Bracketology (11/6)

[bbg95]

Gaming RPI in basketball no longer works. Gaming NET in basketball still does, at least until they hopefully tweak it. I think most agree that RPI is a bad metric that should be abandoned in favor of something better. But until that happens, you can't blame teams and conferences for gaming it.

[n00b]

Nov 8, 2024 22:18:26 GMT -5 c4ndlelight said:

Nov 8, 2024 21:28:11 GMT -5 n00b said:

There are 31 conferences in D1 volleyball. The Big Ten is the EXCEPTION and the only conference that plays 20 matches (I think?) The argument that because the Big Ten does it, everybody should is a weird argument. Maybe the Big Ten shouldn’t be doing something that nobody else does if it hurts their RPI.

The SEC used to play 20 . . . .

The NCAA could just move away from its hyper-focus on a flawed metric that is being gamed that disincentivizes conferences setting up schedules that are good for the sport.

Why are you celebrating a system that literally incentivizes conference teams not playing. SEC used to play 20, then went down to 18, now they're at 16. You've got mid-major conferences shrinking their schedules and pulling all sorts of shenanigans to game the scheduling aspect of RPI.

The NCAA cares enough about basketball to make gaming the RPI irrelevant. The issue is that the selection committee (which is who we are bitching about in the Bracketology thread) does not care enough about women's volleyball to identify and mitigate the effects of gaming the RPI.

The NET for men's basketball is even worse than RPI in my opinion. Very high ranked teams get left out, very low ranked teams very low ranked teams qualify for the tournament. The men's basketball committee simply does a better job of evaluating resumes instead of simply looking at a team's rank in the system. And yes, I'm TOTALLY for a committee that does that.

A system that can be 'gamed' by having the best teams in your conference play each other more often is a good system. A system that could be gamed by minimizing losses for your top teams would be bad.

And for reference:

Team: RPI - Pablo (Difference)
Kentucky: 10 - 13 (+3)
Texas: 11 - 16 (+5)
Missouri: 19 - 14 (-5)
Florida: 24 - 22 (-2)
Texas A&M: 25 - 24 (-1)
Oklahoma: 35 - 34 (-1)
Tennessee: 37 - 33 (-4)
Auburn: 38 - 46 (+8)
Arkansas: 39 - 53 (+14)
South Carolina: 45 - 37 (-8)
Ole Miss: 47 - 49 (+2)
LSU: 59 - 57 (-2)
Mississippi State: 100 - 57 (-43!)
Alabama: 131 - 116 (-15)

SEC RPI Average: 46.7
SEC Pablo Average: 41.7
Teams with a better RPI than Pablo: 5
Teams with a better Pablo than RPI: 9

If the SEC is gaming the system, they're gaming Pablo even more.

[bbg95]

Nov 9, 2024 10:51:16 GMT -5 n00b said:

Nov 8, 2024 22:18:26 GMT -5 c4ndlelight said:

The SEC used to play 20 . . . .

The NCAA could just move away from its hyper-focus on a flawed metric that is being gamed that disincentivizes conferences setting up schedules that are good for the sport.

Why are you celebrating a system that literally incentivizes conference teams not playing. SEC used to play 20, then went down to 18, now they're at 16. You've got mid-major conferences shrinking their schedules and pulling all sorts of shenanigans to game the scheduling aspect of RPI.

The NCAA cares enough about basketball to make gaming the RPI irrelevant. The issue is that the selection committee (which is who we are bitching about in the Bracketology thread) does not care enough about women's volleyball to identify and mitigate the effects of gaming the RPI.

The NET for men's basketball is even worse than RPI in my opinion. Very high ranked teams get left out, very low ranked teams very low ranked teams qualify for the tournament. The men's basketball committee simply does a better job of evaluating resumes instead of simply looking at a team's rank in the system. And yes, I'm TOTALLY for a committee that does that.

A system that can be 'gamed' by having the best teams in your conference play each other more often is a good system. A system that could be gamed by minimizing losses for your top teams would be bad.

And for reference:

Team: RPI - Pablo (Difference)
Kentucky: 10 - 13 (+3)
Texas: 11 - 16 (+5)
Missouri: 19 - 14 (-5)
Florida: 24 - 22 (-2)
Texas A&M: 25 - 24 (-1)
Oklahoma: 35 - 34 (-1)
Tennessee: 37 - 33 (-4)
Auburn: 38 - 46 (+8)
Arkansas: 39 - 53 (+14)
South Carolina: 45 - 37 (-8)
Ole Miss: 47 - 49 (+2)
LSU: 59 - 57 (-2)
Mississippi State: 100 - 57 (-43!)
Alabama: 131 - 116 (-15)

SEC RPI Average: 46.7
SEC Pablo Average: 41.7
Teams with a better RPI than Pablo: 5
Teams with a better Pablo than RPI: 9

If the SEC is gaming the system, they're gaming Pablo even more.

This is a good point. Thanks for posting the Pablo numbers for all the teams.

[uofaGRAD]

trojansc where does Arizona stand now? the resume is weird

[maigrey]

Nov 9, 2024 12:47:10 GMT -5 uofaGRAD said:

trojansc where does Arizona stand now? the resume is weird

So many upsets! I'm eagerly awaiting the next installment (this isn't meant to be pressure, just excitement)

[trojansc]

Nov 9, 2024 12:47:10 GMT -5 uofaGRAD said:

trojansc where does Arizona stand now? the resume is weird

Arizona is now in a position where they just can't afford to lose at all. They have bad remaining RPI matches. They've got to win out IMO. One loss, maybe depending on how other RPI stuff shakes out, but, I highly doubt it.

[mikegarrison]

The emphasis on "out of conference" play and wins and RPI and SOS is horribly misguided, in my opinion. So what if a match was "in-conference" or "out of conference"? It's still the same match, with the same two teams playing. It should count the same.

I've heard the argument that teams have a choice who they play out of conference (which isn't true anyway -- you can't just force some other team to play you just because you want to), and for some reason that's supposed to somehow make those matches more important?

And also the flip side -- just because teams happen to be in the same conference, is that supposed to make matches between them less important? Why? What justifies devaluing a match between, say, Nebraska and Wisconsin just because they are in the same conference?

[trojansc]

Nov 9, 2024 13:00:58 GMT -5 mikegarrison said:

The emphasis on "out of conference" play and wins and RPI and SOS is horribly misguided, in my opinion. So what if a match was "in-conference" or "out of conference"? It's still the same match, with the same two teams playing. It should count the same.

I've heard the argument that teams have a choice who they play out of conference (which isn't true anyway -- you can't just force some other team to play you just because you want to), and for some reason that's supposed to somehow make those matches more important?

And also the flip side -- just because teams happen to be in the same conference, is that supposed to make matches between them less important? Why? What justifies devaluing a match between, say, Nebraska and Wisconsin just because they are in the same conference?

If you have a good W/L record - teams are begging to schedule you. You can take the WKU approach and play weaker schedules - or you can schedule like Creighton and Marquette CHOOSE to.

I think non-conference is important. I think overall body of work matters. You believe conference play is more important and should overshadow non-conference play. If there were no non-conference matches, RPI does not work. We saw this happen during COVID.

The way our committee selects teams, it's important. But we've seen before, teams are able to schedule complete cupcakes or brutal schedules and still get tournament seeds/at-larges. It depends on what you do out-of-conference that determines how good you need to be in conference to accomplish your goal.

Also want to point out that conferences can manipulate RPI. See: Sun Belt getting 4 teams in. But part of the reason they were able to get those in is that they won a couple of important non-conference matches so the committee couldn't just say their RPI was like that only because of their schedule manipulation.

[mikegarrison]

Nov 9, 2024 13:13:06 GMT -5 trojansc said:

You believe conference play is more important and should overshadow non-conference play.

Please don't insert words into my mouth that DIRECTLY contradict what I just said.

It's still the same match, with the same two teams playing. It should count the same.

I did not say conference matches should count more. I said they should count the same.

[mikegarrison]

I do think that when seeding teams from the same conference, their respective conference placement should matter. But obviously that has no meaning at all for seeding decisions between teams that are in different conferences.

[trojansc]

Nov 9, 2024 13:33:22 GMT -5 mikegarrison said:

Nov 9, 2024 13:13:06 GMT -5 trojansc said:

You believe conference play is more important and should overshadow non-conference play.

Please don't insert words into my mouth that DIRECTLY contradict what I just said.

It's still the same match, with the same two teams playing. It should count the same.

I did not say conference matches should count more. I said they should count the same.

Not in this post, but you’ve said it before when we had this discussion.

I’ll go back and see if I can find the discussion. I particularly remember you arguing that its silly how teams finish and perform worse in conference and get in the tournament over them. (Like, when Pacific took 2nd place WCC and LMU took 5th and lost H2H to Pacific twice). LMU got in because they did better in the non-conf. Pacific was left out.

[bluepenquin]

Nov 9, 2024 12:50:57 GMT -5 trojansc said:

Nov 9, 2024 12:47:10 GMT -5 uofaGRAD said:

trojansc where does Arizona stand now? the resume is weird

Arizona is now in a position where they just can't afford to lose at all. They have bad remaining RPI matches. They've got to win out IMO. One loss, maybe depending on how other RPI stuff shakes out, but, I highly doubt it.

I thank this is basically right.  Futures has them at ~ 64 after splitting this week.  If they win out - they would average around 41.  If they lose 1 more they average at 51.

I would go with the working assumption - win out they are 99% chance of being in.  They lose 1 - then they are at the mercy of where the RPI breaks out for them.  Given what happened to Kansas State last year (whose late wins were way more impressive and still didn't get in) - I wouldn't be too optimistic about Arizona's chances if that final RPI creeps into the 53+ range.

Lose 2 - and forget it.

[n00b]

Nov 9, 2024 13:00:58 GMT -5 mikegarrison said:

The emphasis on "out of conference" play and wins and RPI and SOS is horribly misguided, in my opinion. So what if a match was "in-conference" or "out of conference"? It's still the same match, with the same two teams playing. It should count the same.

I've heard the argument that teams have a choice who they play out of conference (which isn't true anyway -- you can't just force some other team to play you just because you want to), and for some reason that's supposed to somehow make those matches more important?

And also the flip side -- just because teams happen to be in the same conference, is that supposed to make matches between them less important? Why? What justifies devaluing a match between, say, Nebraska and Wisconsin just because they are in the same conference?

I generally agree with this, at least on the micro scale.

HOWEVER, I do think the NCAA is a bizarre system, where there are 350 teams, a massive talent gap between the top and bottom, and the freedom to schedule whoever you want for 1/3 of your matches. And in that system, it’s important to have incentives for good teams to schedule other good teams and increase the number of data points in which we compare teams from different conferences.

[mikegarrison]

Nov 9, 2024 16:56:34 GMT -5 n00b said:

And in that system, it’s important to have incentives for good teams to schedule other good teams and increase the number of data points in which we compare teams from different conferences.

"Good teams" already have plenty of incentive to schedule each other. The top-8 type teams already seek each other out as much as possible because wins over each other are what get them into that coveted top-4 position. They don't need extra weighting on OOC results in order to have that happen.

And I'll point out that extra weighting for OOC is the same as devaluing in-conference. When those good teams play in-conference (Pitt, Stanford, Louisville; Wisconsin, Nebraska, Penn State) should it count less for their seeding then when they play across conferences? Why?

[maigrey]

So. Much. Chaos! Wait until Washington beats Nebraska.


🤣🤣🤣🤣🤣