Games of Importance for UC's Resume

[sedziobs]

Wednesday Jan 24th 2018

South Carolina @ Florida
7pm SEC Network
Florida looking like a safe Group 1 loss, at RPI 26. They would need to go 5-7 or worse to drop out of the top 50. 82% chance to win.

Nevada @ Wyoming
11pm ESPNU
Wyoming at RPI 74. Need to finish 7-5 to stay top 100. 24% chance to win.

Upsets in both games. Florida falls to 40 with the loss. Wyoming wins in double OT and improves to 58.

 

[waterhead]

Upsets in both games. Florida falls to 40 with the loss. Wyoming wins in double OT and improves to 58.

That is a big drop for FLA. An equally nice jump for Wyoming.

It actually wouldn't bother me too much if FLA falls to group 2 (not that I think they will). I am not rooting for it but we lost to them and our record against group 1 would look better as a result. It would not be a bad loss and our record against group 2 should be very solid anyway. I know it affects SOS a little bit...but just sayin...not the end of the world if it happens.

I think Cincrulz11 brought this subject up. What does the committee see as more important in groups (columns)? Win loss %? Or total wins and total chances in that group?

 

[waterhead]

That is a big drop for FLA. An equally nice jump for Wyoming.

It actually wouldn't bother me too much if FLA falls to group 2 (not that I think they will). I am not rooting for it but we lost to them and our record against group 1 would look better as a result. It would not be a bad loss and our record against group 2 should be very solid anyway. I know it affects SOS a little bit...but just sayin...not the end of the world if it happens.

I think Cincrulz11 brought this subject up. What does the committee see as more important in groups (columns)? Win loss %? Or total wins and total chances in that group?

Let me give an example

Team A has a record of 4-2 in group 1 66%

Team B has a record of 5-5 in group 1 50%

Which is better? Edit...let's assume the general ranking of the teams they play in RPI is roughly the same.

 

[cincrulz11]

Let me give an example

Team A has a record of 4-2 in group 1 66%

Team B has a record of 5-5 in group 1 50%

Which is better? Edit...let's assume the general ranking of the teams they play in RPI is roughly the same.

yeah its a thought that crossed my mind. of course it probably isn't better to lose to a group 2 team than it is to lose to a group 1 team.


however i could see a scenario where a team that is 4-1 vs group 1 could be talked about as being "better" than a team that is 4-2 vs group 1 cause they have a higher winning percentage vs group 1.


just never know how it will look to committee members.

 

[waterhead]

yeah its a thought that crossed my mind. of course it probably isn't better to lose to a group 2 team than it is to lose to a group 1 team.


however i could see a scenario where a team that is 4-1 vs group 1 could be talked about as being "better" than a team that is 4-2 vs group 1 cause they have a higher winning percentage vs group 1.


just never know how it will look to committee members.

I posted something about one committee member answering questions a while back but forget the thread.

He said they consider total wins and losses in group 1. Record % in that group. How many were away from home. They don't really dock too much for bad losses unless you have a few. He said they don't consider who your coach is or the prestige of your conference other than it's strength. They don't consider past tourney results etc. He claims they basically erase the chalkboard and there is no preferential treatment.

I just don't know how much weight they put on win % vs totals in groups.

 

[sedziobs]

If you're going to do this comparison, you need to include the Group 2 record as well. The committee will see both together. How about a blind resume test:

Team A: 3-1 Group 1, 1-2 Group 2, 1 bad loss
Team B: 4-4 Group 1, 0-0 Group 2, 1 bad loss

Who would you take?

 

[waterhead]

If you're going to do this comparison, you need to include the Group 2 record as well. The committee will see both together. How about a blind resume test:

Team A: 3-1 Group 1, 1-2 Group 2, 1 bad loss
Team B: 4-4 Group 1, 0-0 Group 2, 1 bad loss

Who would you take?

I don't need more hypotheticals. I need answers damnit

 

[sedziobs]

Lol. Team A is Houston (11 seed on bracketmatrix). Team B is Seton Hall (5 seed on bracketmatrix).

That would indicate that Florida staying in Group 1 is preferred.

 

[waterhead]

Lol. Team A is Houston (11 seed on bracketmatrix). Team B is Seton Hall (5 seed on bracketmatrix).

That would indicate that Florida staying in Group 1 is preferred.

Thanks. This season the RPI has not played out all the way yet.

I was curious about past ncaa bubble results but did not want to do the research. A team like Maryland last year would be an interesting subject. Maybe some other bubble teams? I am not asking you to do the research either but maybe someone knows off hand.

 

[sedziobs]

Looking at past years for evidence is tougher now since the Group system is new this year.

 

[cincrulz11]

Lol. Team A is Houston (11 seed on bracketmatrix). Team B is Seton Hall (5 seed on bracketmatrix).

That would indicate that Florida staying in Group 1 is preferred.

im not sure that conclusion can be drawn just from the two things listed though. houston is rocking out 7 250+ wins to seton halls 2 which is dragging their rpi down tremendously.


i would say their resumes are far too different to know for sure the gap is from group 1 and group 2 games.





i mean your point could still be right, but im not sure we can get there from that data.

edit: also thats 4 more group 1 games which is gonna help the SOS and all that.

 

[waterhead]

Looking at past years for evidence is tougher now since the Group system is new this year.

Why? All they did was change the parameters for the groups. They didn't change the criteria for how they view the groups.

 

[sedziobs]

cincrulz, I agree. My admittedly cherry-picked example wasn't meant as proof, but rather as a counter-example to the idea that Group 1 win percentage outweighs Group 2 losses. Any resume can be dissected further. That true of Houston, Seton Hall and UC.

 

[sedziobs]

Why? All they did was change the parameters for the groups. They didn't change the criteria for how they view the groups.

Previously, it was up to each person to consider how to rate a game that was home, away or neutral. Very subjective. The new Groups do that automatically. It's much easier to compare resumes now with only Group records. You can't come close to comparing simple top 50 RPI and 51-100 records without seeing where those games were played.

 

[waterhead]

Previously, it was up to each person to consider how to rate a game that was home, away or neutral. Very subjective. The new Groups do that automatically. It's much easier to compare resumes now with only Group records. You can't come close to comparing simple top 50 RPI and 51-100 records without seeing where those games were played.

Sure. But the new groups do not tell us whether win loss % is better or straight up quantity is better within a group

 

[sedziobs]

Sure. But the new groups do not tell us whether win loss % is better or straight up quantity is better within a group

Ugh. We're going in circles again. I'm saying that it's difficult to go back to previous years to research that question as you suggested because they didn't use the current Group system in previous years. Previously simple top 50 record was not a good resume indicator - it had to be considered alongside road wins. That's no longer the case. Now, Group 1 and 2 records alone are a much better proxy for a team's quality wins than the old system.

I used the Houston/Seton Hall example because there is a consensus that Seton Hall has a better resume than Houston. Considering only Group records (the point of this conversation), Seton Hall would look a lot like Houston if you moved a couple of their Group 1 wins down to Group 2. So the thought process is that it generally hurts your resume to move losses from Group 1 to Group 2, if those are the only two things you're looking at. That's my opinion and one piece of evidence. It is not proof.

 

[waterhead]

Ugh. We're going in circles again. I'm saying that it's difficult to go back to previous years to research that question as you suggested because they didn't use the current Group system in previous years. Previously simple top 50 record was not a good resume indicator - it had to be considered alongside road wins. That's no longer the case. Now, Group 1 and 2 records alone are a much better proxy for a team's quality wins than the old system.

I used the Houston/Seton Hall example because there is a consensus that Seton Hall has a better resume than Houston. Seton Hall would look a lot like Houston if you moved a couple of their Group 1 wins down to Group 2. So the thought process is that it generally hurts your resume to move losses from Group 1 to Group 2, if those are the only two things you're looking at. That's my opinion and one piece of evidence. It is not proof.

I appreciate the feedback. I am just wondering if Bracket Matrix is predicting future results in RPI. Of course it is. How else can we explain some of the seeding predictions?

Bracket Matrix probably considers the committee tendencies by being an aggregate. Some of the tendencies can maybe be eliminated...but there is still the human element I am trying to understand.

 

[Jon Rycek]

I'm just gonna by T-Rank from now on, they have us as a 2 lol... what's with the brodd having us as a 7. Haha

 

[MacBearcat]

Who does UC want tonight in the UCF vs WSU game? I'd assume a win for WSU so that if/when we beat them they are higher ranked?

 

[sedziobs]

Bracketmatrix is crowdsourced from people with all kinds of methodologies. Most of them don't list any reasoning whatsoever, so we have no idea what they're really doing. In general it's a decent approximation.