC?

Never smarmy ... I promise ya

Got a smile on my face when I'm typing those wisecracks -- same as when I read a (usually accurate and well-deserved) swipe at my Canes. All in fun!

04-cheers

... and why wouldn't Asheville make a good ACC travel partner for App State? 03-wink

Cuz those good folks in Cullowhee would split their britches!
05-16-2022 05:59 PM
Find all posts by this user Quote this message in a reply
Hokie Mark Offline
Hall of Famer
*

Posts: 20,342
Joined: Sep 2011
Reputation: 979
I Root For: VT, ACC teams
Location: Charlotte, NC
Post: #184
RE: ACC eliminating divisions 2023
(05-16-2022 03:29 PM)PeteTheChop Wrote:  
(05-16-2022 02:50 PM)Hokie Mark Wrote:  I don't really care lol, it was all based on responding to a smarmy fantasy post by Pete. You think UNC Asheville is joining the ACC?

Never smarmy ... I promise ya

Got a smile on my face when I'm typing those wisecracks -- same as when I read a (usually accurate and well-deserved) swipe at my Canes. All in fun!

04-cheers

... and why wouldn't Asheville make a good ACC travel partner for App State? 03-wink

I can't take any credit for UNC Asheville - that was esayem (check the original post).
05-17-2022 01:08 AM
Visit this user's website Find all posts by this user Quote this message in a reply
Wedge Offline
Moderator
*

Posts: 19,712
Joined: May 2010
Reputation: 960
I Root For: California
Location: IV, V, VI, IX
Post: #185
RE: ACC eliminating divisions 2023
(05-16-2022 04:27 PM)Crayton Wrote:  
(05-16-2022 01:42 PM)Hokie Mark Wrote:  
(05-16-2022 11:44 AM)Crayton Wrote:  
(05-15-2022 06:32 PM)Hokie Mark Wrote:  
(05-15-2022 07:49 AM)Crayton Wrote:  At some point this week I may do a sim on divisionless to see how often non-head-to-head tie-breakers need to be employed. I imagine for divisions it is ~1%, but for divisionless I have no idea. 20%?

put me down for no more than 5% to 10% tiebreakers without divisions.
oh, and bump up tiebreakers WITH divisions to at least 5%, too.

Results are in:
Divisionless ("pods" North/NC/South): 34.46%
Divisionless (Wahoowa's rivals): 32.59%
Divisionless (but using current schedule): 30.85%
Divisions (current schedule): 3.29%

5% was a better guess than 1%. But with no divisions (under a variety of rivalry structures), every 3rd year will require the ACC to use a non-head-to-head tie-breaker.

Other metrics to track? Frequency of rematches? Average CFP rank of champion?

Which teams in particular did you show ending up tied? I just can't imagine that many ties among teams which wouldn't play each other in the regular season.

The factor you may be missing is 3-way ties. Unless 1 team sweeps the other 2 you have to use more esoteric tie-breakers. 3-way ties at 7-1 were slightly more likely that 3-way ties at 6-2 (for the #2 spot).

I used a mix of current and historic strength. Clemson-Pitt is probably the most common. Because the NC teams largely play each other, one of them can sometimes luck their way to 7-1. I thought to track which pairs were most commonly tied for #2 without head-to-head, but that may need to be done a different day.

The bolded scenario will be most likely to occur.

The unbreakable 3-way tie in which each team is 1-1 vs the other two is less likely in the no-division format because of the possible situations in which the 3 teams didn’t all play each other. There is only a very small possibility of a 3-way tie in which none of the 3 teams played either of the others.
05-17-2022 02:17 AM
Find all posts by this user Quote this message in a reply
Crayton Offline
1st String
*

Posts: 1,786
Joined: Feb 2019
Reputation: 104
I Root For: Florida
Location:
Post: #186
RE: ACC eliminating divisions 2023
(05-17-2022 02:17 AM)Wedge Wrote:  
(05-16-2022 04:27 PM)Crayton Wrote:  The factor you may be missing is 3-way ties. Unless 1 team sweeps the other 2 you have to use more esoteric tie-breakers. 3-way ties at 7-1 were slightly more likely that 3-way ties at 6-2 (for the #2 spot).

I used a mix of current and historic strength. Clemson-Pitt is probably the most common. Because the NC teams largely play each other, one of them can sometimes luck their way to 7-1. I thought to track which pairs were most commonly tied for #2 without head-to-head, but that may need to be done a different day.

The bolded scenario will be most likely to occur.

The unbreakable 3-way tie in which each team is 1-1 vs the other two is less likely in the no-division format because of the possible situations in which the 3 teams didn’t all play each other. There is only a very small possibility of a 3-way tie in which none of the 3 teams played either of the others.

You actually get MORE unbreakable 3-way ties at 7-1 with a Single Division. Team A is 1-0, Team B is 0-1 and Team C is 0-0 or 1-1. I was thinking of tracking which teams tied but did not play.

The more I think of it the more I consider next doing some version of the “flex” week scenario, where 4 of the top 5 teams (after 6 conference games) are paired that week, home sites pre-determined. Should bring the likelihood of an unbreakable tie down to… maybe 10% but may increase the rematch rate and increase the average loss-count per champ.

Also, was clicking through individual sims and saw one with THREE undefeated teams. Yikes! I think the only way to avoid that is to have hidden scheduling divisions so no set of 3 teams avoid playing each other.
(This post was last modified: 05-17-2022 10:08 AM by Crayton.)
05-17-2022 05:47 AM
Find all posts by this user Quote this message in a reply
Crayton Offline
1st String
*

Posts: 1,786
Joined: Feb 2019
Reputation: 104
I Root For: Florida
Location:
Post: #187
RE: ACC eliminating divisions 2023
(05-17-2022 05:47 AM)Crayton Wrote:  The more I think of it the more I consider next doing some version of the “flex” week scenario, where 4 of the top 5 teams (after 6 conference games) are paired that week, home sites pre-determined. Should bring the likelihood of an unbreakable tie down to… maybe 10% but may increase the rematch rate and increase the average loss-count per champ.

Got it down to 21.65% with a crude "flex" week re-pairing protocol. May have to do discrete math to see if there is an optimal solution. Not sure I'm up for that this week.
(This post was last modified: 05-17-2022 10:08 AM by Crayton.)
05-17-2022 10:03 AM
Find all posts by this user Quote this message in a reply
Wedge Offline
Moderator
*

Posts: 19,712
Joined: May 2010
Reputation: 960
I Root For: California
Location: IV, V, VI, IX
Post: #188
RE: ACC eliminating divisions 2023
(05-17-2022 05:47 AM)Crayton Wrote:  
(05-17-2022 02:17 AM)Wedge Wrote:  
(05-16-2022 04:27 PM)Crayton Wrote:  The factor you may be missing is 3-way ties. Unless 1 team sweeps the other 2 you have to use more esoteric tie-breakers. 3-way ties at 7-1 were slightly more likely that 3-way ties at 6-2 (for the #2 spot).

I used a mix of current and historic strength. Clemson-Pitt is probably the most common. Because the NC teams largely play each other, one of them can sometimes luck their way to 7-1. I thought to track which pairs were most commonly tied for #2 without head-to-head, but that may need to be done a different day.

The bolded scenario will be most likely to occur.

The unbreakable 3-way tie in which each team is 1-1 vs the other two is less likely in the no-division format because of the possible situations in which the 3 teams didn’t all play each other. There is only a very small possibility of a 3-way tie in which none of the 3 teams played either of the others.

You actually get MORE unbreakable 3-way ties at 7-1 with a Single Division. Team A is 1-0, Team B is 0-1 and Team C is 0-0 or 1-1. I was thinking of tracking which teams tied but did not play.

Those ties are not unbreakable. Team A has the best head-to-head record among those three teams and thus wins the tiebreaker.

Is that "fair"? It's no more or less fair than using head-to-head between two teams-- if two teams played an 8 or 9 game conference schedule and finished with the same conference record, why should we reward the team that won their head-to-head game while losing a game to one of the conference's lesser teams? The answer is that we use that tiebreaker not because it is or isn't fair, but because it is clear-cut. The same answer applies to the above scenario in which the tiebreaker winner is Team A.
05-17-2022 11:30 AM
Find all posts by this user Quote this message in a reply
Crayton Offline
1st String
*

Posts: 1,786
Joined: Feb 2019
Reputation: 104
I Root For: Florida
Location:
Post: #189
RE: ACC eliminating divisions 2023
(05-17-2022 11:30 AM)Wedge Wrote:  
(05-17-2022 05:47 AM)Crayton Wrote:  
(05-17-2022 02:17 AM)Wedge Wrote:  The unbreakable 3-way tie in which each team is 1-1 vs the other two is less likely in the no-division format because of the possible situations in which the 3 teams didn’t all play each other. There is only a very small possibility of a 3-way tie in which none of the 3 teams played either of the others.

You actually get MORE unbreakable 3-way ties at 7-1 with a Single Division. Team A is 1-0, Team B is 0-1 and Team C is 0-0 or 1-1. I was thinking of tracking which teams tied but did not play.

Those ties are not unbreakable. Team A has the best head-to-head record among those three teams and thus wins the tiebreaker.

Is that "fair"? It's no more or less fair than using head-to-head between two teams-- if two teams played an 8 or 9 game conference schedule and finished with the same conference record, why should we reward the team that won their head-to-head game while losing a game to one of the conference's lesser teams? The answer is that we use that tiebreaker not because it is or isn't fair, but because it is clear-cut. The same answer applies to the above scenario in which the tiebreaker winner is Team A.

I certainly see how even in a two-team tie, favoring the team who lost to a worse team can be perceived as arbitrary or even backward (of course in a one division scenario that additional loss will likely be to the 8-0 regular season champ). But that is the tie-break we choose because fans prefer the on-the-field result between the two teams to be determinative.

BUT I don't see much difference between using win% among tied teams (A wins because they are 1-0) and using record vs. common opponent and then record of uncommon opponents to break a tie when there is no head-to-head sweep. That'd arguably have settled even the Big 12 fiasco of 2008... though you know conference heads are going to want to use a national ranking when they can (ex. the AAC tiebreakers would have gone to the CFP in our A,B,C scenario).

I may not have framed my simulated pursuit coherently. It is not: how often is a CFP-ranking used as a tie-breaker (with the right tiebreakers that can be <5% in any system), but how often must a team be chosen when they did not play another team they are tied with.

(05-15-2022 09:57 AM)bullet Wrote:  
(05-15-2022 07:46 AM)Crayton Wrote:  Every coach will favor divisions because that means they compete head-to-head against teams with very similar schedules. In divisionless they could lose out to a team they did not play and who played 3 or 4 “easier” teams. Hopefully they put in the Big Ten tie breakers where record vs. common opponents then record of uncommon opponents are used, instead of some G5 “ranking in most reason CFP poll.”
Its why many of us fans prefer divisions. I like seeing it decided as much as possible on the field by teams who earned their way there, instead of random tiebreaks necessitated by disparate schedules or beauty contests deciding who goes.

How often will a team miss a chance on playing in the CCG because of a scheduling quirk rather than losing it on the field? 33%
(This post was last modified: 05-17-2022 01:13 PM by Crayton.)
05-17-2022 12:52 PM
Find all posts by this user Quote this message in a reply
Wedge Offline
Moderator
*

Posts: 19,712
Joined: May 2010
Reputation: 960
I Root For: California
Location: IV, V, VI, IX
Post: #190
RE: ACC eliminating divisions 2023
(05-17-2022 12:52 PM)Crayton Wrote:  BUT I don't see much difference between using win% among tied teams (A wins because they are 1-0) and using record vs. common opponent and then record of uncommon opponents to break a tie when there is no head-to-head sweep. That'd arguably have settled even the Big 12 fiasco of 2008... though you know conference heads are going to want to use a national ranking when they can (ex. the AAC tiebreakers would have gone to the CFP in our A,B,C scenario).

IMO, any way to use on-the-field results to break a tie is better than using the rankings of 12 people who enjoy the expense-account life at some luxe hotel near DFW Airport while using who-knows-which statistics to rank teams whose games they may not have watched.

Head to head results are better, any win percentage among tied teams is better, record vs. common opponents is better, even point differential in conference games is better because all of those things come from the results of actual football games and not from some group's maybe-informed-maybe-not ranking of football teams.
05-17-2022 01:16 PM
Find all posts by this user Quote this message in a reply
Crayton Offline
1st String
*

Posts: 1,786
Joined: Feb 2019
Reputation: 104
I Root For: Florida
Location:
Post: #191
RE: ACC eliminating divisions 2023
^And I suppose it all boils down to conference preference. In the event of a 3-way tie the AAC chose to go by CFP Ranking, because they are competing against the other G5 for 1 NY6 spot. But the Big Ten chose to go by record of uncommon opponents, because they don’t mind if a lower-ranked team makes the CCG. What will the ACC do? Right now it looks like they’ll defer to SportSource Analytics, a ranking.
05-17-2022 01:21 PM
Find all posts by this user Quote this message in a reply
jimrtex Offline
Special Teams
*

Posts: 758
Joined: Aug 2021
Reputation: 27
I Root For: Houston, Tulsa, Colorado
Location:
Post: #192
RE: ACC eliminating divisions 2023
(05-17-2022 05:47 AM)Crayton Wrote:  
(05-17-2022 02:17 AM)Wedge Wrote:  
(05-16-2022 04:27 PM)Crayton Wrote:  The factor you may be missing is 3-way ties. Unless 1 team sweeps the other 2 you have to use more esoteric tie-breakers. 3-way ties at 7-1 were slightly more likely that 3-way ties at 6-2 (for the #2 spot).

I used a mix of current and historic strength. Clemson-Pitt is probably the most common. Because the NC teams largely play each other, one of them can sometimes luck their way to 7-1. I thought to track which pairs were most commonly tied for #2 without head-to-head, but that may need to be done a different day.

The bolded scenario will be most likely to occur.

The unbreakable 3-way tie in which each team is 1-1 vs the other two is less likely in the no-division format because of the possible situations in which the 3 teams didn’t all play each other. There is only a very small possibility of a 3-way tie in which none of the 3 teams played either of the others.

You actually get MORE unbreakable 3-way ties at 7-1 with a Single Division. Team A is 1-0, Team B is 0-1 and Team C is 0-0 or 1-1. I was thinking of tracking which teams tied but did not play.

The more I think of it the more I consider next doing some version of the “flex” week scenario, where 4 of the top 5 teams (after 6 conference games) are paired that week, home sites pre-determined. Should bring the likelihood of an unbreakable tie down to… maybe 10% but may increase the rematch rate and increase the average loss-count per champ.

Also, was clicking through individual sims and saw one with THREE undefeated teams. Yikes! I think the only way to avoid that is to have hidden scheduling divisions so no set of 3 teams avoid playing each other.
Maybe they should switch from horizontal to vertical divisions. Based on average wins over last 10 seasons (to compensate for Louisville, Pitt, Syracuse, and Boston College):

Clemson 7.2
Miami(FL) 5.0
FSU 5.0
Pitt 4.8
Virginia Tech 4.5
North Carolina 4.2
Louisville 4.1
=========
NC State 3.8
Georgia Tech 3.7
Wake Forest 3.1
Virginia 3.0
Boston College 3.0
Duke 2.9
Syracuse 2.4

Play a round-robin in division which determines standings and play 3 cross-over games, which might be a blend of rival match-ups and long-skipped opponents. For example, UNC might play Virginia and Duke, and infrequent opponent Boston College.

Let a computer choose H/A based on optimizing switches from previous location. Season-to-season there will be some interchange since teams in upper division play better schedules, while those in lower division play weaker opponents.
05-17-2022 04:10 PM
Find all posts by this user Quote this message in a reply
jimrtex Offline
Special Teams
*

Posts: 758
Joined: Aug 2021
Reputation: 27
I Root For: Houston, Tulsa, Colorado
Location:
Post: #193
RE: ACC eliminating divisions 2023
(05-17-2022 12:52 PM)Crayton Wrote:  
(05-17-2022 11:30 AM)Wedge Wrote:  
(05-17-2022 05:47 AM)Crayton Wrote:  
(05-17-2022 02:17 AM)Wedge Wrote:  The unbreakable 3-way tie in which each team is 1-1 vs the other two is less likely in the no-division format because of the possible situations in which the 3 teams didn’t all play each other. There is only a very small possibility of a 3-way tie in which none of the 3 teams played either of the others.

You actually get MORE unbreakable 3-way ties at 7-1 with a Single Division. Team A is 1-0, Team B is 0-1 and Team C is 0-0 or 1-1. I was thinking of tracking which teams tied but did not play.

Those ties are not unbreakable. Team A has the best head-to-head record among those three teams and thus wins the tiebreaker.

Is that "fair"? It's no more or less fair than using head-to-head between two teams-- if two teams played an 8 or 9 game conference schedule and finished with the same conference record, why should we reward the team that won their head-to-head game while losing a game to one of the conference's lesser teams? The answer is that we use that tiebreaker not because it is or isn't fair, but because it is clear-cut. The same answer applies to the above scenario in which the tiebreaker winner is Team A.

I certainly see how even in a two-team tie, favoring the team who lost to a worse team can be perceived as arbitrary or even backward (of course in a one division scenario that additional loss will likely be to the 8-0 regular season champ). But that is the tie-break we choose because fans prefer the on-the-field result between the two teams to be determinative.

BUT I don't see much difference between using win% among tied teams (A wins because they are 1-0) and using record vs. common opponent and then record of uncommon opponents to break a tie when there is no head-to-head sweep. That'd arguably have settled even the Big 12 fiasco of 2008... though you know conference heads are going to want to use a national ranking when they can (ex. the AAC tiebreakers would have gone to the CFP in our A,B,C scenario).

I may not have framed my simulated pursuit coherently. It is not: how often is a CFP-ranking used as a tie-breaker (with the right tiebreakers that can be <5% in any system), but how often must a team be chosen when they did not play another team they are tied with.

(05-15-2022 09:57 AM)bullet Wrote:  
(05-15-2022 07:46 AM)Crayton Wrote:  Every coach will favor divisions because that means they compete head-to-head against teams with very similar schedules. In divisionless they could lose out to a team they did not play and who played 3 or 4 “easier” teams. Hopefully they put in the Big Ten tie breakers where record vs. common opponents then record of uncommon opponents are used, instead of some G5 “ranking in most reason CFP poll.”
Its why many of us fans prefer divisions. I like seeing it decided as much as possible on the field by teams who earned their way there, instead of random tiebreaks necessitated by disparate schedules or beauty contests deciding who goes.

How often will a team miss a chance on playing in the CCG because of a scheduling quirk rather than losing it on the field? 33%
Another approach is to use a double round-robin using the latest games available.

Each season schedule the 5 opponents not played the previous season, and then choose 3 or 4 other opponents.
05-17-2022 04:26 PM
Find all posts by this user Quote this message in a reply
Post Reply 




User(s) browsing this thread: 3 Guest(s)


Copyright © 2002-2022 Collegiate Sports Nation Bulletin Board System (CSNbbs), All Rights Reserved.
CSNbbs is an independent fan site and is in no way affiliated to the NCAA or any of the schools and conferences it represents.
This site monetizes links. FTC Disclosure.
We allow third-party companies to serve ads and/or collect certain anonymous information when you visit our web site. These companies may use non-personally identifiable information (e.g., click stream information, browser type, time and date, subject of advertisements clicked or scrolled over) during your visits to this and other Web sites in order to provide advertisements about goods and services likely to be of greater interest to you. These companies typically use a cookie or third party web beacon to collect this information. To learn more about this behavioral advertising practice or to opt-out of this type of advertising, you can visit http://www.networkadvertising.org.
Powered By MyBB, © 2002-2022 MyBB Group.