View previous topic :: View next topic |
Author |
Message |
myrtle
Joined: 02 May 2008 Posts: 32341
Back to top |
Posted: 02/11/19 8:09 pm ::: top 16 reveal |
Reply |
|
ok did anyone take a screen shot or have who went where and why Charlie was grumbling about it?
_________________ For there is always light,
if only we’re brave enough to see it.
If only we’re brave enough to be it.
- Amanda Gorman
|
|
myrtle
Joined: 02 May 2008 Posts: 32341
Back to top |
Posted: 02/11/19 8:28 pm ::: |
Reply |
|
anybody?
IMO there are six top teams and a big gap to the next bunch so whichever regions has two of those teams is obviously going to be 'tougher' than the other two. Now the two 'easier' regions should have IMO the toughest 3 & 4 seeds.
_________________ For there is always light,
if only we’re brave enough to see it.
If only we’re brave enough to be it.
- Amanda Gorman
|
|
pilight
Joined: 23 Sep 2004 Posts: 67052 Location: Where the action is
Back to top |
|
myrtle
Joined: 02 May 2008 Posts: 32341
Back to top |
Posted: 02/11/19 8:31 pm ::: |
Reply |
|
Well, Portland is clearly the cupcakiest...
If NC State ends up as a two seed I'll eat my hat.
Maryland is the likely three to replace them but they would of course much prefer to play in Greensboro as a three rather than Portland as a two.
_________________ For there is always light,
if only we’re brave enough to see it.
If only we’re brave enough to be it.
- Amanda Gorman
|
|
Nixtreefan
Joined: 14 Nov 2012 Posts: 2539
Back to top |
|
Marquette Fan
Joined: 06 Mar 2005 Posts: 3582
Back to top |
|
Nixtreefan
Joined: 14 Nov 2012 Posts: 2539
Back to top |
Posted: 02/11/19 10:03 pm ::: |
Reply |
|
I don't think it will just be NC st, but hey you never know after looking at this.
|
|
CBiebel
Joined: 23 Dec 2004 Posts: 1058 Location: PA
Back to top |
Posted: 02/11/19 10:18 pm ::: |
Reply |
|
Some weird stuff there. Miami is there (#20 AP #20 RPI), yet Syracuse (#16 AP, #9 RPI) isn't?
|
|
myrtle
Joined: 02 May 2008 Posts: 32341
Back to top |
Posted: 02/11/19 10:21 pm ::: |
Reply |
|
CBiebel wrote: |
Some weird stuff there. Miami is there (#20 AP #20 RPI), yet Syracuse (#16 AP, #9 RPI) isn't? |
yeah, I did a double take on that as well. I don't think the 4 seeds are set at all. I think the 1 - 3 seeds are all likely to at least get one of the top 16s...except NC State. It would surprise me if they hang on to a #4 but they could surprise me and win a couple more games than I think they will here in the home stretch.
_________________ For there is always light,
if only we’re brave enough to see it.
If only we’re brave enough to be it.
- Amanda Gorman
|
|
patsweetpat
Joined: 14 Jul 2010 Posts: 2313 Location: Culver City, CA
Back to top |
Posted: 02/11/19 10:39 pm ::: |
Reply |
|
CBiebel wrote: |
Some weird stuff there. Miami is there (#20 AP #20 RPI), yet Syracuse (#16 AP, #9 RPI) isn't? |
Pretty sure they don't look at AP ratings. So that's immaterial. But Miami has two wins over Top 10 RPI teams, while Syracuse's very best win is over #20 Texas A&M. Plus Miami beat Syracuse head-to-head.
I think it's a pretty defensible choice.
But I genuinely don't care either way. Whatevs.
|
|
Howee
Joined: 27 Nov 2009 Posts: 15754 Location: OREGON (in my heart)
Back to top |
Posted: 02/11/19 10:57 pm ::: |
Reply |
|
Now, obviously the lines were drawn before UConn's blowout of SC tonight, but the committee isn't stupid...(are they?) Tell me, statisticians, how MS-State got the 1 seed over UConn. I take it they've considered the SOS/RPI/etc in their seedings. My (more irregular) thinking is MS-State and UConn deserve the #1s, and Baylor goes to a #2, all things considered....no?
_________________ Oregon: Go Ducks!
"Inévitablement, les canards voleront"
|
|
pilight
Joined: 23 Sep 2004 Posts: 67052 Location: Where the action is
Back to top |
Posted: 02/11/19 11:13 pm ::: |
Reply |
|
Howee wrote: |
Now, obviously the lines were drawn before UConn's blowout of SC tonight, but the committee isn't stupid...(are they?) Tell me, statisticians, how MS-State got the 1 seed over UConn. I take it they've considered the SOS/RPI/etc in their seedings. My (more irregular) thinking is MS-State and UConn deserve the #1s, and Baylor goes to a #2, all things considered....no? |
Coming into tonight, Miss St had more Quadrant One wins and more Quadrant Two wins. Most of UConn's games have been against sub-100 RPI opponents (14 out of 23 before tonight's game) whil most of Miss St's have been against top 100 RPI teams (12 out of 23). Plus UConn has two losses compared to Miss St's one and Miss St was ahead of UConn in RPI.
_________________ I'm a lonely frog
I ain't got a home
|
|
Howee
Joined: 27 Nov 2009 Posts: 15754 Location: OREGON (in my heart)
Back to top |
Posted: 02/11/19 11:22 pm ::: |
Reply |
|
pilight wrote: |
Howee wrote: |
Now, obviously the lines were drawn before UConn's blowout of SC tonight, but the committee isn't stupid...(are they?) Tell me, statisticians, how MS-State got the 1 seed over UConn. I take it they've considered the SOS/RPI/etc in their seedings. My (more irregular) thinking is MS-State and UConn deserve the #1s, and Baylor goes to a #2, all things considered....no? |
Coming into tonight, Miss St had more Quadrant One wins and more Quadrant Two wins. Most of UConn's games have been against sub-100 RPI opponents (14 out of 23 before tonight's game) whil most of Miss St's have been against top 100 RPI teams (12 out of 23). Plus UConn has two losses compared to Miss St's one and Miss St was ahead of UConn in RPI. |
Okay. I get that. But--using the same metric--how do Baylor and UConn then compare?
_________________ Oregon: Go Ducks!
"Inévitablement, les canards voleront"
|
|
pilight
Joined: 23 Sep 2004 Posts: 67052 Location: Where the action is
Back to top |
Posted: 02/11/19 11:28 pm ::: |
Reply |
|
Howee wrote: |
pilight wrote: |
Howee wrote: |
Now, obviously the lines were drawn before UConn's blowout of SC tonight, but the committee isn't stupid...(are they?) Tell me, statisticians, how MS-State got the 1 seed over UConn. I take it they've considered the SOS/RPI/etc in their seedings. My (more irregular) thinking is MS-State and UConn deserve the #1s, and Baylor goes to a #2, all things considered....no? |
Coming into tonight, Miss St had more Quadrant One wins and more Quadrant Two wins. Most of UConn's games have been against sub-100 RPI opponents (14 out of 23 before tonight's game) whil most of Miss St's have been against top 100 RPI teams (12 out of 23). Plus UConn has two losses compared to Miss St's one and Miss St was ahead of UConn in RPI. |
Okay. I get that. But--using the same metric--how do Baylor and UConn then compare? |
Baylor has three more quadrant one wins, has most of their games against the top 100, has only one loss, has a better RPI, and won head-to-head.
_________________ I'm a lonely frog
I ain't got a home
|
|
CBiebel
Joined: 23 Dec 2004 Posts: 1058 Location: PA
Back to top |
Posted: 02/12/19 12:11 am ::: |
Reply |
|
myrtle wrote: |
CBiebel wrote: |
Some weird stuff there. Miami is there (#20 AP #20 RPI), yet Syracuse (#16 AP, #9 RPI) isn't? |
I think the 1 - 3 seeds are all likely to at least get one of the top 16s...except NC State. It would surprise me if they hang on to a #4 but they could surprise me and win a couple more games than I think they will here in the home stretch. |
Considering that their remaining schedule consists of:
at #16 Syracuse
#6 ND
at UNC (who beat them 64-51 in Raleigh)
at #2 Louisville
#20 Miami
Oh, and at Wake Forest.
That sounds like they will be lucky to win 3 of the remaining 6 games.
|
|
Howee
Joined: 27 Nov 2009 Posts: 15754 Location: OREGON (in my heart)
Back to top |
Posted: 02/12/19 12:20 am ::: |
Reply |
|
Thanks, Pilight.
_________________ Oregon: Go Ducks!
"Inévitablement, les canards voleront"
|
|
Phil
Joined: 22 Oct 2011 Posts: 1275
Back to top |
Posted: 02/12/19 10:46 am ::: |
Reply |
|
myrtle wrote: |
If NC State ends up as a two seed I'll eat my hat.
|
I agree.
|
|
Phil
Joined: 22 Oct 2011 Posts: 1275
Back to top |
Posted: 02/12/19 10:50 am ::: |
Reply |
|
[quote="patsweetpat"]
CBiebel wrote: |
But Miami has two wins over Top 10 RPI teams, while Syracuse's very best win is over #20 Texas A&M.
|
Charlie said that as well, but I count three. Am I missing something?
I agree with you that it is a very defensible selection.
|
|
SDHoops
Joined: 09 Nov 2007 Posts: 1183
Back to top |
Posted: 02/12/19 11:24 am ::: |
Reply |
|
I'm a big UConn fan...but for once, just once...especially if the Huskies are a two seed, I'd like to see them go to a region other than Albany. Ship them off and make them win on the road! Whoever gets the #1 seed for that region should earn the right to be a #1 seed and not worry about a possible home crowd for a lower seed. In the early rounds I can see it, but not for regionals!
|
|
Nixtreefan
Joined: 14 Nov 2012 Posts: 2539
Back to top |
|
SDHoops
Joined: 09 Nov 2007 Posts: 1183
Back to top |
|
Phil
Joined: 22 Oct 2011 Posts: 1275
Back to top |
|
ucbart
Joined: 21 Nov 2004 Posts: 2824 Location: New York
Back to top |
Posted: 02/12/19 5:53 pm ::: |
Reply |
|
If UCONN has a superior team, they win....almost every year.
We lost a regional to Duke in 2006 in Bridgeport. It has little to do with the crowd, IMO. It has way more to deal with who the team personnel is.
These regionals need fans and the UCONN fans buy tickets more than anyone. I may be a UCONN homer, but they should be placed where it's easiest for the fans to get to.
|
|
Phil
Joined: 22 Oct 2011 Posts: 1275
Back to top |
Posted: 02/12/19 7:02 pm ::: |
Reply |
|
But you don't need to take my word for it.
I created a quick spreadsheet which can be found here:
https://docs.google.com/spreadsheets/d/16DrOq3P9arQPMCSv1Vn9cf_x5F_jxAcjeVldrgQ64yM/edit?usp=sharing
If I did this right, I set the sharing so that anyone with a link can view and edit it.
I listed all the teams in the top 16 reveal (plus those whose RPI or Massey rating is in the top 16).
I have a column (1) which shows the ranking 1 to 16 as ranked by the committee.
Then there is column (2) which shows there ranking according to the NCAA RPI formula.
Then column (3) is the Massey rating.
There are two takeaways:
The Massey top 16 differs from the committee top 16 by only a single team. Arizona State would be in the top 16 based on Massey and Miami is the top 16 based on the committee. in contrast, the RPI list differs from the committee list in three places. Syracuse, UCF and Florida State have RPI ratings in the top 16 but did not get selected by the committee. Maryland, Oregon State and Miami were selected by the committee but are not in the top 16 RPI.
The second take away is that the Massey ranking is much closer to the committee ranking than the RPI.
A standard way of measuring distances is to look at the squared differences. The final two columns in the spreadsheet show committee versus RPI and committee versus Massey. The square differences for the committee list compared to the RPI list is 486, more than four times the size of the squared differences between the committee list and the Massey list at 113.
[Edit] I'm trying not to overcomplicate; Calculating the sum of squares is a simple calculation and very common. Some would prefer to take the square root (standard deviation) because the units are the same. Still others prefer Mean Absolute Error. I add the other two measures, but all three send the same message -- the Massey metric is far closer to the committee decision than is the RPI metric. (I always liked MAE, but the statistical properties aren't quite as nice as the Variance/Standard Deviation/Squared error measures.)
Read more here if you really want to get into the weeds:
https://stats.stackexchange.com/questions/118/why-square-the-difference-instead-of-taking-the-absolute-value-in-standard-devia
Last edited by Phil on 02/12/19 9:10 pm; edited 2 times in total |
|
GlennMacGrady
Joined: 03 Jan 2005 Posts: 8249 Location: Heisenberg
Back to top |
Posted: 02/12/19 7:45 pm ::: |
Reply |
|
Phil wrote: |
A standard way of measuring distances is to look at the square differences. |
Phil, just curious as to the rationale for using cumulative square differences rather than just cumulative differences. The point made would be the same, except that (Committee-RPI) would look about twice as great as (Committee-Massey) instead of four times as great.
Are square differences a standard method of making comparisons in other statistical areas? |
|
|
|