|
Post by nakedcrayon on Nov 4, 2014 10:18:30 GMT -5
"PAC-12 teams play 2 out of 12 teams only once. SEC teams play 7 out of 13 teams only once.
There's a huge difference in magnitude here. B1G, this year, has a similar amount of teams skipped, but at least they play a 20-game schedule."
So the SEC plays two less conference matches but also has more teams by one then the other conferences. I don't think Mary Wise calls up the SEC office and requests to play Tennessee and Miss State twice and avoid Kentucky and say TAMU twice.
Unbalanced schedules will exist in any conference with 12 or more teams. Now if you want to talk OOC schedule dodging that is an entirely different story. Only Florida Arkansas and TAMU somewhat played a OOC schedule with some meaning behind it this past season.
|
|
|
Post by FreeBall on Nov 4, 2014 10:25:32 GMT -5
With 14 (or more) teams in a conference, there is no way they are going to play a full round robin schedule. That would require each team to play 26 conference matches, which would not leave many dates open to play a non-conference schedule.
I'm not an advocate of the Big 10 having a conference tournament, but something like the following might work if you want to more fairly determine the conference champion:
1. Split the conference into two divisions of seven teams each.
2. Each team plays a round robin within their own division and plays each team in the other division once. That's a total of 19 conference matches for each team.
3. Have a four team tournament with the top two teams from each division. #1's would play #2's one day and the winners would play for the championship the next day.
|
|
|
Post by badgerbreath on Nov 4, 2014 10:36:27 GMT -5
I think something like that would be good...and by having four teams it would hedge a bit against unbalanced schedules related to who gets to play whom home and away in the other division. The downside is that RPI gets heavily affected by who is in your division and how they are playing that year. If you do it geographically, which would make sense, PSU would get stuck playing Maryland and Rutgers twice every year, which would bring down their in conference RPI into the forseeablee future.
|
|
|
Post by n00b on Nov 4, 2014 10:59:04 GMT -5
I think they are advocating for a balanced, full round robin schedule, rather than fewer conference matches. The Big 12 is the only power conference that has that (because they are small enough to do so). The Big Ten and Pac 12 don't, but nobody is accusing them of having Stanford and Washington avoid each other for seeding purposes.
|
|
|
Post by The Bofa on the Sofa on Nov 4, 2014 11:03:55 GMT -5
And I don't think anyone is accusing the SEC of doing it for seeding or selection purposes, just saying that it is a natural consequence of that type of scheduling. It allows for teams to have distorted RPI values, which can benefit them come selection time.
Although the reduction in the number of conference matches is certainly an added benefit to the SEC, because good conferences can fatten their RPIs with more out-of-conference matches.
|
|
|
Post by n00b on Nov 4, 2014 11:04:34 GMT -5
What I'd like to see is schedules designed so that similar strength teams always play each other twice. Several college basketball conferences adopted this a few years ago. Nobody benefits from Penn State playing Rutgers twice in a season, so why should it ever happen? For the sake of "fairness"? Who cares if it's "fair" if it's bad for everybody?
Contrary to what some posters seem to be thinking about the SEC. It's actually much more beneficial for At-Large teams to play more matches against the top half of the conference. Yes, the may incur a few more losses but avoiding Tennessee and Mississippi State is well worth it.
|
|
|
Post by c4ndlelight on Nov 4, 2014 11:23:26 GMT -5
What I'd like to see is schedules designed so that similar strength teams always play each other twice. Several college basketball conferences adopted this a few years ago. Nobody benefits from Penn State playing Rutgers twice in a season, so why should it ever happen? For the sake of "fairness"? Who cares if it's "fair" if it's bad for everybody? Contrary to what some posters seem to be thinking about the SEC. It's actually much more beneficial for At-Large teams to play more matches against the top half of the conference. Yes, the may incur a few more losses but avoiding Tennessee and Mississippi State is well worth it. Losses generally aren't worth it. This schedule would be bad for Florida, who expects to beat pretty much anybody anyways, but for mid-level teams that can expect to lose a good deal of these matches not playing them is better RPI-wise. (And don't forget that avoiding those losses also helps everyone else's SOS). And Alabama, for example, dodges Mississippi St. in addition to 3 of the 4 top teams, so basically their schedule is set up to be able to maximize wins against mid-level SEC teams that aren't hard to beat but have solid W-L records [because they can load up on extra wins OOC] that will help their RPI. With Alabama's resume, they should be a true bubble team, but the RPI has them borderline Top 25 and they probably will go automatically in. Did the SEC set out to pump up Alabama this year? Probably not, but the short schedule artificially raises the RPI of all SEC teams, and each year there will be a couple of teams with the unbalanced schedule that are really going to be rated way over where they should be (see Georgia last year). And frankly, the PAC-12 and Big Ten could do the same thing, but that would be bad for the teams (who actually, you know, want to play each other) and for competition in the conference. Alternatively, the Committee could recognize these implicit RPI imbalances and adjust for them, however, the reliance on RPI (and then using RPI as basis for judging good wins) makes that hard to achieve. It's not just the SEC, the expanded ACC is in a similar position. Miami might actually be getting more of an artificial boost than any SEC team this year. And of course, the Big XII is only pulling one of those levers (shorter conference schedule), but since that is mandated by size of membership there's not really much to ask them to do about it.
|
|
|
Post by ay2013 on Nov 4, 2014 12:42:49 GMT -5
Official OOC matches should be capped at 10.
|
|
|
Post by BeachbytheBay on Nov 4, 2014 12:58:09 GMT -5
Official OOC matches should be capped at 10. why? 20 conference vs 10 non-conf? more non-conf mateches make things more interesting, and give a better comparison of conferences the problem is conference heavy schedules don't mesh well with a RPI rating system Massey or Pablo provides a better analysis as conference sizes expand - don't get penalized as much for scheduling quircks
|
|
|
Post by mikegarrison on Nov 4, 2014 13:11:04 GMT -5
Ideally, for the purpose of rating teams, OoC matches should be spread throughout the season, not all concentrated in one time at the beginning.
Moreover, to get the real ideal input for a rating system that would best let us understand the strength of all teams with equal accuracy, the whole schedule for every team should be randomized. No conferences, no "Big 4" power tournaments, etc. But the reality is, we don't want that:
1) We like our conferences! They are fun traditional rivalries. 2) We don't care about having the best accuracy on knowing the strength of all teams. What we really care about is accuracy of knowing the strength of the top 20 (for seeding) and those between about 35 and 55 (for at-large selection cut-off).
|
|
|
Post by ay2013 on Nov 4, 2014 13:16:52 GMT -5
Official OOC matches should be capped at 10. why? 20 conference vs 10 non-conf? more non-conf mateches make things more interesting, and give a better comparison of conferences the problem is conference heavy schedules don't mesh well with a RPI rating system Massey or Pablo provides a better analysis as conference sizes expand - don't get penalized as much for scheduling quircks So that conferences like the SEC and Big 12 can't manipulate to get unfair ratings for their teams.
|
|
|
Post by BeachbytheBay on Nov 4, 2014 13:26:12 GMT -5
why? 20 conference vs 10 non-conf? more non-conf mateches make things more interesting, and give a better comparison of conferences the problem is conference heavy schedules don't mesh well with a RPI rating system Massey or Pablo provides a better analysis as conference sizes expand - don't get penalized as much for scheduling quircks So that conferences like the SEC and Big 12 can't manipulate to get unfair ratings for their teams. you mean those two conference are actively conspiring to manipulate the system!!?? and reducing out-of-conference matches would be beneficial for the sport?!
|
|
|
Post by The Bofa on the Sofa on Nov 4, 2014 13:27:20 GMT -5
Ideally, for the purpose of rating teams, OoC matches should be spread throughout the season, not all concentrated in one time at the beginning. Moreover, to get the real ideal input for a rating system that would best let us understand the strength of all teams with equal accuracy, the whole schedule for every team should be randomized. No conferences, no "Big 4" power tournaments, etc. But the reality is, we don't want that: 1) We like our conferences! They are fun traditional rivalries. 2) We don't care about having the best accuracy on knowing the strength of all teams. What we really care about is accuracy of knowing the strength of the top 20 (for seeding) and those between about 35 and 55 (for at-large selection cut-off). There's another issue. With conferences, it gives teams with no hope to be in the top 64 something to shoot for. For example, take the teams in the MEAC. None of these teams have any hope (these days) of being in the top 64, so in that respect, have no hope of making the NCAA tournament. But they do get to compete against schools that are nominally similar in terms of student access and budgets to win the conference title. Granted, nowadays winning the conference comes with the added benefit of getting an automatic berth to the NCAA tournament, but even if it didn't, these teams would join up in a conference so that they could compete for the conference title.
|
|
|
Post by BeachbytheBay on Nov 4, 2014 13:33:04 GMT -5
Ideally, for the purpose of rating teams, OoC matches should be spread throughout the season, not all concentrated in one time at the beginning. Moreover, to get the real ideal input for a rating system that would best let us understand the strength of all teams with equal accuracy, the whole schedule for every team should be randomized. No conferences, no "Big 4" power tournaments, etc. But the reality is, we don't want that: 1) We like our conferences! They are fun traditional rivalries. 2) We don't care about having the best accuracy on knowing the strength of all teams. What we really care about is accuracy of knowing the strength of the top 20 (for seeding) and those between about 35 and 55 (for at-large selection cut-off). the same method that addresses the strength of all teams should address the strength of the top 20 - solution = use something better than RPI solution 2 - eliminate all conferences except for BCS - then you only have to worry about say 64 teams!! more conference matches and fewer non-conference matches = play more of the same, and marginalize the smaller conferences (correlary to solution 2) scheduling non-conference aside from the beginning is just not practical, except for the smaller conferences which need a 4th weekend - but it would be an interesting scenario to have mid-majors start earlier, and then add some non-conf matches middle of season - could be an opportunity for Big West, WCC, Mizzou Valley, Big East to schedule some higher quality matches as the season goes on
|
|
|
Post by mikegarrison on Nov 4, 2014 14:52:37 GMT -5
scheduling non-conference aside from the beginning is just not practical It could be done, as long as everyone agreed to do it. Designate the last week of October as "Conference Challenge Week" or something. All (or at least enough) conferences start one week earlier and leave a gap of a week at the same time. Then the teams are free to make OoC plans for that week.
|
|