Bracketology and Team Resumes

[Uspshoosier]

6 minutes ago, brumdog45 said:

I think it would be interesting to see how they would compare in the system that I mentioned above.  Under the system I suggest, a SOS of 50 v an SOS of 70 isn't that meaningful when it doesn't actually mean the SOS of 50 should produce more wins for the reason I stated.  It could be that Mississippi might actually have a higher net win comparison despite a lower SOS because the combination of teams they play actually produces a lower win expectancy.

For me I’m fine with the system they are using.  I was fine with RPI and I’m fine with the NET.   Aside from some bubble team selections every year I think they do a fantastic job.   I might disagree with why they selected a certain team but most of the times I can see their point of view.  I think outside people make a bigger deal of the Quad system then it really is.   Those 10 committee members when comparing resumes  can see that a win at Baylor is a Q1a might be grouped in the same Quad as a win at a team that is 75 but clearly the Baylor win would hold a significant amount more weight.   They have had several interviews stating this  point 

[Southside]

MSU keeping it interesting vs scUM. 

And not to long ago I thought IU would be in, and MSU would miss out... lol

[RaceToTheTop]

2 hours ago, Uspshoosier said:

For me I’m fine with the system they are using.  I was fine with RPI and I’m fine with the NET.   Aside from some bubble team selections every year I think they do a fantastic job.   I might disagree with why they selected a certain team but most of the times I can see their point of view.  I think outside people make a bigger deal of the Quad system then it really is.   Those 10 committee members when comparing resumes  can see that a win at Baylor is a Q1a might be grouped in the same Quad as a win at a team that is 75 but clearly the Baylor win would hold a significant amount more weight.   They have had several interviews stating this  point 

You know that you and I are on different opinions on that......but one thing I don't understand is how you could be fine with RPI and fine with NET.  The two systems are giving completely different results.  

Here are three systems ranking of Big Ten teams (sans Nebraska).  The systems team rankings are listed in the same order for each.  Which of the three do you field most reflects the order the committee would seed them?

Big Ten: system 1/2/3

Michigan 3/3/1

Illinois 4/17/4

Iowa 6/34/14

Ohio State 9/24/7

Purdue 20/25/11

Wisconsin 25/77/29

Maryland 32/61/44

Rutgers 37/65/30

Penn State 48/116/63

Indiana 60/109/56

Michigan State 72/78/45

Minnesota 78/135/70

Northwestern 89/176/86

 

[Uspshoosier]

3 minutes ago, brumdog45 said:

You know that you and I are on different opinions on that......but one thing I don't understand is how you could be fine with RPI and fine with NET.  The two systems are giving completely different results.  

Here are three systems ranking of Big Ten teams (sans Nebraska).  The systems team rankings are listed in the same order for each.  Which of the three do you field most reflects the order the committee would seed them?

Big Ten: system 1/2/3

Michigan 3/3/1

Illinois 4/17/4

Iowa 6/34/14

Ohio State 9/24/7

Purdue 20/25/11

Wisconsin 25/77/29

Maryland 32/61/44

Rutgers 37/65/30

Penn State 48/116/63

Indiana 60/109/56

Michigan State 72/78/45

Minnesota 78/135/70

Northwestern 89/176/86

 

I would use whichever metric the committee is using at the time of the selection process.  Plus I would need more information on each of the teams team sheet before I could project what the committee would do 

[RaceToTheTop]

2 minutes ago, Uspshoosier said:

I would use whichever metric the committee is using at the time of the selection process.  Plus I would need more information on each of the teams team sheet before I could project what the committee would do 

That sounds like circular logic to me -- that you would 'use whatever metric the committee is using at the time' is basically just saying that 'I would use the metric the committee was using becaue they are using it.'  Let me state this another way:  which of the three systems do YOU feel would best be used on its own to seed teams?  Because all of the systems can view the 'more information' afterwards that you are referring to.

I don't think it's very difficult to admit that the middle system is giving ridiculous results and in no way reflect the strength of teams.  Illinois is at 17, Wisconsin at 77......if there is enough 'extra information that would warrant moving Illinois up to a 1 or 2 seed enough information to move Wisky up to where most people project them to be seeded, then the middle system is useless because you are pretty much going to ignore the ratings they provide any way.  The middle system suggests the Big Ten has 5 seedable teams.

 

[Uspshoosier]

2 minutes ago, brumdog45 said:

That sounds like circular logic to me -- that you would 'use whatever metric the committee is using at the time' is basically just saying that 'I would use the metric the committee was using becaue they are using it.'  Let me state this another way:  which of the three systems do YOU feel would best be used on its own to seed teams?  Because all of the systems can view the 'more information' afterwards that you are referring to.

I don't think it's very difficult to admit that the middle system is giving ridiculous results and in no way reflect the strength of teams.  Illinois is at 17, Wisconsin at 77......if there is enough 'extra information that would warrant moving Illinois up to a 1 or 2 seed enough information to move Wisky up to where most people project them to be seeded, then the middle system is useless because you are pretty much going to ignore the ratings they provide any way.  The middle system suggests the Big Ten has 5 seedable teams.

 

Since no one system is used on its own to seed teams I wouldn’t chose any.   

[Hovadipo]

4 minutes ago, brumdog45 said:

That sounds like circular logic to me -- that you would 'use whatever metric the committee is using at the time' is basically just saying that 'I would use the metric the committee was using becaue they are using it.' 

He (and I’m assuming most other bracketologists) uses the set of criteria the committee uses because he’s trying to predict what the committee will do. I really don’t see what there is to debate. He doesn’t care whether the rating systems suck or not. It’s what the committee does, he’s trying to nail down what they’re going to do, so he uses what they use. 

[RaceToTheTop]

3 minutes ago, Uspshoosier said:

Since no one system is used on its own to seed teams I wouldn’t chose any.   

Then why would the NCAA have ever bothered to get rid of the RPI if it wasn't the only system used?  Why would the NCAA bother to tweak the NET rankings?  Would they not tell you that their current NET system is better than their previous?  I'm sorry, I love your work here, but I think you are really begging the question.  

[RaceToTheTop]

1 minute ago, Hovadipo said:

He (and I’m assuming most other bracketologists) uses the set of criteria the committee uses because he’s trying to predict what the committee will do. I really don’t see what there is to debate. He doesn’t care whether the rating systems suck or not. It’s what the committee does, he’s trying to nail down what they’re going to do, so he uses what they use. 

I get that, which is why I reframed my question as to what system he felt properly ordered the teams best and not what he thought the committee would do.

[Hovadipo]

3 minutes ago, brumdog45 said:

I get that, which is why I reframed my question as to what system he felt properly ordered the teams best and not what he thought the committee would do.

Why does it need to be narrowed down to 1 system? I know next to nothing about bracketology, but I’m guessing there is something useful to be pulled from each of the systems. 

[RaceToTheTop]

1 minute ago, Hovadipo said:

Why does it need to be narrowed down to 1 system? I know next to nothing about bracketology, but I’m guessing there is something useful to be pulled from each of the systems. 

It doesn't have to be narrowed down to one system.  But don't we want to have the systems that are used to be as reflective as they can be?  It's obvious that the NCAA feels that way as they switched from RPI to NET to a tweaked NET this year.  

Where I completely disagree with USPS is on the RPI system.  I think it was a horrible, useless system that produced completely inaccurate results.  NET is better, and the newer NET is better than the previous.  But I still think they are missing the mark.  And I think there are still ways that teams can game the system.

[Uspshoosier]

9 minutes ago, brumdog45 said:

Then why would the NCAA have ever bothered to get rid of the RPI if it wasn't the only system used?  Why would the NCAA bother to tweak the NET rankings?  Would they not tell you that their current NET system is better than their previous?  I'm sorry, I love your work here, but I think you are really begging the question.  

They switched because coaches wanted something different.   The RPI was easily manipulated by teams so they asked for it to change.  Ncaa had a meeting with  with a bunch of smart people and even held a meeting with the metric gurus and came up with a new system based on all the information they gathered from those meetings.   They came up with a system that they thought was a balance of predictive and results based.    Nothing perfect but metric gurus and national bracketologist all agreed it was an improvement on the RPI.   I personally don’t care which is better for seeding.  It’s not that deep to me.  I go by what the committee uses and try my best to guess what they wilL do 

[WayneFleekHoosier]

4 minutes ago, brumdog45 said:

It doesn't have to be narrowed down to one system.  But don't we want to have the systems that are used to be as reflective as they can be?  It's obvious that the NCAA feels that way as they switched from RPI to NET to a tweaked NET this year.  

Where I completely disagree with USPS is on the RPI system.  I think it was a horrible, useless system that produced completely inaccurate results.  NET is better, and the newer NET is better than the previous.  But I still think they are missing the mark.  And I think there are still ways that teams can game the system.

Yep, which is why the eye test and game knowledge should/will always be a factor.  Question is whose eye test do can be trusted?  

Edited March 8, 2021 by WayneFleekHoosier

[Uspshoosier]

3 minutes ago, brumdog45 said:

It doesn't have to be narrowed down to one system.  But don't we want to have the systems that are used to be as reflective as they can be?  It's obvious that the NCAA feels that way as they switched from RPI to NET to a tweaked NET this year.  

Where I completely disagree with USPS is on the RPI system.  I think it was a horrible, useless system that produced completely inaccurate results.  NET is better, and the newer NET is better than the previous.  But I still think they are missing the mark.  And I think there are still ways that teams can game the system.

I agree RPI sucked balls but that’s what they used so I thats what I used.    I always said the RPI was terrible when people screamed at me to use Ken Pom and other metics during the RPI years.   Totally agree NET is better and can improve 

[Uspshoosier]

7 minutes ago, WayneFleekHoosier said:

Yep, which is why the eye test and game knowledge should/will always be a factor.  Question is whose eye test do can be trusted?  

I have no problem with people using the eye test argument as long as they actually watch them games instead of box scores. A lot of articles written nationally about why the eye test shouldn’t be used 

[RaceToTheTop]

1 minute ago, Uspshoosier said:

I agree RPI sucked balls but that’s what they used so I thats what I used.    I always said the RPI was terrible when people screamed at me to use Ken Pom and other metics during the RPI years.   Totally agree NET is better and can improve 

That's basically my point.  I had mentioned previously that I thought there were some errors in the way that the NET calculations were being made which is why I came up with a very rudimentary system using nothing more than the data that is used in the NET calculations -- there was no intention of moving certain teams up or down that I felt were over or under ranked.  And the numbers that I'm seeing my raw system produces I think just might be closer to how the committee actually seeds teams than their NET system does.  While obviously the NET rankings are 100% not their end all, be all, I believe my system using their numbers for home, road, quad wins/losses is closer to reflecting what they think than the NET is.  I'll check it after they seed teams.

[Uspshoosier]

8 minutes ago, WayneFleekHoosier said:

Yep, which is why the eye test and game knowledge should/will always be a factor.  Question is whose eye test do can be trusted?  

Chris Reynolds is one of those people who will be deciding teams fates 

[Uspshoosier]

You seem like a numbers guy so here are some other metrics I enjoy looking at but when it comes to projecting the field I don’t use

https://barttorvik.com/#

https://haslametrics.com/ratings.php

[RaceToTheTop]

27 minutes ago, Uspshoosier said:

You seem like a numbers guy so here are some other metrics I enjoy looking at but when it comes to projecting the field I don’t use

https://barttorvik.com/#

https://haslametrics.com/ratings.php

Interesting stuff.  One common theme pretty much in all rating systems:  they don't like Michigan State very well (obviously these ratings were prior to today's game, but one game won't change the rating that much).  I can say that every system really likes Houston.....except my system.  My system has Houston sitting at #25 (likely will move up about 4 spots after today's win over Memphis).  On the opposite end, my system really loves West Virginia.

[Uspshoosier]

Initial ballot results: 25 teams voted into the field, 46 teams on the under consideration board (includes regular-season champs). Committee and staff just finished COVID testing and soon will eat dinner.

david worlock