|
Post by mikegarrison on Mar 1, 2017 14:01:32 GMT -5
I just argued that the criteria for getting the last few teams in or out isn't terribly important (as long as it is reasonably accurate and, most importantly, fair). But for seeding, it's very important that the seeding should be done based on which teams are actually the most likely to win, not which teams had the "best" (in whose eyes?) accomplishments. That's because the purpose of the seeding is to fight against the great weakness of the single-elim format -- top teams meeting too soon and knocking each other out. To get the best chance of having the best teams actually meet in the finals, the seeding *must* be done on the basis of which teams are most likely to win, not on which teams have had the most impressive season. Home court that is attached with a seed makes the seed more important - and now with regional hosts. This is not good for bracket purity - and puts much more importance on the seed order.
But we are still not agreeing with '*must* be done on the basis of which teams are most likely to win'.
In theory - you would have to say that for MLB, NBA or NFL seeding for the playoffs - they should go with an analytic metric to determine the playoff matchups -seeds (record would still determine if you make the playoffs). The Rocket teams from the 90's would have the 6th seed based on record, but Vegas (and metrics if done at the time) would have considered them the #1 or #2 team most likely to win.
It's different for a pro league because they all play each other and pretty much all their opponents in common. In the same way, I don't say to use pablo to decide conference championships or even conference tournament seedings.
|
|
|
Post by ay2013 on Mar 1, 2017 14:04:08 GMT -5
I just argued that the criteria for getting the last few teams in or out isn't terribly important (as long as it is reasonably accurate and, most importantly, fair). But for seeding, it's very important that the seeding should be done based on which teams are actually the most likely to win, not which teams had the "best" (in whose eyes?) accomplishments. That's because the purpose of the seeding is to fight against the great weakness of the single-elim format -- top teams meeting too soon and knocking each other out. To get the best chance of having the best teams actually meet in the finals, the seeding *must* be done on the basis of which teams are most likely to win, not on which teams have had the most impressive season. Going into the tournament, Penn State was 4th in Pablo this year. Would you have been ok with them getting a 4 seed and hosting regionals? Sure, why not? So long as the metric used is known and actually being applied. One of the glaring weakness of the RPI is that it doesn't control for performance beyond a simple win or a loss. Very few people actually thought PSU was what RPI said at the time a 20ish team. The ISSUE is that the criteria used for selection didn't equate to a seed, not how good PSU actually was. I mentioned it earlier but whatever system we rely heavily on for selection should have more control for performance. If PSU played on the road against Nebraska, Wisconsin, Texas, Stanford, and Washington, and lost all 5 in 5 sets, but either out scored those teams in total points or played very close 2 point sets in all of them, The raw RPI would just say 5 losses, period, and then jump teams who DOnT have as a many losses above them, but a more performance based system would probably say, if those are the top 5 teams, PSU is probably within that margin with them also very close to the top 5.
|
|
|
Post by jake on Mar 1, 2017 14:14:30 GMT -5
I hope the NCAA will also add a 'Guideline' to limit the number of teams from any one conference from being selected.
IMO,...45%-50% would be fair.
|
|
|
Post by redbeard2008 on Mar 1, 2017 14:23:46 GMT -5
The real issue for women's volleyball is much less RPI and all its various ramifications, but remaining "regionalism" which extends to the Regionals (explaining both the PSU and Washington seeds), and relative to which RPI, and its well-attested East Coast bias, merely provides convenient rationalizations rather than justifications. RPI is merely the tail, not the dog.
|
|
|
Post by trollhunter on Mar 1, 2017 14:35:51 GMT -5
There are some very smart people that I respect who disagree with this. I also come down on the side of 'accomplishment'. There is clearly no consensus on this issue and it is fundamental to Pomeroy or Pablo not being used in the selection process. I just argued that the criteria for getting the last few teams in or out isn't terribly important (as long as it is reasonably accurate and, most importantly, fair). But for seeding, it's very important that the seeding should be done based on which teams are actually the most likely to win, not which teams had the "best" (in whose eyes?) accomplishments. That's because the purpose of the seeding is to fight against the great weakness of the single-elim format -- top teams meeting too soon and knocking each other out. To get the best chance of having the best teams actually meet in the finals, the seeding *must* be done on the basis of which teams are most likely to win, not on which teams have had the most impressive season. I like and agree with your analysis of how selection of last few teams in a single-elim tournament just needs to be reasonably accurate and fair. Can't please everyone all the time. Regarding seeding, I actually think the same logic applies. While seeding is important, it is more important to just be in the tournament and have a chance. Seedings, rankings, ratings, etc. are an imperfect science, and even if they were perfect, there are still upsets, teams peaking late, injury, matchups, illness, etc. No need to really make them a huge issue for single-elim. In a single-elim, whether you lose in first round or place fifth is really a moot point. By definition, it is a winner take all event. If they wanted to find out the fifth best team, they would go with double-elim or something similar.
|
|
|
Post by ay2013 on Mar 1, 2017 14:50:33 GMT -5
I hope the NCAA will also add a 'Guideline' to limit the number of teams from any one conference from being selected. IMO,...45%-50% would be fair. Why? Under this rule, teams like Baylor and Iowa state, who had RPI in the 30's, Baylor with 22 wins, would be excluded because the big 12 is capped at 4 teams?
|
|
bluepenquin
Hall of Fame
4-Time VolleyTalk Poster of the Year (2019, 2018, 2017, 2016), All-VolleyTalk 1st Team (2021, 2020, 2019, 2018, 2017, 2016)
Posts: 12,447
|
Post by bluepenquin on Mar 1, 2017 15:13:28 GMT -5
Home court that is attached with a seed makes the seed more important - and now with regional hosts. This is not good for bracket purity - and puts much more importance on the seed order.
But we are still not agreeing with '*must* be done on the basis of which teams are most likely to win'.
In theory - you would have to say that for MLB, NBA or NFL seeding for the playoffs - they should go with an analytic metric to determine the playoff matchups -seeds (record would still determine if you make the playoffs). The Rocket teams from the 90's would have the 6th seed based on record, but Vegas (and metrics if done at the time) would have considered them the #1 or #2 team most likely to win.
It's different for a pro league because they all play each other and pretty much all their opponents in common. In the same way, I don't say to use pablo to decide conference championships or even conference tournament seedings. If the goal is to seed based on most likely to win (as you say), why wouldn't you seed a conference tournament using Pablo? Sounds like you only want the most likely to win when we don't have common opponents and it is accomplishment/record when we do have common opponents? Why use 'accomplishment' for one and 'likely to win' for the other.
Maybe it is you have no confidence in an accomplishment metric when we have unbalance schedules - but in theory the goal for both should be the same. Either you seed based on what teams have done (accomplishment) or based on likelihood of winning in the future. I still side with the idea of accomplishment - even if it is made difficult to determine by having unbalanced schedules.
|
|
|
Post by mikegarrison on Mar 1, 2017 15:19:44 GMT -5
It's different for a pro league because they all play each other and pretty much all their opponents in common. In the same way, I don't say to use pablo to decide conference championships or even conference tournament seedings. If the goal is to seed based on most likely to win (as you say), why wouldn't you seed a conference tournament using Pablo? Sounds like you only want the most likely to win when we don't have common opponents and it is accomplishment/record when we do have common opponents? Why use 'accomplishment' for one and 'likely to win' for the other.
Maybe it is you have no confidence in an accomplishment metric when we have unbalance schedules - but in theory the goal for both should be the same. Either you seed based on what teams have done (accomplishment) or based on likelihood of winning in the future. I still side with the idea of accomplishment - even if it is made difficult to determine by having unbalanced schedules.
It's not just "difficult", it's impossible. It's always going to come down to some committee's judgment of which team's accomplishments were better.
|
|
|
Post by tomclen on Mar 1, 2017 15:57:30 GMT -5
If the goal is to seed based on most likely to win (as you say), why wouldn't you seed a conference tournament using Pablo? Sounds like you only want the most likely to win when we don't have common opponents and it is accomplishment/record when we do have common opponents? Why use 'accomplishment' for one and 'likely to win' for the other.
Maybe it is you have no confidence in an accomplishment metric when we have unbalance schedules - but in theory the goal for both should be the same. Either you seed based on what teams have done (accomplishment) or based on likelihood of winning in the future. I still side with the idea of accomplishment - even if it is made difficult to determine by having unbalanced schedules.
It's not just "difficult", it's impossible. It's always going to come down to some committee's judgment of which team's accomplishments were better. Which is why I still maintain there should be a rule saying that tournament seeding can't override conference standing. I don't understand why people think that's an absurd concept, yet they seem okay with some team getting fluky hot during their conference tournament and getting an auto-bid. Remember the team that made the tournament with more than 20 losses a few years ago?
|
|
|
Post by n00b on Mar 1, 2017 17:43:57 GMT -5
It's not just "difficult", it's impossible. It's always going to come down to some committee's judgment of which team's accomplishments were better. Which is why I still maintain there should be a rule saying that tournament seeding can't override conference standing. I don't understand why people think that's an absurd concept, yet they seem okay with some team getting fluky hot during their conference tournament and getting an auto-bid. Remember the team that made the tournament with more than 20 losses a few years ago? Team A: (9-19 ,9-7 in conference, #75 SOS) Team B: (20-8, 8-8 in conference, #100 SOS) Team A should get in before Team B?
|
|
bluepenquin
Hall of Fame
4-Time VolleyTalk Poster of the Year (2019, 2018, 2017, 2016), All-VolleyTalk 1st Team (2021, 2020, 2019, 2018, 2017, 2016)
Posts: 12,447
|
Post by bluepenquin on Mar 1, 2017 18:10:55 GMT -5
It's not just "difficult", it's impossible. It's always going to come down to some committee's judgment of which team's accomplishments were better. Which is why I still maintain there should be a rule saying that tournament seeding can't override conference standing. I don't understand why people think that's an absurd concept, yet they seem okay with some team getting fluky hot during their conference tournament and getting an auto-bid. Remember the team that made the tournament with more than 20 losses a few years ago? I would favor a strong consideration for conference record and conference championship. Also, you have to consider unbalanced schedules and their impact.
But here is an example that could have happened last year in the Big 12. What if, Kansas 'struggling' through the non conference losses their 5th set to Arkansas and Xavier. They narrowly lose in the 5th set to Creighton which takes away a top 25 win. Those were all 5 set matches that Kanas won. They have a record of 24-6 instead of 27-3 and their RPI is in the mid 20's or low 30's. But they beat Texas for the Big 12 championship. Despite finishing 2nd in conference - Texas has a #4 RPI and if not 'deserving' of a #4 seed, certainly a top 8 seed. By the current metrics - Kansas would not be worthy of a seed. What do you do then. How far do you downgrade Texas to get them below Kansas or how far do you move Kansas up to get ahead of Texas. Or do both go unseeded?
|
|
|
Post by mikegarrison on Mar 1, 2017 18:59:31 GMT -5
Which is why I still maintain there should be a rule saying that tournament seeding can't override conference standing. I don't understand why people think that's an absurd concept, yet they seem okay with some team getting fluky hot during their conference tournament and getting an auto-bid. Remember the team that made the tournament with more than 20 losses a few years ago? Team A: (9-19 ,9-7 in conference, #75 SOS) Team B: (20-8, 8-8 in conference, #100 SOS) Team A should get in before Team B? Team A can't get an at-large bid by rule, because they are under .500 for the season.
|
|
|
Post by trollhunter on Mar 1, 2017 21:15:34 GMT -5
Which is why I still maintain there should be a rule saying that tournament seeding can't override conference standing. I don't understand why people think that's an absurd concept, yet they seem okay with some team getting fluky hot during their conference tournament and getting an auto-bid. Remember the team that made the tournament with more than 20 losses a few years ago? I would favor a strong consideration for conference record and conference championship. Also, you have to consider unbalanced schedules and their impact.
But here is an example that could have happened last year in the Big 12. What if, Kansas 'struggling' through the non conference losses their 5th set to Arkansas and Xavier. They narrowly lose in the 5th set to Creighton which takes away a top 25 win. Those were all 5 set matches that Kanas won. They have a record of 24-6 instead of 27-3 and their RPI is in the mid 20's or low 30's. But they beat Texas for the Big 12 championship. Despite finishing 2nd in conference - Texas has a #4 RPI and if not 'deserving' of a #4 seed, certainly a top 8 seed. By the current metrics - Kansas would not be worthy of a seed. What do you do then. How far do you downgrade Texas to get them below Kansas or how far do you move Kansas up to get ahead of Texas. Or do both go unseeded?
As I have said many times, Blue knows what's what.
|
|
|
Post by ay2013 on Mar 1, 2017 22:11:32 GMT -5
It's not just "difficult", it's impossible. It's always going to come down to some committee's judgment of which team's accomplishments were better. Which is why I still maintain there should be a rule saying that tournament seeding can't override conference standing. I don't understand why people think that's an absurd concept, yet they seem okay with some team getting fluky hot during their conference tournament and getting an auto-bid. Remember the team that made the tournament with more than 20 losses a few years ago? I agree that conference championships should be given weight, but for the standings to have veto authority over seeding? That's absurd.
|
|
|
Post by lo4um on Mar 1, 2017 22:20:13 GMT -5
I hope the NCAA will also add a 'Guideline' to limit the number of teams from any one conference from being selected. IMO,...45%-50% would be fair. That's absurd. Teams would get punished for being in tough conferences like the Big Ten. Michigan was like 7th or 8th in their conference and they were a sweet 16 team
|
|