We know the rankings process is flawed.. Quick reference being when a ranked team loses to a higher ranked team they fall in the rankings where as a ranked team that beats an unranked bad rpi/sos team they move up in the rankings...  So know the new NT projection report further illuminates the flaw where as teams will be discredited more in seeding due to the original flaw of the ranking process.. Case in point:

	School	Conf	Coach	Rank	RPI	SOS	Record	Home	Away	Neutral	Last 10	NT Projection
1.	Miami (FL)	ACC	cheeznsweet	3	1	5	18-2	4-1	14-1	0-0	8-2	Lock
2.	Michigan	Big 10	tdvy31	5	8	42	20-0	6-0	14-0	0-0	10-0	Lock
3.	Syracuse	Big East	grecianfox	1	6	38	20-0	9-0	11-0	0-0	10-0	Lock
4.	Arizona	PAC 10	amsiegel	2	15	65	19-1	10-0	9-1	0-0	10-0	Lock
5.	Michigan St.	Big 10	professor17	8	4	7	17-3	4-2	13-1	0-0	8-2	Lock
6.	UNC	ACC	scottyj74	6	3	2	16-4	9-1	7-3	0-0	6-4	Lock
7.	Texas	Big 12	dwoelflin07	4	11	17	17-3	6-3	11-0	0-0	9-1	Lock
8.	Duke	ACC	coach_ms	9	2	6	17-3	5-0	12-3	0-0	8-2	Lock
9.	Georgia Tech	ACC	oldman	11	7	9	17-4	5-2	12-2	0-0	6-4	Lock
10.	Utah	MWC	shqipta	7	5	22	18-2	7-0	11-2	0-0	8-2	Lock
11.	Florida St.	ACC	sully712	14	14	15	15-5	6-2	9-3	0-0	7-3	Lock
12.	N. Carolina St.	ACC	joehof	10	10	36	18-2	10-0	8-2	0-0	8-2	Lock

Arizona is a really good team, but if the NT was to begin today, they would be a 1 Seed where as I feel Michigan St. has played a way tougher schedule but would be a 2 Seed due to their rank.  I feel the RPI/SOS then W/L then Rank should be used in NT Seeding.. Same goes for Texas (obviously ahead of my Dukies) but still.. We have the same record, Texas RPI is 9 lower while their SOS is 11 lower..  And yet they would be a higher 2 Seed than Duke..  We've complained for years about the Seeding be bogus.. Now the proof is that its the Rankings that are screwing up our Seeding..

Your entire "argument" assumes that the projections are accurate and the best way to measure teams. What evidence is there of that?

I think you are right, but it wont take long for people to point out that according the FAQ rankings are not used, although the evidance suggests otherwise.

coach, I'm not even so sure that ranking plays a direct role in the seeding process. I think they may be on there as a reference and for informational purposes.

EDIT: Good call, cburton!

Bigger picture, I hardly think that quibbling over whether the team at 4 or 5 should be getting the #1 seed is representative of a "flawed" proces (let alone one team getting a "higher" 2 seed than the other). It's a pretty two-sided issue, a judgment call. If you had teams that should be on the 5 line geting a 1 seed, well, that's flawed. But you're referring to very, very subtle shades of gray and hardly a flawed process.

Although cheez being on the top line may in itself be proof that something is amiss ...

Posted by zbrent716 on 12/9/2011 12:14:00 PM (view original):
Your entire "argument" assumes that the projections are accurate and the best way to measure teams. What evidence is there of that?

If the projections are not a basis for the eventual NT selection process then why did they add it?

The evidence doesn't suggest that rankings are used, it suggests that the same factors that go into rankings go into determining NT seeds.  That might be splitting hairs to some degree, but it's clearly not the same thing.

Is it rankings screwing up the seeding? If so, why isn't Utah ahead of Duke and GA Tech? Higher rank, more wins, and better rpi than GA Tech.

I really feel like all of the "team X is getting screwed by the new seeding" arguments are rooted in a disconnect between seed and RPI. 

People have become way too dependent on RPI for all things NT.  It's a metric, and a flawed one.  Using it as the primary basis for seeding, as you are suggesting, without the context of vs. top-100, vs. top-50, etc., isn't really a strong argument, in my opinion, for a "flawed" system.

Agreed wholeheartedly with ike.

The bigger question here is how did Utah get into that mix? 

recent dev chat

do rankings now affect tourny seeds? will rankings affect them after the changes? OR is it that rankings correlate with many of the variables used for seeding now and therefore APPEAR to determine seeds? (metsmax - Hall of Famer - 3:00 PM)

A team's rank won't directly determine seeding, but the logic uses similar inputs, so there will definitely be a correlation there. But it won't be a direct mapping of rank to tournament seed by any means.

Posted by ike1024 on 12/9/2011 12:42:00 PM (view original):
I really feel like all of the "team X is getting screwed by the new seeding" arguments are rooted in a disconnect between seed and RPI. 

People have become way too dependent on RPI for all things NT.  It's a metric, and a flawed one.  Using it as the primary basis for seeding, as you are suggesting, without the context of vs. top-100, vs. top-50, etc., isn't really a strong argument, in my opinion, for a "flawed" system.

This.

So RPI is flawed.  I think many agree with that.  It's a statement that is on the forums just as often as "my seed is wrong."  Can we provide WIS with a better system then to judge teams?  It's easy to pick apart the system, but we need solutions also if we want it fixed.  I, for one, do not have the answer, but it seems there are a lot of smart people here who understand math better than I do.

There is no right answer barret, irl or in HD.  That is why irl there are disappointments and arguments about who is in and who is out.  That is why there are multiple computer index ratings out there (RPI, Pomeroy, Sagarin and others have their own propriatary formulas that differ from each other slightly.)  There is subjectivity in the "objective" formulas, but at the end of the day it's not black and white.  Agreed that RPI rank was overused.  Some years the difference in spots 40-80 might be wide whereas the next year that same 40 spot gap might be pretty narrow.  Many times one or two games (a blow-out win) could lower your RPI compared to if you had never even played the game...that doesn't make sense either.  So I think Seble is on this and is committed to getting it as "right" as right can be.  I noticed that 6 of the 12 spots are ACC teams--is that playing a role?

I fully agree that RPI is flawed and I am not complaining about my team's rating.  My team is nowhere near good enough for postseason play and I know it.  It is only 11 games but this is still a bit interesting.

I think I'm fine where my team is ranked but it is a bit odd to see Seton Hall ranked higher.

Seton Hall is the better team so the projections probably have it right in getting the better team ranked higher.  I am just surprised the projection sees it that way.

	School	Conf	Coach	Rank	RPI	SOS	Record	Home	Away	Neutral	Last 10	NT Projection
11.	Connecticut	Big East	oldman	20	14	72	11-0	0-0	11-0	0-0	10-0	Lock
17.	Syracuse	Big East	steginman	13	25	83	10-1	4-1	6-0	0-0	9-1	Lock
58.	Villanova	Big East	swenske04		65	101	8-3	4-2	4-1	0-0	7-3	Bubble
77.	Providence	Big East	cappelgp		82	121	7-4	2-2	5-2	0-0	7-3	Bubble
78.	Cincinnati	Big East	bscollier		112	78	6-5	1-5	5-0	0-0	6-4	Bubble
84.	Seton Hall	Big East	mjwilson		192	322	9-2	1-1	8-1	0-0	8-2	Out
85.	Louisville	Big East	beachedwhale		35	57	7-4	0-1	7-3	0-0	6-4	Out
92.	St. Johns	Big East	dogget		67	162	8-3	6-0	2-3	0-0	7-3	Out