This is our old blog. It hasn't been active since 2011. Please see the link above for our current blog or click the logo above to see all of the great data and content on this site.
«
»

The Real Field of 64?

Posted by Neil Paine on March 16, 2010

Note: This post was originally published at College Basketball Reference, S-R's brand-new College Hoops site, so when you're done reading, go over and check it out!

As much fun as the next few weeks will be, I think we can all acknowledge that the NCAA Tournament field does not represent the 64 best D-I basketball teams in the country (nor does it represent the 64 most "accomplished" teams in the land, however you want to define that). Automatic bids to teams from small conferences give the tourney a feeling of equality and tiny schools a chance to shine on a big stage, but the most talented team in the NIT field would still be favored by a wide margin against many of the NCAA Tourney's lower seeds. I'm not complaining about this reality, mind you, and I certainly don't begrudge the NCAA for giving an opportunity to small-conference and mid-major teams. But do you ever wonder what the NCAA field would look like if it did only include the 64 "best" or "most accomplished" teams in the nation?

For this exercise, we'll rank teams two different ways: by SRS (adjusted for home-court advantage), which should be the truest measure of a team's "talent" over the course of the entire season because point margin is the best predictor of future W-L%; and by the BBR Ranking method, which also accounts for schedule strength and game location but focuses on a team's wins and losses only (point differential is only factored into the SOS adjustment). As a subset of each ranking, we'll take a look at how the bracket would look if the field ranked teams 1-64 straight-up, as well as if every conference still had an automatic bid -- but that bid went to the conference's highest-ranked team instead of the tournament champ. First, here are the 64 best teams in the country by SRS:

Rank Team Conf G W L MOV SOS SRS
1 Kansas B12 34 32 2 17.26 7.93 25.18
2 Duke ACC 34 29 5 15.55 8.88 24.44
3 Syracuse BE 32 28 4 13.51 7.32 20.84
4 Kentucky SEC 34 32 2 13.05 6.43 19.49
5 Brigham Young MWC 33 28 5 16.83 2.62 19.45
6 Kansas St B12 32 25 7 9.60 9.55 19.15
7 Ohio State B10 34 27 7 12.32 6.44 18.75
8 Wisconsin B10 31 23 8 10.51 7.85 18.36
9 West Virginia BE 33 27 6 8.76 9.42 18.18
10 Texas B12 33 24 9 10.92 7.26 18.18
11 Maryland ACC 30 22 8 10.54 7.49 18.03
12 Georgetown DC BE 33 23 10 7.60 9.66 17.26
13 Villanova BE 31 24 7 9.23 7.86 17.09
14 Baylor B12 31 24 7 9.60 7.46 17.06
15 Missouri B12 32 22 10 10.88 6.03 16.91
16 Purdue B10 32 27 5 9.54 7.25 16.79
17 California P10 33 23 10 9.16 7.49 16.65
18 Clemson ACC 31 21 10 8.89 7.20 16.09
19 Xavier OH A10 32 24 8 10.64 5.12 15.76
20 Michigan St B10 32 24 8 8.40 7.04 15.44
21 Minnesota B10 34 21 13 8.17 7.14 15.31
22 Washington P10 33 24 9 8.90 6.36 15.26
23 Tennessee SEC 33 25 8 8.07 7.01 15.08
24 Marquette BE 33 22 11 8.13 6.87 15.00
25 Florida St ACC 31 22 9 7.76 7.10 14.87
26 Vanderbilt SEC 31 23 8 6.98 7.66 14.63
27 Texas A&M B12 31 22 9 5.93 8.69 14.61
28 Utah St WAC 33 26 7 12.97 1.43 14.40
29 Georgia Tech ACC 33 21 12 6.00 8.19 14.19
30 St Mary's CA WCC 30 25 5 12.07 1.65 13.72
31 Louisville BE 32 20 12 5.88 7.82 13.71
32 Temple A10 34 29 5 8.69 4.56 13.25
33 Pittsburgh BE 32 24 8 5.75 7.47 13.22
34 Butler Horz 32 28 4 9.85 3.31 13.16
35 UTEP CUSA 32 26 6 10.84 2.26 13.09
36 Arizona St P10 32 22 10 7.31 5.58 12.89
37 New Mexico MWC 33 29 4 8.95 3.90 12.84
38 Mississippi SEC 31 21 10 7.17 5.58 12.74
39 UNLV MWC 32 24 8 8.13 4.41 12.54
40 Mississippi St SEC 34 23 11 7.55 4.98 12.53
41 Memphis CUSA 32 23 9 9.51 2.75 12.26
42 Notre Dame BE 34 23 11 6.03 6.22 12.25
43 Virginia Tech ACC 31 23 8 7.39 4.83 12.22
44 Gonzaga WCC 31 25 6 9.09 3.12 12.20
45 Old Dominion CAA 34 26 8 10.28 1.91 12.19
46 Oklahoma St B12 32 22 10 5.57 6.60 12.17
47 Connecticut BE 32 17 15 2.81 9.13 11.94
48 San Diego St MWC 31 23 8 7.79 4.14 11.92
49 Wake Forest ACC 29 19 10 4.16 7.65 11.81
50 Florida SEC 33 21 12 5.45 6.06 11.51
51 Miami FL ACC 32 19 13 5.83 5.62 11.45
52 Dayton A10 32 20 12 6.73 4.59 11.31
53 Illinois B10 33 19 14 3.01 8.25 11.26
54 Northern Iowa MVC 32 28 4 8.81 2.13 10.95
55 Seton Hall BE 31 19 12 4.72 5.84 10.56
56 Richmond A10 34 26 8 6.86 3.70 10.56
57 Va Commonwealth CAA 31 22 9 9.26 1.24 10.50
58 North Carolina ACC 32 16 16 2.01 8.48 10.49
59 Murray St OVC 32 28 4 15.28 -5.15 10.13
60 Cincinnati BE 33 18 15 1.96 7.92 9.88
61 Siena MAAC 33 27 6 9.97 -0.37 9.60
62 Alabama SEC 32 17 15 2.85 6.44 9.30
63 Michigan B10 31 14 17 0.91 8.35 9.26
64 Alabama Birmingham CUSA 31 23 8 5.97 3.28 9.25

That ranking would yield a bracket of:

1 Kansas
16 Alabama Birmingham
8 Temple
9 Pittsburgh
5 California
12 San Diego St
4 Purdue
13 Wake Forest
6 Marquette
11 Memphis
3 West Virginia
14 Richmond
7 Florida St
10 Mississippi St
2 Wisconsin
15 Va Commonwealth
1 Kentucky
16 Siena
8 Georgia Tech
9 Arizona St
5 Michigan St
12 Old Dominion
4 Villanova
13 Dayton
6 Minnesota
11 Gonzaga
3 Georgetown DC
14 Illinois
7 Utah St
10 New Mexico
2 Brigham Young
15 Cincinnati
1 Duke
16 Michigan
8 Louisville
9 Butler
5 Clemson
12 Connecticut
4 Missouri
13 Florida
6 Tennessee
11 Notre Dame
3 Texas
14 Seton Hall
7 Vanderbilt
10 UNLV
2 Ohio State
15 North Carolina
1 Syracuse
16 Alabama
8 St Mary's CA
9 UTEP
5 Xavier OH
12 Oklahoma St
4 Baylor
13 Miami FL
6 Washington
11 Virginia Tech
3 Maryland
14 Northern Iowa
7 Texas A&M
10 Mississippi
2 Kansas St
15 Murray St

Okay, but maybe that's not fair -- I mean, the power conferences own that bracket. What if we granted an automatic bid to each conference's best team by SRS and gave at-large bids to the next 32 best teams by SRS?

1 Kansas
16 South Dakota
8 Temple
9 Pittsburgh
5 California
12 Siena
4 Purdue
13 Cornell NY
6 Marquette
11 Memphis
3 West Virginia
14 Western Kentucky
7 Florida St
10 Mississippi St
2 Wisconsin
15 Belmont
1 Kentucky
16 Quinnipiac
8 Georgia Tech
9 Arizona St
5 Michigan St
12 Old Dominion
4 Villanova
13 Montana
6 Minnesota
11 Gonzaga
3 Georgetown DC
14 Sam Houston St
7 Utah St
10 New Mexico
2 Brigham Young
15 Coastal Carolina
1 Duke
16 Arkansas Pine Bluff
8 Louisville
9 Butler
5 Clemson
12 Murray St
4 Missouri
13 Kent St
6 Tennessee
11 Notre Dame
3 Texas
14 IUPUI
7 Vanderbilt
10 UNLV
2 Ohio State
15 Boston U
1 Syracuse
16 Lehigh
8 St Mary's CA
9 UTEP
5 Xavier OH
12 Northern Iowa
4 Baylor
13 Wofford
6 Washington
11 Virginia Tech
3 Maryland
14 Pacific CA
7 Texas A&M
10 Mississippi
2 Kansas St
15 Morgan St

That looks more equitable, but what if we felt using the point differential-based SRS was wrong because it encourages teams to run up the score on their opponents? Well, we can also rank the teams by their BBR ranking:

Rank Team Conf G W L BBRrk
1 Kansas B12 34 32 2 16.38
2 West Virginia BE 33 27 6 15.61
3 Duke ACC 34 29 5 15.50
4 Kansas St B12 32 25 7 14.74
5 Kentucky SEC 34 32 2 14.68
6 Syracuse BE 32 28 4 14.03
7 Purdue B10 32 27 5 13.74
8 Georgetown DC BE 33 23 10 13.23
9 Villanova BE 31 24 7 13.11
10 Baylor B12 31 24 7 12.59
11 Texas A&M B12 31 22 9 12.48
12 Wisconsin B10 31 23 8 12.09
13 Vanderbilt SEC 31 23 8 12.01
14 Ohio State B10 34 27 7 11.83
15 Temple A10 34 29 5 11.78
16 Tennessee SEC 33 25 8 11.73
17 Michigan St B10 32 24 8 11.69
18 Pittsburgh BE 32 24 8 11.69
19 Maryland ACC 30 22 8 11.64
20 Texas B12 33 24 9 11.35
21 California P10 33 23 10 11.16
22 New Mexico MWC 33 29 4 11.13
23 Florida St ACC 31 22 9 10.90
24 Butler Horz 32 28 4 10.88
25 Georgia Tech ACC 33 21 12 10.60
26 Wake Forest ACC 29 19 10 10.52
27 Clemson ACC 31 21 10 10.33
28 Washington P10 33 24 9 10.14
29 Xavier OH A10 32 24 8 9.98
30 Oklahoma St B12 32 22 10 9.85
31 Brigham Young MWC 33 28 5 9.75
32 Marquette BE 33 22 11 9.71
33 Northern Iowa MVC 32 28 4 9.70
34 Louisville BE 32 20 12 9.55
35 San Diego St MWC 31 23 8 9.49
36 Virginia Tech ACC 31 23 8 9.40
37 South Florida BE 32 20 12 9.22
38 Illinois B10 33 19 14 9.19
39 Missouri B12 32 22 10 9.05
40 Connecticut BE 32 17 15 9.03
41 Richmond A10 34 26 8 9.00
42 UNLV MWC 32 24 8 8.96
43 Gonzaga WCC 31 25 6 8.92
44 Minnesota B10 34 21 13 8.87
45 Notre Dame BE 34 23 11 8.77
46 Mississippi SEC 31 21 10 8.59
47 Arizona St P10 32 22 10 8.50
48 Cincinnati BE 33 18 15 8.34
49 Texas Tech B12 32 17 15 8.29
50 Florida SEC 33 21 12 8.16
51 Mississippi St SEC 34 23 11 8.14
52 St Mary's CA WCC 30 25 5 8.11
53 Alabama Birmingham CUSA 31 23 8 8.09
54 UTEP CUSA 32 26 6 7.99
55 Arizona P10 31 16 15 7.76
56 Rhode Island A10 32 23 9 7.65
57 North Carolina ACC 32 16 16 7.61
58 St John's NY BE 32 17 15 7.59
59 Seton Hall BE 31 19 12 7.52
60 North Carolina St ACC 34 19 15 7.51
61 Old Dominion CAA 34 26 8 7.30
62 Miami FL ACC 32 19 13 7.03
63 Utah St WAC 33 26 7 6.99
64 Michigan B10 31 14 17 6.90

Now we get this bracket:

1 Kansas
16 Michigan
8 Marquette
9 Northern Iowa
5 Michigan St
12 Cincinnati
4 Tennessee
13 Texas Tech
6 Butler
11 Richmond
3 Villanova
14 Rhode Island
7 Georgia Tech
10 Connecticut
2 Georgetown DC
15 North Carolina
1 Kansas St
16 Old Dominion
8 Xavier OH
9 Virginia Tech
5 Texas
12 Notre Dame
4 Vanderbilt
13 St Mary's CA
6 California
11 Minnesota
3 Wisconsin
14 Alabama Birmingham
7 Washington
10 South Florida
2 Kentucky
15 North Carolina St
1 West Virginia
16 Utah St
8 Brigham Young
9 Louisville
5 Pittsburgh
12 Arizona St
4 Temple
13 Florida
6 Florida St
11 UNLV
3 Baylor
14 Arizona
7 Wake Forest
10 Missouri
2 Purdue
15 St John's NY
1 Duke
16 Miami FL
8 Oklahoma St
9 San Diego St
5 Maryland
12 Mississippi
4 Ohio State
13 Mississippi St
6 New Mexico
11 Gonzaga
3 Texas A&M
14 UTEP
7 Clemson
10 Illinois
2 Syracuse
15 Seton Hall

Again, though, it's dominated by the big conferences because they get massive SOS adjustments; just look at Michigan, who went 14-17 vs. D-I competition but gets into this bracket because their SOS was one of the nation's best. Nothing against Michigan, but if we again reward each conference's best team by SRS with an automatic bid, teams with bad records but SOS boosts (like the Wolverines) will drop out:

1 Kansas
16 South Dakota
8 Marquette
9 Northern Iowa
5 Michigan St
12 Cornell NY
4 Tennessee
13 Kent St
6 Butler
11 Richmond
3 Villanova
14 Western Kentucky
7 Georgia Tech
10 Connecticut
2 Georgetown DC
15 Jacksonville FL
1 Kansas St
16 Coastal Carolina
8 Xavier OH
9 Virginia Tech
5 Texas
12 Old Dominion
4 Vanderbilt
13 UC Santa Barbara
6 California
11 Alabama Birmingham
3 Wisconsin
14 Murray St
7 Washington
10 South Florida
2 Kentucky
15 Robert Morris PA
1 West Virginia
16 Arkansas Pine Bluff
8 Brigham Young
9 Louisville
5 Pittsburgh
12 Siena
4 Temple
13 Oakland
6 Florida St
11 UNLV
3 Baylor
14 Sam Houston St
7 Wake Forest
10 Missouri
2 Purdue
15 Morgan St
1 Duke
16 Lehigh
8 Oklahoma St
9 San Diego St
5 Maryland
12 Utah St
4 Ohio State
13 Wofford
6 New Mexico
11 Gonzaga
3 Texas A&M
14 Weber St
7 Clemson
10 Illinois
2 Syracuse
15 Boston U

The only question now is, which bracket do you like best, the committee's or one of the four above? Which one is the fairest? Which one do you think would produce the "truest" national champion?

No related posts.

This entry was posted on Tuesday, March 16th, 2010 at 5:42 pm and is filed under NCAA. You can follow any responses to this entry through the RSS 2.0 feed. Both comments and pings are currently closed.

4 Responses to “The Real Field of 64?”

  1. iamse7en Says:
    March 16th, 2010 at 8:07 pm

    I like the one that ranks my team as a 2 seed, not as an 8 seed.

  2. DSMok1 Says:
    March 17th, 2010 at 1:18 pm

    Here, I think, is the most equitable: my own version of the RPI.

    The concept: have a fixed value for wins and losses, and add that to the adjusted efficiency differential of opponents (calculated by Ken Pomeroy).

    This is like your "BBR" system, only a little more refined.

    To get the fixed value of the wins and losses, I ran a regression over the last 6 years to best approximate the Ken Pomeroy Adjusted Efficiency Margin for the whole NCAA: the result was that a win was worth 25.45 points of margin (so any undefeated team gets +25.45 to add to their opponent average efficiency differential). How does that play out?

    1. Kansas (32-2) Win Margin: (32*25.45-2*29.45)/34 = 22.46 Average Opponent Adjusted Efficiency Margin: 10.2 OVERALL RATING: 32.66
    2. Duke (29-5) Win Margin: 17.96 Opp Mar: 12.2 OVERALL RATING: 30.16
    3. Kentucky (32-2) Win Mar: 22.46 Opp Mar: 7.6 OVERALL RATING: 30.06
    4. West Virginia (27-6) Win Mar: 16.20 Opp Mar: 13.4 OVERALL RATING: 29.60
    5. Syracuse (28-4) Win Mar: 19.09 Opp Mar: 9.4 OVERALL RATING: 28.49
    6. Kansas St. (26-7) Win Mar: 14.65 Opp Mar: 12.8 OVERALL RATING: 27.45
    7. Purdue 27.30
    8. Villanova 25.26
    9. Temple 25.16
    10. Baylor 24.32
    11. Butler 24.09
    12. New Mexico 24.08
    13. Georgetown 23.63
    14. Ohio St. 23.57
    15. Texas A&M 22.63
    16. Wisconsin 22.61
    17. Northern Iowa 22.49
    18. Vanderbilt 22.43
    19. Pittsburgh 22.43
    20. Maryland 22.41
    21. Michigan St. 22.33
    22. Tennessee 22.11
    23. Brigham Young 21.56
    24. Texas 21.17
    25. Florida St. 20.87
    26. California 20.33
    27. Xavier 20.33
    28. Richmond 19.47
    29. Washington 19.47
    30. San Diego St. 19.41
    31. Clemson 19.33
    32. Virginia Tech 19.31
    33. Gonzaga 19.31
    34. Georgia Tech 19.19
    35. Wake Forest 19.00
    36. St. Mary's 18.74
    37. Texas El Paso 18.51
    38. Nevada Las Vegas 18.31
    39. Oklahoma St. 18.14
    40. Marquette 17.88
    41. Missouri 17.14
    42. Old Dominion 16.97
    43. UAB 16.91
    44. Rhode Island 16.83
    45. Siena 16.70
    46. Cornell 16.68
    47. Louisville 16.46
    48. Notre Dame 16.38
    49. Utah St. 16.37
    50. South Florida 16.36
    51. Arizona St. 16.24
    52. Mississippi 15.93
    53. Mississippi St. 15.78
    54. Minnesota 15.59
    55. Illinois 15.16
    56. William & Mary 14.74
    57. Marshall 14.53
    58. Florida 14.44
    59. Wichita St. 14.38
    60. Connecticut 14.09
    61. Seton Hall 14.05
    62. Memphis 13.63
    63. Murray St. 13.46
    64. Kent St. 13.33

    This allows for the concept of simply rewarding what a team has done, since the objective is winning the game, not running up the score. It also provides a true opposition assessment.

    What do you all think? I think Florida did not deserve to be an at-large, and Notre Dame and Marquette are massively over-seeded. And Virginia Tech, who I have as an 8 seed, should have beat out Florida (who I have as a 15 seed).

  3. Brank_Manderbeak Says:
    March 17th, 2010 at 2:02 pm

    Neil, does the predictive power of point differential carry over to the college game? I know it works well enough in the NBA because each team plays each other at least twice, but with all the wonky scheduling of college seasons, I imagine it would reflect, to a greater degree, the team's schedule rather than the team's talent (and there seems to be less parity in college ball too, but that's just my impression).

  4. DSMok1 Says:
    March 17th, 2010 at 2:12 pm

    Brank: It is possible to adjust point differential according to the opposition; this is what the SRS does (and KenPom.com does, also adjusting for pace using tempo-free efficiencies). Otherwise, no, point differential is nearly meaningless in the NCAA environment.

«
»