Ultimate College Softball
Register Calendar Latest Topics
 
 
 


Reply
  Author   Comment   Page 5 of 6     «   Prev   2   3   4   5   6   Next
3leftturns

Registered:
Posts: 11,255
Reply with quote  #121 
Quote:
Originally Posted by eeyore


Note that your analysis is circular; by saying that the top teams play a lot of games against other teams near the top of RPI, you are saying that RPI is a very good predictor of itself.  That should surprise no one.
Nothing circular. SEC and Pac teams prove their value out of conference before they start beating up each other.

Records vs. teams outside their conference
SEC
Florida: 30-2
Auburn: 28-2
Texas A&M: 26-2
Tennessee: 28-2
Alabama: 28-4
LSU: 26-5
Kentucky: 23-4
Ole Miss: 25-4
South Carolina: 24-7
Georgia: 27-3
Arkansas: 24-4
Miss. State: 26-4
Mizzou: 22-9
337-52 (26-4 average per team) (.867)

Pac 12
Oregon 30-0
UCLA 26-5
Washington 27-3
Arizona 30-1
Utah 20-5
Arizona State 21-5
Cal 24-5
Oregon State 19-10
Stanford 17-10
214-44 (24-5) (.829)

Big 12
Oklahoma 30-7
Baylor: 28-6
Oklahoma State: 21-16
Texas: 23-12

Kansas 21-14
Texas Tech 14-20
Iowa State 17-20
154-95 (22-14) (.618)


ACC
FSU 24-6-1
UNC 21-9
Notre Dame 19-9
Louisville 17-10
Pitt 17-6
BC 17-12
Syracuse 21-8
NC State 7-23
Virginia 13-16
Georgia Tech 11-17
Virginia Tech 14-15
181-131-1 (17-12) (.580)

Big Ten
Minnesota 27-2
Michigan 21-7-1
Illinois 22-9
Wisconsin 20-3
Ohio State 18-7
Northwestern 14-15
Nebraska 11-18
Michigan State 18-10
Indiana 14-16
Penn State 14-17
Purdue 11-21
Iowa 10-17
Maryland 7-18-1
Rutgers 9-18
216-172 (15-12) (.557)

BigOrangeSoftball

Registered:
Posts: 220
Reply with quote  #122 
Quote:
Originally Posted by 3leftturns
Nothing circular. SEC and Pac teams prove their value out of conference before they start beating up each other.

Records vs. teams outside their conference
SEC
Florida: 30-2
Auburn: 28-2
Texas A&M: 26-2
Tennessee: 28-2
Alabama: 28-4
LSU: 26-5
Kentucky: 23-4
Ole Miss: 25-4
South Carolina: 24-7
Georgia: 27-3
Arkansas: 24-4
Miss. State: 26-4
Mizzou: 22-9
337-52 (26-4 average per team) (.867)

Pac 12
Oregon 30-0
UCLA 26-5
Washington 27-3
Arizona 30-1
Utah 20-5
Arizona State 21-5
Cal 24-5
Oregon State 19-10
Stanford 17-10
214-44 (24-5) (.829)

Big 12
Oklahoma 30-7
Baylor: 28-6
Oklahoma State: 21-16
Texas: 23-12

Kansas 21-14
Texas Tech 14-20
Iowa State 17-20
154-95 (22-14) (.618)


ACC
FSU 24-6-1
UNC 21-9
Notre Dame 19-9
Louisville 17-10
Pitt 17-6
BC 17-12
Syracuse 21-8
NC State 7-23
Virginia 13-16
Georgia Tech 11-17
Virginia Tech 14-15
181-131-1 (17-12) (.580)

Big Ten
Minnesota 27-2
Michigan 21-7-1
Illinois 22-9
Wisconsin 20-3
Ohio State 18-7
Northwestern 14-15
Nebraska 11-18
Michigan State 18-10
Indiana 14-16
Penn State 14-17
Purdue 11-21
Iowa 10-17
Maryland 7-18-1
Rutgers 9-18
216-172 (15-12) (.557)



Pretty much speaks for itself. Oklahoma fans don't seem to realize that not everyone "took their lumps." And if they did, they gave out more lumps to elite level teams than oklahoma did.
eeyore

Registered:
Posts: 46
Reply with quote  #123 
Quote:
Originally Posted by 3leftturns
Nothing circular. SEC and Pac teams prove their value out of conference before they start beating up each other.

Records vs. teams outside their conference . . .



You can keep reposting that as often as you like.  That won't make it relevant to the point I was making, which is that RPI is fatally flawed mathematically.  Any ranking system in which it is possible for a win to hurt your rating has no business being used.

RahOKU

Registered:
Posts: 1,203
Reply with quote  #124 
Well, I think we've talked ourselves out on this one -- and here's shocker: No one's mind was changed!

So, on the first full day of post-season play, I say good luck to everyone. Here's hoping that there's some great softball played today (unlike last night in SLC), and that every team plays to and beyond its ability and that no game is decided by a bum call, bloop double off the fists or some other unfortunate happenstance. 

We've arrived at the point where nothing matters but winning on the field, going head to head with another team and hitting better, fielding better and pitching better than the other guy. We can stow talk of RPI, WP, OWP, OOWP and all the rest of it until next February. For that I'm glad. Some reminders of what our teams are playing for:

https://ouathletics.exposure.co/sooners-at-the-wcws/photos/2809497

https://ouathletics.exposure.co/sooners-at-the-wcws/photos/2809588

https://ouathletics.exposure.co/sooners-at-the-wcws/photos/2809584




Southie

Registered:
Posts: 53
Reply with quote  #125 
Gasso is just upset because they are not a Top 8 national seed and may have to leave the state of Oklahoma for tourney play.  She knows that if they make it to OKC (huge home team advantage), they have as good a shot as anyone to win it all.  But, having to possibly go to Auburn for supers takes away their normal home field advantage.
3leftturns

Registered:
Posts: 11,255
Reply with quote  #126 
Quote:
Originally Posted by eeyore


You can keep reposting that as often as you like.  That won't make it relevant to the point I was making, which is that RPI is fatally flawed mathematically.  Any ranking system in which it is possible for a win to hurt your rating has no business being used.
If a win is against Suzy's t-ball team, who never should be scheduled by a team hoping to make the field of 64, then yes, they should be anything but rewarded.

It seems you love math, but may not actually watch much softball.

Also, in the long run, after 2700 games have been entered to compute a team's OWP (as opposed to only 53 for a much more easily moveable WP), how does one series against a 250RPI team hurt you, assuming you win?
AleDawg

Registered:
Posts: 401
Reply with quote  #127 

Quote:
Originally Posted by Southie
Gasso is just upset because they are not a Top 8 national seed and may have to leave the state of Oklahoma for tourney play.  She knows that if they make it to OKC (huge home team advantage), they have as good a shot as anyone to win it all.  But, having to possibly go to Auburn for supers takes away their normal home field advantage.


This is what's ironic about this. A Sooner fan complaining that the RPI is "rigged" in favor of the SEC and PAC when one could also say that the WCWS location is "rigged" in favor of teams from Oklahoma. Hahaha.

Silly points both.

The committee was wrong, not the RPI. People need to stop conflating their dislike of one for the other.


__________________
"Never argue with a fool, they will lower you to their level and then beat you with experience."
CajunAmos

Registered:
Posts: 1,012
Reply with quote  #128 
Quote:
Originally Posted by Bama_CF



Really?? Well the actual RPI sure says otherwise. The group of teams ranked at the top are all among the teams that played the most games against the top 25. AND the order of their ranking matches almost perfectly to the percentage of those games they WON. 

Florida is #1  - .762 in their 21 games vs top 25
Oregon is #2 - .727 in their 22 games vs top 25

With the exception of UCLA, the next group of teams is ordered almost perfectly in line with their winning % vs top teams. 

Then you get to Ole Miss, which played 25 games vs top 25 but winning % was .440. They are #18.

The RPI, year after year, is a fairly accurate predictor of the WCWS and how teams will do when they get there.


Hence the concern about the committee's choice to ignore those number in favor of something that fits their idea of selection.
Mangler

Registered:
Posts: 205
Reply with quote  #129 
All this whining doesn't change the fact that no one has a better idea, besides 'help my favorite team'
Current system is not perfect but it the best one going, at least to people who manage it and trying to keep everyone happy at the same time.
Nothing is going to make everyone happy.
1janiedough

Registered:
Posts: 2,495
Reply with quote  #130 
Quote:
Originally Posted by Mangler
All this whining doesn't change the fact that no one has a better idea, besides 'help my favorite team'
Current system is not perfect but it the best one going, at least to people who manage it and trying to keep everyone happy at the same time.
Nothing is going to make everyone happy.



 This right here!  Numbers don't fail...people do.[smile]
Bama_CF

Registered:
Posts: 2,277
Reply with quote  #131 
Quote:
Originally Posted by CajunAmos


Hence the concern about the committee's choice to ignore those number in favor of something that fits their idea of selection.



Yes, but it seems most of the complaining is about the RPI, when most of it should be about the committee IMO.

__________________

 

3leftturns

Registered:
Posts: 11,255
Reply with quote  #132 
The committee saying Minny was 7 on May 6 and dropping them 10 spots.... I think we have unanimity in that opinion
lovsofbal

Registered:
Posts: 1,485
Reply with quote  #133 
Quote:
Originally Posted by 1janiedough



 This right here!  Numbers don't fail...people do.[smile]


and who puts these formulas together?  People do lol
1janiedough

Registered:
Posts: 2,495
Reply with quote  #134 
Once again, the rpi, imo, is a GREAT indicator, an accurate indicator, and the best system around.  The selection committee, however, is another thing altogether AND they did NOT create the rpi so completely irrelevant.
eeyore

Registered:
Posts: 46
Reply with quote  #135 
Quote:
Originally Posted by 3leftturns
If a win is against Suzy's t-ball team, who never should be scheduled by a team hoping to make the field of 64, then yes, they should be anything but rewarded.


A win should never cause your rating to go down.  There are certainly cases in which should do little or nothing to improve it, but there should never be an instance in which a win hurts you relative to not having played the game.  The purpose of a purely mathematical rating system is to sort the information available about the teams in order to assess their relative quality.  A win cannot ever produce information that provides evidence that a team is worse than you thought it was before the game was played.

[Caveat: that's only true of ratings system that go purely by W/L record.  In a system that considers margin of victory, a close win against a bad team can produce such evidence.  However, using scoring differential creates its own problems, both conceptual and mathematical, and I've never found one that I think is as successful as the best W/L only systems.  Robin Lock's CHODR has conceptual appeal, but that's an awful lot of moving parts to get lined up in the regressions.]

There are much better rating systems out there than RPI.  None of them are, or could be, perfect, but that's not an excuse to continue using such a flawed system.

Quote:
It seems you love math, but may not actually watch much softball.


I watch a lot of sports.  I probably don't watch as much softball as a lot of people here, because it's hard to watch that much softball up here and my travel budget is largely consumed following the Gopher women's hockey team, but those other sports also use RPI as the basis for their selection processes.  It's underlying flaws remain constant regardless of sport, though the specific manifestations change depending upon the specific coefficients used by that sport and its conference structure.

But, if you want to be able to evaluate a mathematical rating system, understanding math is a prerequisite.

Quote:
Also, in the long run, after 2700 games have been entered to compute a team's OWP (as opposed to only 53 for a much more easily moveable WP), how does one series against a 250RPI team hurt you, assuming you win?


It can hurt you quite a bit.  I really wish the NCAA's ranking of teams by RPI would provide the actual values rather than just the order of teams.  Without those numbers, it's impossible to make any sort of definitive statement, and I'm not set up to calculate them myself, absent someone pointing me to a database from which I could pull the raw numbers myself.  My guess, though, is that by the time you get down into the teens, the gaps between teams are pretty narrow (that's how Ole Miss could jump five spots in one weekend) and so a couple of bad wins could have significant effects on a team's rating.

The hockey selection committee finally recognized this problem and adjusts the raw RPI values by discarding from the calculations they use to select and seed teams any win that lowers their ratings.  It's an unsatisfactory response, given that it introduces new holes in the logic, but it's a lot better than not adjusting it at all.  And you have no idea how much it pains me to be praising the hockey committee.
SoonerFan

Registered:
Posts: 249
Reply with quote  #136 

RPI is weighted:

25% WINNING PERCENTAGE
50% opponent's winning percentage
25% of those opponents' opponents winning percentage

I just think the team's winning percentage should account for more, such as 35% and the opponent' opponents winning percentage should be less, such as 15%.

I don't expect anyone from the SEC or Pac12 to agree with me. 

The SEC and the Pac12 have it good this year. And until the other teams in the Big 10 and the Big 12 can step up to match the level of play we've seen from Minn, Mich, Baylor and OU, we will not see much change. Because if you do not win a majority of those non-conference games in Feb and early March you are toast with getting a Top 8 seed. 


__________________
Tradition is just fine. Until something better comes along. 
1janiedough

Registered:
Posts: 2,495
Reply with quote  #137 
Quote:
Originally Posted by SoonerFan

RPI is weighted:

25% WINNING PERCENTAGE
50% opponent's winning percentage
25% of those opponents' opponents winning percentage

I just think the team's winning percentage should account for more, such as 35% and the opponent' opponents winning percentage should be less, such as 15%.

I don't expect anyone from the SEC or Pac12 to agree with me. 

The SEC and the Pac12 have it good this year. And until the other teams in the Big 10 and the Big 12 can step up to match the level of play we've seen from Minn, Mich, Baylor and OU, we will not see much change. Because if you do not win a majority of those non-conference games in Feb and early March you are toast with getting a Top 8 seed. 




This is precisely the thing...once the 1 or 2 team conferences start stacking, they will see the playing field even out.  It favors those who play the game consistently at a high level...just as EVERYTHING in life does.
AustinSoftball

Registered:
Posts: 115
Reply with quote  #138 
OU may simply be the victim of some members of the committee wanting to put the Myers softball operation to a stern test early on.  It is not the first time it has happened.
eeyore

Registered:
Posts: 46
Reply with quote  #139 
Quote:
Originally Posted by Mangler
All this whining doesn't change the fact that no one has a better idea, besides 'help my favorite team'


Actually, there are a lot of better ideas out there.  Many ratings systems superior to RPI exist.  Of the ones where I'm familiar enough with the mathematics involved, I prefer KRACH, but there are plenty of alternatives that I don't like as much but are still much better than RPI.

As for whether it would help my team, I honestly don't know the answer to that.  In the iteration of RPI that I'm most familiar with, Minnesota softball is exactly the kind of team that gets an undeserved boost from RPI relative to better systems: a good team that plays in a weak conference.  However, soft ball has different coefficients for the RPI elements, and a completely different conference structure; I'm used to using a higher coefficient on your own winning percentage and a lower one on OWP, in a structure that has 2-3 roughly equally strong conferences and 1-2 significantly weaker ones, which is pretty much the opposite of softball, with 2-3 very strong conferences and a lot of weaker ones.

So, my suspicion is that the Gophers would do worse using KRACH than RPI, but that's nothing more than a weak hunch.

The whole bit about counting top 25 wins to decide who hosts is simply indefensible.  The issues with that are both more obvious than the ones with RPI, and more severe.

Quote:
Current system is not perfect but it the best one going, at least to people who manage it and trying to keep everyone happy at the same time.
Nothing is going to make everyone happy.


The problem is that there's a lot of evidence that suggests that the people who manage the process don't understand the tools that they are using.  That's never a good set up.
1janiedough

Registered:
Posts: 2,495
Reply with quote  #140 
Quote:
Originally Posted by AustinSoftball
OU may simply be the victim of some members of the committee wanting to put the Myers softball operation to a stern test early on.  It is not the first time it has happened.



I'm about sick of that douchenozzle Clint Myers.  He is an embarrassment to my home state of Arizona, as is his douchebag son Cory...and even Casey for being thought of as "weird".  Wtf does that even mean?  Inbred?
eeyore

Registered:
Posts: 46
Reply with quote  #141 
Quote:
Originally Posted by SoonerFan
RPI is weighted:

25% WINNING PERCENTAGE
50% opponent's winning percentage
25% of those opponents' opponents winning percentage

I just think the team's winning percentage should account for more, such as 35% and the opponent' opponents winning percentage should be less, such as 15%.



The flaws of RPI run too deep to be fixed by this.  Changing the coefficients merely changes the relative degree of the various problems.  The problem with increasing the weight of winning percentage is that you increase the chances that a good team in a weak conference ends up looking like a great team.  There is  no happy medium at which to set them.

Also, as a practical matter, OOWP has a much lower than 25% weighting in the calculation.  The differences between the best OOWP and the worst are much smaller than with the other two components, and so it needs a larger coefficient in order to have the same weight.  Again, without the raw numbers, no one can say anything for certain, but my guess is that OOWP constitutes well under 10% of the overall differences in RPI values.

rudymartinez

Registered:
Posts: 386
Reply with quote  #142 
Douche nozzle. Will have to add that to my insult list. Thanks.
3leftturns

Registered:
Posts: 11,255
Reply with quote  #143 

Quote:
Originally Posted by eeyore

In the iteration of RPI that I'm most familiar with, Minnesota softball is exactly the kind of team that gets an undeserved boost from RPI relative to better systems: a good team that plays in a weak conference.
Minnesota got "an undeserved boost from RPI"?

Also the KRACH site terms RPI as 51 percent OOWP.


 

eeyore

Registered:
Posts: 46
Reply with quote  #144 
Quote:
Originally Posted by 3leftturns
Minnesota got "an undeserved boost from RPI"?


Also the KRACH site terms RPI as 51 percent OOWP.



You really didn't read my whole comment, did you?

ChinMusic

Registered:
Posts: 562
Reply with quote  #145 
Quote:
Originally Posted by AustinSoftball
OU may simply be the victim of some members of the committee wanting to put the Myers softball operation to a stern test early on.  It is not the first time it has happened.


That's the root of everyone's concerns.  "...members of the committee wanting to ...."  They shouldn't have any agenda.
3leftturns

Registered:
Posts: 11,255
Reply with quote  #146 
Quote:
Originally Posted by eeyore


You really didn't read my whole comment, did you?
No, mea culpa.

Then, why are are crying about the softball RPI?
3leftturns

Registered:
Posts: 11,255
Reply with quote  #147 
The RPI has done an excellent job for softball. What are these RPI screw jobs that everyone intimates?

Who is the worst RPI to win the tournament?

Of course it isn't perfect. NOTHING that is human-programmed could be where there is even a scintilla of subjectivity involved

The thing that was done wrong this year was the idiot committee saying Minnesota was 7th (crazy high) and then dropping them 10 spots in 8 days. Just have no clue why, if Minnesota was 17 in their eyes, the Gophers were in that reveal.
eeyore

Registered:
Posts: 46
Reply with quote  #148 
Quote:
Originally Posted by 3leftturns
No, mea culpa.

Then, why are are crying about the softball RPI?


I'd explain it again, but all evidence suggests that you'd ignore it.
3leftturns

Registered:
Posts: 11,255
Reply with quote  #149 
All evidence suggests that the RPI has done an excellent job in softball, so yes. I don't let the pursuit of perfect be the enemy of excellent

You wanna fix hockey, the sport most dependent on luck to begin with, more power to you

But, it has been a fun couple of days with you registered. You can go back to your original handle. Perhaps pull your head off the calculator and watch some softball while you switch back
gonegolfin

Registered:
Posts: 353
Reply with quote  #150 
Quote:
Originally Posted by SoonerFan

RPI is weighted:

25% WINNING PERCENTAGE
50% opponent's winning percentage
25% of those opponents' opponents winning percentage

I just think the team's winning percentage should account for more, such as 35% and the opponent' opponents winning percentage should be less, such as 15%.

I don't expect anyone from the SEC or Pac12 to agree with me. 

The SEC and the Pac12 have it good this year. And until the other teams in the Big 10 and the Big 12 can step up to match the level of play we've seen from Minn, Mich, Baylor and OU, we will not see much change. Because if you do not win a majority of those non-conference games in Feb and early March you are toast with getting a Top 8 seed. 



Certainly winning is under emphasized in the RPI (just over 25%) and SOS is over emphasized (just over 75%) ... one of the principal complaints I have been voicing for years. But modifying the weights will not solve the problems. A core problem is that two levels of interconnectivity is simply not enough to capture real SOS. I have provided many examples over the years supporting why. The problem with RPI is structural and cannot be fixed with tweaks.

Brian
Previous Topic | Next Topic
Print
Reply

Quick Navigation:

Easily create a Forum Website with Website Toolbox.