Pool C

Dave 'd-mac' McHugh · February 04, 2010, 01:11:28 PM

Maybe I am lost. While I appreciate the RPI information, why does it seem people are convinced the NCAA is using RPI to rank and thus select teams to the tournament? We all realize they do not use RPI, right?

KnightSlappy · February 04, 2010, 01:38:38 PM

Quote from: Dave 'd-mac' McHugh on February 04, 2010, 01:11:28 PM
Maybe I am lost. While I appreciate the RPI information, why does it seem people are convinced the NCAA is using RPI to rank and thus select teams to the tournament? We all realize they do not use RPI, right?

I think everyone is aware that neither RPI nor an RPI like number is expressly mentioned in the handbook, but why would they bother with a numerical wining percentage and strength of schedule criteria if they weren't going to compare the two in some sort of numerical way?

Even if the committee doesn't use an exact calculation to compare WP and SOS, there should exist some sort of correlation between an RPI-like calculation and the NCAA's rankings (before head to head adjustments and the such). Right?

IMO, I think the NCAA does do some sort of hard calculation with WP and SOS before they adjust for head to head and results vs. ranked opponents.

AO · February 04, 2010, 01:57:54 PM

Quote from: KnightSlappy on February 04, 2010, 01:38:38 PM
Quote from: Dave 'd-mac' McHugh on February 04, 2010, 01:11:28 PM
Maybe I am lost. While I appreciate the RPI information, why does it seem people are convinced the NCAA is using RPI to rank and thus select teams to the tournament? We all realize they do not use RPI, right?

I think everyone is aware that neither RPI nor an RPI like number is expressly mentioned in the handbook, but why would they bother with a numerical wining percentage and strength of schedule criteria if they weren't going to compare the two in some sort of numerical way?

Even if the committee doesn't use an exact calculation to compare WP and SOS, there should exist some sort of correlation between an RPI-like calculation and the NCAA's rankings (before head to head adjustments and the such). Right?

IMO, I think the NCAA does do some sort of hard calculation with WP and SOS before they adjust for head to head and results vs. ranked opponents.

I don't know how this rpi150 is working out in the other regions, but in the West, the rankings favor strength of schedule over rpi as Gustavus is ranked ahead of Chapman.

Titan Q · February 04, 2010, 01:59:29 PM

Quote from: Dave 'd-mac' McHugh on February 04, 2010, 01:11:28 PM
Maybe I am lost. While I appreciate the RPI information, why does it seem people are convinced the NCAA is using RPI to rank and thus select teams to the tournament? We all realize they do not use RPI, right?

David, I think most realize there is no actual RPI calculation used. But this type of metric was a pretty good indicator of Pool C bids last year (see below), so it seems to be pretty useful.
-----------------------

Final numbers from Selection Sunday 2009, courtesy of the late Patrick Abegg, with the selected Pool C teams in red...

Reg Conf Rank Prior RPI OWP OOWP School Natl Status Reg Overall

NE 16 01 02 0.6259 0.5300 0.5270 Middlebury 004 A in 22-2 24-3
NE 13 02 03 0.6110 0.5167 0.5257 Mass-Dartmouth 010 A in 23-3 25-3
NE 17 03 01 0.6257 0.5643 0.5411 Worcester Polytech 013 C 4 20-4 20-5
NE 18 04 05 0.5878 0.4537 0.4823 Elms 014 B 1 25-1 26-1
NE 13 05 04 0.6096 0.5467 0.5234 Rhode Island College 021 C 9 23-5 23-5
NE 14 06 07 0.6052 0.5297 0.5354 Bridgewater State 022 A in 19-4 20-6
NE 16 07 08 0.6013 0.5507 0.5438 Amherst 031 C 15 19-6 21-6
NE 14 08 06 0.6081 0.5690 0.5249 Salem State 034 C 17 20-6 21-6
NE 11 09 09 0.5719 0.4717 0.4923 University of New England 037 A in 23-4 24-4
NE 17 10 11 0.5773 0.5038 0.5416 MIT 042 A in 19-6 20-8
NE 90 11 12 0.5958 0.5841 0.5349 Brandeis 050 C 23 17-8 17-8
NE 12 12 nr 0.5697 0.5131 0.5025 St. Joseph's (Maine) 051 A in 21-7 21-7
NE 16 13 10 0.5936 0.5951 0.5305 Bowdoin 059 C 29 17-9 17-9

EA 21 01 01 0.6305 0.5339 0.5374 Ithaca 003 C 1 22-2 24-2
EA 24 02 02 0.5953 0.5129 0.5135 Hamilton 020 C 8 16-3 18-7
EA 24 03 03 0.6035 0.5755 0.5128 St. Lawrence 038 C 18 18-6 20-6
EA 21 04 04 0.5761 0.5673 0.5159 Utica 067 C 35 17-9 17-9
EA 21 05 07 0.5852 0.5770 0.5346 Rochester Tech 070 A in 15-8 19-8
EA 61 06 08 0.5488 0.4892 0.4779 Medaille 076 A in 17-6 21-6
EA 23 07 06 0.5506 0.4953 0.5035 Fredonia State 077 17-7 18-9
EA 90 08 05 0.5657 0.5452 0.5323 Rochester 080 16-9 16-9

AT 32 01 01 0.6134 0.5152 0.5068 Richard Stockton 007 A in 22-2 25-2
AT 33 02 02 0.5939 0.5015 0.4837 SUNY-Farmingdale 015 C 5 24-3 24-3
AT 33 03 03 0.5754 0.4615 0.4940 St. Joseph's (L.I.) 027 A in 23-3 24-3
AT 31 04 05 0.5621 0.4622 0.4842 Baruch 046 C 22 21-4 23-5
AT 32 05 06 0.5715 0.5201 0.5052 William Paterson 055 C 26 20-7 20-7
AT 32 06 04 0.5748 0.5363 0.4993 Montclair State 058 C 28 16-6 20-6
AT 31 07 07 0.5417 0.4323 0.4873 Brooklyn 066 A in 22-5 23-5
AT 31 08 08 0.5141 0.4653 0.4859 Lehman 126 16-9 18-9

MA 45 01 01 0.6031 0.5352 0.5345 Franklin and Marshall 019 C 7 21-5 22-5
MA 43 02 04 0.5825 0.5051 0.5122 DeSales 028 A in 21-5 22-5
MA 41 03 02 0.5760 0.4755 0.4957 St. Mary's (Md.) 030 C 14 18-3 21-5
MA 42 04 03 0.5943 0.5303 0.5167 Widener 033 A in 20-5 22-5
MA 44 05 06 0.5647 0.4648 0.4894 Gwynedd-Mercy 044 A in 21-4 22-5
MA 45 06 05 0.5832 0.5527 0.5318 McDaniel 053 C 25 16-7 18-8
MA 46 07 07 0.5644 0.4997 0.4889 Scranton 054 B 2 20-6 21-6
MA 45 08 08 0.5905 0.5686 0.5327 Gettysburg 057 A in 18-8 18-8
MA 41 09 09 0.5589 0.4902 0.4933 Wesley 060 A in 16-5 18-9
MA 44 10 10 0.5489 0.4892 0.4765 Cabrini 073 20-7 20-7
MA 46 11 11 0.5386 0.4847 0.4894 Susquehanna 090 B 5 16-7 18-8

SO 51 01 03 0.6068 0.5343 0.5125 Texas-Dallas 017 A in 22-4 24-4
SO 54 02 04 0.6137 0.5594 0.5100 Centre 018 A in 19-4 23-4
SO 53 03 01 0.5828 0.4817 0.5108 Randolph-Macon 024 C 11 18-3 20-6
SO 54 04 02 0.5846 0.4946 0.5158 Trinity (Texas) 029 C 13 20-4 23-4
SO 53 05 05 0.5757 0.4958 0.5112 Guilford 043 C 21 20-5 21-5
SO 55 06 08 0.5751 0.5152 0.4975 Averett 047 A in 17-5 20-8
SO 51 07 09 0.5706 0.5289 0.5047 Mississippi College 052 C 24 18-7 20-7
SO 51 08 07 0.5580 0.4873 0.5073 McMurry 063 C 32 18-6 19-8
SO 54 09 06 0.5626 0.5131 0.5100 DePauw 065 C 34 15-6 19-7
SO 51 10 11 0.5612 0.5198 0.5015 Mary Hardin-Baylor 071 C 38 19-8 19-8
SO 55 11 10 0.5571 0.5133 0.5016 Christopher Newport 075 14-6 18-9

GL 64 01 02 0.6174 0.5369 0.5209 John Carroll 009 A in 21-3 23-4
GL 64 02 01 0.6095 0.5381 0.5156 Capital 016 C 6 22-4 24-4
GL 63 03 04 0.5792 0.4870 0.5031 Wooster 035 A in 21-4 22-6
GL 62 04 03 0.5785 0.4960 0.4984 Calvin 039 C 19 14-3 19-8
GL 62 05 06 0.5827 0.5271 0.4988 Hope 040 A in 14-4 21-7
GL 90 06 05 0.5918 0.5449 0.5273 Carnegie Mellon 041 C 20 15-5 19-6
GL 64 07 08 0.5577 0.5223 0.5196 Ohio Northern 081 16-8 18-9
GL 63 08 07 0.5480 0.4992 0.4979 Ohio Wesleyan 084 16-7 17-8
GL 61 09 09 0.5353 0.4870 0.4854 Penn State-Behrend 096 15-7 17-9

MW 90 01 01 0.6335 0.5390 0.5396 Washington U. 002 A in 22-2 23-2
MW 71 02 02 0.6590 0.5982 0.5702 Wheaton (Ill.) 006 A in 20-3 24-3
MW 72 03 03 0.6154 0.5479 0.5085 Transylvania 011 A in 18-3 21-5
MW 71 04 04 0.6250 0.5991 0.5609 Elmhurst 032 C 16 20-7 20-7
MW 74 05 09 0.5655 0.5005 0.5011 Lawrence 049 A in 19-6 19-6
MW 71 06 06 0.5824 0.5409 0.5679 Augustana 061 C 30 17-8 18-8
MW 71 07 05 0.6108 0.6241 0.5588 North Central 062 C 31 14-8 16-10
MW 74 08 07 0.5520 0.4835 0.5017 St. Norbert 072 17-6 18-6
MW 74 09 10 0.5544 0.5050 0.4931 Grinnell 074 15-6 18-7
MW 74 10 08 0.5580 0.5283 0.5089 Carroll 078 16-8 16-8
MW 71 11 nr 0.5628 0.5285 0.5578 Millikin 086 14-8 16-9

WE 82 01 01 0.6357 0.5117 0.5193 St. Thomas 001 A in 26-0 27-0
WE 83 02 02 0.6183 0.4914 0.5360 Puget Sound 005 C 2 21-1 24-3
WE 86 03 03 0.6501 0.6007 0.5528 UW-Stevens Point 008 A in 22-4 23-4
WE 86 04 04 0.6221 0.5607 0.5594 UW-Whitewater 012 C 3 21-5 22-5
WE 86 05 06 0.6246 0.5820 0.5616 UW-Platteville 023 C 10 17-5 22-5
WE 81 06 05 0.5786 0.4657 0.5133 Buena Vista 025 C 12 20-3 23-3
WE 83 07 07 0.5927 0.5304 0.5184 Whitworth 036 A in 19-5 22-5
WE 84 08 08 0.5662 0.4781 0.4827 Claremont-Mudd-Scripps 045 A in 19-4 21-6
WE 81 09 10 0.5711 0.5096 0.5049 Cornell 048 A in 19-6 21-6
WE 82 10 09 0.5691 0.5261 0.5041 Bethel 056 C 27 18-7 20-7
WE 89 11 nr 0.5132 0.3529 0.4972 Chapman 089 B 4 17-3 24-3

Reg Region
Conf Conference number
Rank Regional ranking
Prior Prior regional ranking
School
Natl National ranking based on regional results
Status
B + number: Pool B ranking (top 4 in tournament)
C + number: Pool C ranking of 18 teams in tournament
C second: second tier Pool C (spots 19-28)
C third: third tier Pool C (spots 29-38)
A in: clinched Pool A bid
blank: lower level Pool C

Titan Q · February 04, 2010, 02:29:25 PM

Quote from: KnightSlappy on February 04, 2010, 01:38:38 PM
IMO, I think the NCAA does do some sort of hard calculation with WP and SOS before they adjust for head to head and results vs. ranked opponents.

This is not correct actually. I just spoke with Pat Cunningham (head coach, Trinity-Tx), the national committee chair, and he confirmed that WP and SOS are not mathmatically combined in anyway during the regional or national discussions. (He did say OWP and OOWP are somehow combined.)

Towards the end of our conversation I heard another line ringing. Pat looks at the caller I.D. and says, "It's Dave McHugh calling."

Titan Q · February 04, 2010, 02:53:27 PM

Another couple notes from that conversation with Pat...

* NCAA committee rep John Williams has charged the group to "look inside the numbers." So while they are very much bound by the established criteria, there are a lot of "basketball conversations" (my term, not Pat's) that take place during the process.

* It's clear that everyone involved in administering the process is aware of the challenges presented by SOS - including 1) issues surrounding using regional criteria to compare teams nationally, and 2) the problems presented due to the fact that some teams play a lot of non-conference games (like the CCIW teams - 11) and others do not (like the ASC), and the impact this has on OWP. That doesn't mean OWP/OOWP are not used, but rather just that committee members realize they are not a perfect measure of SOS in Division III...especially when used to compare teams nationally.

KnightSlappy · February 04, 2010, 03:24:54 PM

Quote from: Titan Q on February 04, 2010, 02:29:25 PM
Quote from: KnightSlappy on February 04, 2010, 01:38:38 PM
IMO, I think the NCAA does do some sort of hard calculation with WP and SOS before they adjust for head to head and results vs. ranked opponents.

This is not correct actually. I just spoke with Pat Cunningham (head coach, Trinity-Tx), the national committee chair, and he confirmed that WP and SOS are not mathmatically combined in anyway during the regional or national discussions. (He did say OWP and OOWP are somehow combined.)

Towards the end of our conversation I heard another line ringing. Pat looks at the caller I.D. and says, "It's Dave McHugh calling."

Yeah, well, you know, that was just, like, my opinion, man.

Fair enough.

Ralph Turner · February 04, 2010, 04:20:37 PM

Quote from: Titan Q on February 04, 2010, 02:53:27 PM
Another couple notes from that conversation with Pat...

* NCAA committee rep John Williams has charged the group to "look inside the numbers." So while they are very much bound by the established criteria, there are a lot of "basketball conversations" (my term, not Pat's) that take place during the process.

* It's clear that everyone involved in administering the process is aware of the challenges presented by SOS - including 1) issues surrounding using regional criteria to compare teams nationally, and 2) the problems presented due to the fact that some teams play a lot of non-conference games (like the CCIW teams - 11) and others do not (like the ASC), and the impact this has on OWP. That doesn't mean OWP/OOWP are not used, but rather just that committee members realize they are not a perfect measure of SOS in Division III...especially when used to compare teams nationally.

If the coaches are reviewing what the data are saying deeper down, then I can accept that.

The isolated conferences such as the ASC, the NWC, the SCIAC and to a lesser degree, the MIAC (which plays 20 conference games) are just rying to fill 25 game schedules with D-III opponents.

Those conference numbers are different than OWP/OOWP values that can be generated by the CCIW, the WIAC and especially the UAA and the NESCAC.

Dave 'd-mac' McHugh · February 04, 2010, 05:33:13 PM

It was indeed me calling!

Though, Titan Q... Pat isn't the chair of the committee, he chairs the South Region and is on the national committee. The chair of the national committee will actually be on Hoopsville tonight: Charlie Brock from Springfield (MA).

I already knew there is no RPI, but Pat confirmed that as Titan Q mentioned. We will talk more with Brock about this tonight, but there isn't a hard and fast determination which should be taken at a higher "value" or factor, WL or the SOS. You can see this in each of the regional rankings... different emphasis or reward... whether it is SOS or WL. I also know that the regional committees "advise" their rankings to the naional committee which then goes and tweaks and changes if necessary before the rankings are officially released.

The point I am trying to make is... feel free to have an RPI... but tweaking your mathmetics to compensate for convesations and thinking by many coaches isn't going to get you what you are looking for. Each week, these conversations will change and adapt to more and more date being placed in front of the committees. Their decisions this week may change next week when numbers like OWP and OOWP change with more results.

Remember one thing in all of this... the committees actually factor in five criteria if they have all five to consider (which makes their jobs easier). When they don't, the main two are WL and SOS... and while the NCAA says that SOS is determined by a formula of (OWP x .667) + (OOWP x .333)... it does NOT indicate if the WL or the SOS should have more weight over the other when considering rankings. And I am quite sure it will never happen. The RPI works very well in Division I since there is so much "cross pollination" of teams. Teams on the west coast either play the teams on the east coast, or a team they have played has played those teams. As a result... you can get a good national "marker" and SOS in Division I. You simply can't compare a WL, OWP, and OOWP from a team like Whitworth to a team like Guilford... and it wouldn't be fair if we start to expect the NCAA to impliment such an RPI in Division III... because it wouldn't be a fair comparison.

I know we all look for hard numbers to get an idea of who is in and who is out... but there is only so far we can take these numbers before we have to just let conversations and smart people make the decisions. If it was all about hard numbers, we wouldn't have any excitement when it comes to the selections anyway... and even in DI they don't go exactly by the RPI!

Ralph Turner · February 04, 2010, 06:32:42 PM

Quote from: Dave 'd-mac' McHugh on February 04, 2010, 05:33:13 PM
It was indeed me calling! Though, Titan Q... Pat isn't the chair of the committee, he chairs the South Region and is on the national committee. The chair of the national committee will actually be on Hoopsville tonight: Charlie Brock from Springfield (MA).

I already knew there is no RPI, but Pat confirmed that as Titan Q mentioned. We will talk more with Brock about this tonight, but there isn't a hard and fast determination which should be taken at a higher "value" or factor, WL or the SOS. You can see this in each of the regional rankings... different emphasis or reward... whether it is SOS or WL. I also know that the regional committees "advise" their rankings to the naional committee which then goes and tweaks and changes if necessary before the rankings are officially released.

The point I am trying to make is... feel free to have an RPI... but tweaking your mathmetics to compensate for convesations and thinking by many coaches isn't going to get you what you are looking for. Each week, these conversations will change and adapt to more and more date being placed in front of the committees. Their decisions this week may change next week when numbers like OWP and OOWP change with more results.

Remember one thing in all of this... the committees actually factor in five criteria if they have all five to consider (which makes their jobs easier). When they don't, the main two are WL and SOS... and while the NCAA says that SOS is determined by a formula of (OWP x .667) + (OOWP x .333)... it does NOT indicate if the WL or the SOS should have more weight over the other when considering rankings. And I am quite sure it will never happen. The RPI works very well in Division I since there is so much "cross pollination" of teams. Teams on the west coast either play the teams on the east coast, or a team they have played has played those teams. As a result... you can get a good national "marker" and SOS in Division I. You simply can't compare a WL, OWP, and OOWP from a team like Whitworth to a team like Guilford... and it wouldn't be fair if we start to expect the NCAA to impliment such an RPI in Division III... because it wouldn't be a fair comparison.

I know we all look for hard numbers to get an idea of who is in and who is out... but there is only so far we can take these numbers before we have to just let conversations and smart people make the decisions. If it was all about hard numbers, we wouldn't have any excitement when it comes to the selections anyway... and even in DI they don't go exactly by the RPI!

Sanity!

Mr. Ypsi · February 04, 2010, 07:49:47 PM

Quote from: Ralph Turner on February 04, 2010, 06:32:42 PM
Quote from: Dave 'd-mac' McHugh on February 04, 2010, 05:33:13 PM
It was indeed me calling! Though, Titan Q... Pat isn't the chair of the committee, he chairs the South Region and is on the national committee. The chair of the national committee will actually be on Hoopsville tonight: Charlie Brock from Springfield (MA).

I already knew there is no RPI, but Pat confirmed that as Titan Q mentioned. We will talk more with Brock about this tonight, but there isn't a hard and fast determination which should be taken at a higher "value" or factor, WL or the SOS. You can see this in each of the regional rankings... different emphasis or reward... whether it is SOS or WL. I also know that the regional committees "advise" their rankings to the naional committee which then goes and tweaks and changes if necessary before the rankings are officially released.

The point I am trying to make is... feel free to have an RPI... but tweaking your mathmetics to compensate for convesations and thinking by many coaches isn't going to get you what you are looking for. Each week, these conversations will change and adapt to more and more date being placed in front of the committees. Their decisions this week may change next week when numbers like OWP and OOWP change with more results.

Remember one thing in all of this... the committees actually factor in five criteria if they have all five to consider (which makes their jobs easier). When they don't, the main two are WL and SOS... and while the NCAA says that SOS is determined by a formula of (OWP x .667) + (OOWP x .333)... it does NOT indicate if the WL or the SOS should have more weight over the other when considering rankings. And I am quite sure it will never happen. The RPI works very well in Division I since there is so much "cross pollination" of teams. Teams on the west coast either play the teams on the east coast, or a team they have played has played those teams. As a result... you can get a good national "marker" and SOS in Division I. You simply can't compare a WL, OWP, and OOWP from a team like Whitworth to a team like Guilford... and it wouldn't be fair if we start to expect the NCAA to impliment such an RPI in Division III... because it wouldn't be a fair comparison.

I know we all look for hard numbers to get an idea of who is in and who is out... but there is only so far we can take these numbers before we have to just let conversations and smart people make the decisions. If it was all about hard numbers, we wouldn't have any excitement when it comes to the selections anyway... and even in DI they don't go exactly by the RPI!

Sanity!

True, but as a SINGLE number to begin sorting out teams, I think KnightSlappy's RPI50 is a great start. We can't know how the committee will weigh W% vs. SOS, so a 50-50 split is as good as we can do.

Dave, your warnings are good for 'newbies' (and appreciated!), but most of us realize there is NO simple formula to predict selections. (And are grateful that there isn't!)

Ralph Turner · February 04, 2010, 08:42:53 PM

If Knightslappy's number gets the regional committees to winnowing down the number to the ranked number plus 3-4 more, then okay. Each region can deal with the characteristics that the PRI/SOS/OOWP/OWP gives.

gordonmann · February 04, 2010, 09:24:36 PM

I spoke with a member of the Mid Atlantic regional selection committee who said two things on this subject:

- I wish we had an RPI like Division I does

- There has to be something better than the OWP and OOWP criteria we use.

sac · February 04, 2010, 09:29:44 PM

Quote from: gordonmann on February 04, 2010, 09:24:36 PM
I spoke with a member of the Mid Atlantic regional selection committee who said two things on this subject:

- I wish we had an RPI like Division I does

- There has to be something better than the OWP and OOWP criteria we use.

OWP and OOWP are components of RPI, they were never intended to be used as stand alone metrics. Thats one thing I've never understood about using these numbers as the current criteria suggest.

RPI would be more useful if they counted all D3 games instead of just regional ones.

Ryan Scott (Hoops Fan) · February 04, 2010, 11:22:58 PM

Quote from: gordonmann on February 04, 2010, 09:24:36 PM
I spoke with a member of the Mid Atlantic regional selection committee who said two things on this subject:

- I wish we had an RPI like Division I does

- There has to be something better than the OWP and OOWP criteria we use.

I think an RPI type number would probably work for teams within one region, specifically those geographically smaller regions in which the teams all play each other. It wouldn't work nationally. It would probably be helpful for MA, E, NE, and Atlantic regions. It might work for some of the others.