drhoops

[Editted -- this is actually 2002 not last year (2003)]

I ran ARPI on the results from last year, to see if teams actually do get screwed up by the use of the RPI in bubble decisions and seeding. Here is how the top 100 teams stack up from last year followed by analysis of who got hurt and who didnt, and whether the committee did a good job adjusting for RPI problems according to ARPI:

Team (Conf.) W- L (SG) adjusted_rpi (Rank Difference)
1. Maryland (ACC ) 32- 4 (15) 0.718905 (+1)
2. Kansas (B12 ) 31- 4 (18) 0.711107 (-1)
3. Oklahoma (B12 ) 31- 5 (19) 0.709416 (+0)
4. Cincinnati (CUSA ) 31- 4 (20) 0.688290 (+0)
5. Duke (ACC ) 31- 4 (20) 0.684075 (+0)
6. Connecticut (BE ) 27- 7 (22) 0.665775 (+0)
7. Pittsburgh (BE ) 29- 6 (22) 0.660786 (+7)
8. Arizona (P10 ) 24-10 (25) 0.655181 (-1)
9. Gonzaga (WCC ) 28- 4 (16) 0.646200 (+17)
10. Xavier (A10 ) 26- 6 (17) 0.644616 (+3)
11. Mississippi St. (SEC ) 27- 8 (28) 0.641097 (+1)
12. Indiana (B10 ) 24-12 (31) 0.639406 (-3)
13. Alabama (SEC ) 27- 8 (30) 0.639097 (-5)
14. Illinois (B10 ) 26- 9 (27) 0.638922 (-3)
15. Kentucky (SEC ) 21-10 (25) 0.633430 (-5)
16. Marquette (CUSA ) 25- 7 (20) 0.629917 (+8)
17. Kent St. (MAC ) 29- 6 (25) 0.627473 (-2)
18. Miami FL (BE ) 24- 8 (22) 0.625181 (+3)
19. Texas Tech (B12 ) 23- 9 (22) 0.621932 (+0)
20. Hawaii (WAC ) 27- 6 (19) 0.620006 (+12)
21. Texas (B12 ) 22-12 (27) 0.618475 (-5)
22. Ohio St. (B10 ) 24- 8 (24) 0.616923 (-5)
23. Tulsa (WAC ) 25- 7 (20) 0.616341 (+4)
24. California (P10 ) 23- 9 (24) 0.614994 (+5)
25. Western Kentucky (SB ) 25- 4 (18) 0.614871 (+5)
26. Central Connecticut (NEC ) 27- 5 (17) 0.613939 (+27)
27. Oklahoma St. (B12 ) 23- 9 (23) 0.612520 (-5)
28. Florida (SEC ) 22- 9 (26) 0.612475 (-10)
29. UCLA (P10 ) 21-12 (25) 0.611971 (-4)
30. Oregon (P10 ) 24- 9 (25) 0.610421 (-7)
31. Southern California (P10 ) 22-10 (22) 0.607522 (+2)
32. Georgia (SEC ) 22-10 (30) 0.607268 (-12)
33. Wake Forest (ACC ) 21-13 (28) 0.605690 (-5)
34. North Carolina St. (ACC ) 23-11 (22) 0.604112 (-3)
35. Pennsylvania (Ivy ) 25- 7 (23) 0.601155 (+6)
36. Stanford (P10 ) 20-10 (20) 0.600742 (-2)
37. Notre Dame (BE ) 21-11 (23) 0.597261 (-2)
38. Southern Illinois (MVC ) 28- 8 (24) 0.596923 (+5)
39. Missouri (B12 ) 24-12 (27) 0.595802 (-2)
40. Creighton (MVC ) 23- 9 (23) 0.594270 (+5)
41. Utah (MWC ) 19- 9 (20) 0.590439 (-5)
42. Michigan St. (B10 ) 19-12 (26) 0.589962 (+4)
43. Pepperdine (WCC ) 21- 9 (19) 0.588653 (+17)
44. Memphis (CUSA ) 26- 9 (27) 0.588075 (+0)
45. Villanova (BE ) 19-13 (27) 0.587968 (-6)
46. South Carolina (SEC ) 21-15 (32) 0.585985 (-8)
47. Mississippi (SEC ) 18-11 (26) 0.585881 (+1)
48. Wyoming (MWC ) 21- 9 (19) 0.585578 (+15)
49. Charlotte (CUSA ) 18-12 (24) 0.584519 (-9)
50. Butler (Horz ) 26- 6 (22) 0.584036 (+19)
51. Boston College (BE ) 19-12 (26) 0.581373 (-9)
52. St. John's (BE ) 20-12 (23) 0.581189 (-5)
53. San Diego St. (MWC ) 20-12 (24) 0.577705 (-4)
54. Georgetown (BE ) 18-11 (21) 0.575777 (+7)
55. Bowling Green (MAC ) 23- 9 (25) 0.574434 (+3)
56. Wisconsin (B10 ) 18-13 (27) 0.572136 (-6)
57. Syracuse (BE ) 23-13 (29) 0.570519 (+2)
58. Vanderbilt (SEC ) 17-15 (30) 0.570176 (-1)
59. Louisville (CUSA ) 19-13 (24) 0.570087 (+7)
60. Dayton (A10 ) 21-11 (24) 0.569939 (-4)
61. Nevada Las Vegas (MWC ) 21-11 (24) 0.569920 (+3)
62. Utah St. (BW ) 22- 8 (22) 0.569261 (+8)
63. NC Wilmington (CAA ) 23-10 (29) 0.569130 (-11)
64. Arkansas (SEC ) 14-15 (26) 0.567188 (-13)
65. Ball St. (MAC ) 23-12 (26) 0.567018 (+10)
66. Richmond (A10 ) 22-14 (29) 0.566473 (-4)
67. Louisiana St. (SEC ) 18-15 (29) 0.565251 (+7)
68. Brigham Young (MWC ) 17-12 (25) 0.564921 (-14)
69. Tennessee Tech (OVC ) 24- 7 (22) 0.564877 (+13)
70. Valparaiso (MCon ) 24- 8 (21) 0.563717 (+17)
71. Temple (A10 ) 19-15 (26) 0.563253 (-6)
72. Louisiana Tech (WAC ) 20-10 (21) 0.562397 (-4)
73. Minnesota (B10 ) 18-13 (26) 0.561881 (-2)
74. Virginia (ACC ) 17-12 (21) 0.561668 (-19)
75. Tennessee (SEC ) 14-16 (29) 0.558931 (-8)
76. St. Joseph's (A10 ) 19-12 (24) 0.558776 (-4)
77. Iowa (B10 ) 19-16 (31) 0.557114 (-4)
78. McNeese St. (Slnd ) 20- 9 (24) 0.556165 (+7)
79. Fresno St. (WAC ) 19-15 (25) 0.554788 (-1)
80. New Mexico (MWC ) 16-14 (26) 0.550235 (-4)
81. Rutgers (BE ) 18-13 (25) 0.549558 (+10)
82. Princeton (Ivy ) 15-12 (21) 0.548634 (-3)
83. South Florida (CUSA ) 19-13 (27) 0.547768 (-6)
84. Cal Irvine (BW ) 20-11 (21) 0.546614 (+10)
85. Cal Santa Barbara (BW ) 19-11 (24) 0.546305 (+5)
86. Davidson (SC ) 18-10 (22) 0.546059 (-5)
87. Detroit (Horz ) 18-13 (25) 0.545985 (+2)
88. College of Charleston (SC ) 20- 8 (22) 0.544971 (-5)
89. Virginia Commonwealth (CAA ) 21-11 (30) 0.544726 (-9)
90. Hampton (MEAC ) 25- 7 (24) 0.543654 (+13)
91. Georgia St. (ASun ) 19-11 (26) 0.540891 (+2)
92. St. Bonaventure (A10 ) 17-13 (25) 0.540813 (-8)
93. Louisiana Monroe (Slnd ) 19-12 (25) 0.538490 (+8)
94. George Mason (CAA ) 19-10 (26) 0.538440 (-2)
95. Louisiana Lafayette (SB ) 19-11 (21) 0.537576 (+0)
96. Wagner (NEC ) 19-10 (19) 0.537394 (+25)
97. Yale (Ivy ) 21-10 (21) 0.536693 (+8)
98. New Mexico St. (SB ) 18-11 (21) 0.535877 (+1)
99. Montana St. (BSky ) 18-10 (24) 0.535157 (+3)
100. Georgia Tech (ACC ) 15-16 (24) 0.535152 (-14)

drhoops

Good news: Overall, it seems that the difference between ARPI and RPI becomes smaller as the season progresses.

Bad news: Not for all teams...

Most underrated by RPI are:
Pitt (7,14), ARPI seed 2, RPI seed 4, actual seed 3
Gonzaga (9,26), 3, 7, 6 (underseeded)
Marquette (16,24), 4, 6, 5 (slightly underseeded)
Hawaii (20,32), 5, 8, 10 (severely underseeded)
Central Connecticut (26,53), 7, out, 14 (severely underseeded)
Pepperdine (43,60), 11/bubble, out, 10
Wyoming (48,63), 12/bubble, out, 11
Butler (50,69), 13/bubble, out, out

Most these teams underrated by RPI (getting penalized for wins against cupcakes)
were underseeded, sometimes badly, or passed over for an at large bid.

Most Overrated:
Alabama (13,8), 4, 2, 2 (overseeded)
Kentucky (15,10) 4, 3, 4 (good job committee)
Texas (21,16) 6, 4, 6 (good job committee)
Ohio St. (27,22) 7, 6, 4 (overseeded)
Florida (28,18) 7, 5, 5 (overseeded)
Oregon (30,23), 8, 6, 2 (way overseeded)
Georgia (32,20) 8, 5, 3 (way overseeded)
Villanova (45,39), 12, 10, out
South Carolina (46, 38), 12, 10, out
Charlotte (49, 40), 13, 10, 9 (way overseeded)
Boston College (51,42) 13, 11, 11 (overseeded)

All these teams overrated by RPI (except Kentucky and Marquette), were rewarded by getting overseeded or invited over other deserving teams.

So, the committee IS depending on RPI (no big news), and it IS being unfair and hurting some deserving teams while rewarding some undeserving ones.

xubrew

you're putting too much emphasis on the rpi. a lot of the teams you see as being overseeded got those seeds for reasons other than their rpi. alabama was ranked really high in the coaches poll that year, and eight of their last ten games were against rpi top 25 teams. i think a #2 seed was fair.

oregon had some quality wins, but more importantly they were #1 in the pac ten standings. conference standings are important to the committee. it is rare for a team to be seeded lower than another team if they are higher in the conference standings.


a lot of the teams you list as being underseeded didn't really beat anybody. everyone is in an uproar that year because butler didn't get in. they had more bad losses than they had quality wins and two of their last three games were losses to sub 100 rpi teams. that isn't going to get anyone in because the committee doesn't reward teams for winning 25 games if they all come against sub 100 rpi teams. nor should they. butler shouldn't have been in that year imho.


bottom line is, contrary to popular belief, the rpi is not all that important. is it flawed?? well, it would be if its main purpose was to seed the teams. however, that is NOT the purpose of the rpi. taking that into consideration, i think the rpi does what it is supposed to do and if you look at the other criteria, i don't think anyone has been cheated because of it. as far as seeding the teams goes, the rpi is never consistent with where a lot of teams wind up being seeded.

__________________
IT'S MILLER TIME!!!

JuniorNoboa

This is how I see the RPI.

If you are in the top 40 in RPI you are in the tourney (there is the odd exception)

If you are not in the top 70 in RPI, you are out of the tourney.

Does that mean RPI is important? Well, if you are top 40 you must have other factors going in favour of you beyond RPI, and if you are out of the top 70 and you have a good record, you probably have a poor schedule / lack of quality wins.

The bubble always comes down to teams in that 40-70 range. And I would think that for those teams the RPI does not even enter the discussion as to deciding who is in and who is not. The RPI is just a guide to the committee for teams that need to be further discussed (those in the area of 40-70). The deciding factors for those team will never be the RPI.

drhoops

You're right -- I realize that RPI is not the only factor and that the committee looks at a lot of other things such as conference record, road record, recent games, polls etc...

However, almost every team overseeded by RPI was overseeded by committee, while almost every team underseeded by RPI was underseeded by the committee. Yes, some or even many of these might be due to factors outside RPI, but the trend is what I was trying to show.

You may question using ARPI as the "true" measure of a team --it is not --. However, it is definitely a better measure than RPI -- wins should never hurt a teams rating. I was hoping that the committee's seeding would better correlate with ARPI than RPI. All I am saying is that it doesnt...

Also, it looks that the teams most hurt by RPI are the ones in weaker conferences such as Hawaii, Central Conn. and Gonzaga. The ones most helped are ok teams in power conferences such as Georgia, Oregon, Charlotte and BC. It maybe that RPI/ARPI is not a good measure for such teams, but it could be RPIs bias in the work...

xubrew

if you say that the rpi or arpi helps the midlevel teams in the major conferences while bringing down the higher teams of some of the lower level conferences, you are right.

however, it isn't just the rpi that has that bias. the committee looks at quality wins, bad losses, head to head competition, strength of schedule, etc. obviously a team from a stronger conference is probably going to have a stronger schedule. they will also have more chances at quality wins. they will also have the opportunity to face a lot of teams fighting for ncaa tournament spots in head to head competition, which gives them more control over their own destiny.

on the flip side, using some teams from this year as an example. college of charleston, murray state, utah state and western michigan do not have those opportunities. if they receive at large bids, they are likely to be low seeds. if they do lose a game as murray state and western michigan have done, it is to a sub 100 team and the loss looks a lot worse than say georgia losing to mississippi state.

the ultimate argument is (and there really isn't enough head to head competition to answer this for sure) if georgia or iowa state (two bbbnet bubble teams) were in the big west and played utah state's schedule, would they have as good a record?? is it really that impressive to post that kind of a record when they aren't playing the better teams?? western michigan only has two losses, but is it fair to put them in the tournament over an lsu or georgia team because had lsu been in the mac, they would have run the table as well??

personally, i think a lot of the time the smaller conferences are not given the consideration they deserve. part of the ncaa selection process looks at quality wins. i think the regular season conference champion/division champion should automatically be credited with an additional quality win against an rpi top 25 team. in doing that, it would give utah state, murray state, and western michigan a boost over a team that finishes toward the middle of a major conference. that is one of the changes i'd like to see implimented.

__________________
IT'S MILLER TIME!!!

drhoops

XU: I agree with most your points above. I dont think our positions are that far.

On good wins and bad losses and opportunities from the schedule: even though the mid-majors have fewer opportunities than the average teams in power conferences, they should be judged on how they perform in these limited opportunities.

So, if Creighton plays only 8 games against good competition and wins 6 of them, while Georgia plays 20 games against quality competition and wins 8, Creighton is clearly the better team here. A similar argument can be made with bad losses. RPI does a very bad job of capturing this IMO -- it rewards/penalizes the schedule with too little regard to how you do against it.

xubrew

i think i completely agree with you drhoops. that is why i think ALL regular season conference champs, even the swac winnier, should be credited with an additional quality win against an rpi top 25 team. it is difficult to win in conference. unlike most ooc opponents, in a lot of cases the teams play each other twice a year and are extremely familiar with each other and that can level the playing field a little bit. regular season conference champs should be given some more recognition. that way, murray state, western michigan, utah state, etc could negate a conference loss and still be in position for an at large.

__________________
IT'S MILLER TIME!!!