Looking quickly at the NCAA tournament. . ..in the first round, our ranking system would have predicted:
Connecticut
|
+24 over Chattanooga
|
|
Carolina
|
+24 over Radford
|
|
Duke
|
+20 over Binghamton
|
|
Louisville
|
+18 over Morehead
|
|
Memphis
|
+17 over Cal State Northridge
|
|
Pittsburgh
|
+17 over East Tennessee State
|
|
Oklahoma
|
+16 over Morgan State
|
|
Michigan State
|
+14 over Robert Morris
|
|
|
|
|
|
Missouri
|
+14 over Cornell
|
|
Gonzaga
|
+14 over Akron
|
|
Villanova
|
+13 over American
|
|
West Virginia
|
+12 over Dayton
|
|
UCLA
|
+10 over VCU
|
|
Syracuse
|
+10 over Little Stevie Austin
|
|
Purdue
|
+9 over Northern Iowa
|
|
Xavier
|
+9 over Portland State
|
|
|
|
|
|
Washington
|
+9 over Mississippi State
|
|
Illinois
|
+9 over Western Kentucky
|
|
Kansas
|
+8 over North Dakota State
|
|
Wake Forest
|
+8 over Cleveland State
|
|
Brigham Young
|
+7 over Texas A&M
|
|
Arizona State
|
+6 over Temple
|
|
Clemson
|
+6 over Michigan
|
|
Marquette
|
+4 over Utah State
|
|
|
|
|
|
California
|
+4 over Maryland
|
|
USC
|
+3 over Boston College
|
|
|
|
(despite the seedings)
|
|
Texas
|
+3 over Minnesota
|
|
Utah
|
+2 over Arizona
|
|
Ohio State
|
+2 over Sienna
|
|
Wisconsin
|
+1 over Florida State
|
|
|
|
(despite the seedings)
|
|
Oklahoma State
|
+0.3 over Tennessee
|
|
Butler
|
+0.0 over LSU
|
|
|
|
(despite the seedings)
|
|
Our system would have agreed with “seeding” as to who should win the game for 29 of the 32 matchups. Although USC was the 10 seed and Boston College the 7, our system thought USC was about three points better (112.2 to 108.8); our system was right about that. Although Florida State was the 5 seed and Wisconsin the 12 seed, our system thought that Wisconsin was half a point better, which it turned out that they were. And, although Butler was a 9 seed and LSU the 8 seed, our system thought. ..well, our system had them both at 111.2, but we had Butler higher by some very small fraction. Which turned out to be incorrect.
This is same chart, but we’ve added how the games actually turned out:
Connecticut
|
+24 over Chattanooga
|
U Conn by 56
|
Carolina
|
+24 over Radford
|
UNC by 43
|
Duke
|
+20 over Binghamton
|
Duke by 24
|
Louisville
|
+18 over Morehead
|
Louisville by 20
|
Memphis
|
+17 over Cal State Northridge
|
Memphis by 11
|
Pittsburgh
|
+17 over East Tennessee State
|
Pittsburgh by 10
|
Oklahoma
|
+16 over Morgan State
|
Oklahoma by 26
|
Michigan State
|
+14 over Robert Morris
|
Michigan State by 15
|
In this group, then, all of the “better” or higher-ranked teams (by us) did win. We had expected these 8 teams to win by a total of 150 points; they actually won by a total of 205, with UConn and Carolina providing almost all of the overage. In the second group:
Missouri
|
+14 over Cornell
|
Missouri by 19
|
Gonzaga
|
+14 over Akron
|
Gonzaga by 13
|
Villanova
|
+13 over American
|
Villanova by 13
|
West Virginia
|
+12 over Dayton
|
Dayton by 8 (UPSET)
|
UCLA
|
+10 over VCU
|
UCLA by 1
|
Syracuse
|
+10 over Stephen F.
|
Syracuse by 15
|
Purdue
|
+9 over Northern Iowa
|
Purdue by 5
|
Xavier
|
+9 over Portland State
|
Xavier by 18
|
So in this group the higher seeds went 7-1. The favored teams were expected to outscore the underdogs by a total of 91 points; they actually outscored the other teams by only 76 (84 in the wins, -8 in the upset.) In the third group:
Washington
|
+9 over Mississippi State
|
Washington by 13
|
Illinois
|
+9 over Western Kentucky
|
W Kentucky by 4 (UPSET)
|
Kansas
|
+8 over North Dakota State
|
Kansas by 10
|
Wake Forest
|
+8 over Cleveland State
|
Cleveland St by 15 (UPSET)
|
Brigham Young
|
+7 over Texas A&M
|
Texas A&M by 13 (UPSET)
|
Arizona State
|
+6 over Temple
|
ASU by 9
|
Clemson
|
+6 over Michigan
|
Michigan by 3 (UPSET)
|
Marquette
|
+4 over Utah State
|
Marquette by 1
|
In this group the higher-ranked teams went just 4-4. They favored teams were expected to outscore the others by 57. They actually were outscored by 2. And in the games that were expected to be closest:
California
|
+4 over Maryland
|
Turtles by 13 (UPSET)
|
USC
|
+3 over Boston College
|
USC by 17
|
Texas
|
+3 over Minnesota
|
Texas by 14
|
Utah
|
+2 over Arizona
|
Arizona by 13 (UPSET)
|
Ohio State
|
+2 over Sienna
|
Sienna by 2 (UPSET)
|
Wisconsin
|
+1 over Florida State
|
Wisconsin by 2
|
Oklahoma State
|
+0.3 over Tennessee
|
Oklahoma State by 2
|
Butler
|
+0.0 over LSU
|
LSU by 4 (Quasi-Upset)
|
 
So in this group, again, our method was 4-4. The expected winners were expected to outscore the expected victims by 15, but actually outscored them by only 3.
In toto, then, our system was 23-9 at picking the first-round games—not great, not terrible. Probably about average, but at least we beat the President. The teams we had expected to win, we had expected to outscore their opponents by a total of 313 points. They actually outscored them by 282.
Looking ahead to the second round of games, we would predict:
|
Gonzaga over Western Kentucky by 15
|
|
Memphis over Maryland by 12
|
|
Tarheels over LSU by 12
|
|
Kansas over Dayton by 11
|
|
|
|
|
Connecticut over Texas A & M by 11
|
|
Louisville over Sienna by 10
|
|
Pittsburgh over Oklahoma State by 8
|
|
Oklahoma by Michigan by 7
|
|
|
|
|
Duke over Texas by 6
|
|
Arizona over Cleveland State by 4
|
|
Missouri over Marquette by 4
|
|
Michigan State over USC by 4
|
|
|
|
|
UCLA over Villanova by 3
|
|
|
(despite the seedings)
|
|
Xavier over Wisconsin by 2
|
|
Arizona State over Syracuse by 2
|
|
|
(despite the seedings)
|
|
Washington over Purdue by 1
|