Tournament Report Card
78Predictor Score
Record
49-14
63 games
Accuracy
77.8%
49 of 63
vs Chalk
+6.4%
Chalk: 45/63
Brier Score
0.138
Excellent
Analysis
Biggest lock: Duke at 99% confidence, a 15-seed gap that played out as expected. Closest call: Kentucky at 52%, basically a coin flip. Called 3 close games correctly, missed 9 upsets. Brier score of 0.138, better than picking by seed alone.
By Round
Calibration
Game Results
(4) Kansas 68,(13) California Baptist 60
(11) Texas 77,(2) Purdue 79
(5) St. John's 79,(12) Northern Iowa 53
(6) Tennessee 78,(11) Miami (Ohio) 56
(1) Michigan 95,(6) Tennessee 62
(5) Wisconsin 82,(12) High Point 83
(5) Vanderbilt 78,(12) McNeese 68
(5) Texas Tech 91,(12) Akron 71
(7) Kentucky 63,(2) Iowa State 82
(9) Iowa 59,(3) Illinois 71
(1) Florida 72,(9) Iowa 73
(1) Michigan 90,(4) Alabama 77
(1) Arizona 109,(4) Arkansas 88
(2) Connecticut 63,(1) Michigan 69
(1) Duke 80,(5) St. John's 75
(6) Louisville 83,(11) South Florida 79
(7) UCLA 75,(10) Central Florida 71
(7) UCLA 57,(2) Connecticut 73
(1) Duke 72,(2) Connecticut 73
(11) Texas 74,(3) Gonzaga 68
(8) Ohio State 64,(9) TCU 66
(2) Connecticut 71,(3) Illinois 62
(5) St. John's 67,(4) Kansas 65
(6) North Carolina 78,(11) VCU 82
(6) Tennessee 76,(2) Iowa State 62
(8) Clemson 61,(9) Iowa 67
(7) Miami (FLA.) 27,(10) Missouri 66
(5) Vanderbilt 72,(4) Nebraska 74
(1) Arizona 79,(2) Purdue 64
(1) Arizona 73,(1) Michigan 91
(7) St. Mary's (CA) 50,(10) Texas A&M 63
(6) Louisville 69,(3) Michigan State 77
(5) Texas Tech 65,(4) Alabama 90
(3) Michigan State 63,(2) Connecticut 67
(6) BYU 71,(11) Texas 79
(6) Tennessee 79,(3) Virginia 72
(8) Villanova 76,(9) Utah State 86
(9) Iowa 77,(4) Nebraska 71
(3) Illinois 65,(2) Houston 55
(7) Kentucky 89,(10) Santa Clara 84
(8) Georgia 77,(9) Saint Louis 102
Biggest Misses
High PointoverWisconsin
(12) vs (5) 82-83
IowaoverFlorida
(9) vs (1) 72-73
ConnecticutoverDuke
(2) vs (1) 72-73
TexasoverGonzaga
(11) vs (3) 74-68
TCUoverOhio State
(9) vs (8) 64-66
Final Four Review
How the Model Did
Pre-tournament predictions vs. the actual Final Four field
21.6%
pre-tourney
Our highest-probability team from the start. Won every tournament game by double digits. Dusty May has not been seriously challenged.
11.8%
pre-tourney
Picked 11.8% pre-tournament. Won four games cleanly. Tommy Lloyd reaches the Final Four in his fifth season.
13.2%
pre-tourney
Correctly picked through all four rounds. Defense has been the tournament story. First Final Four for Brad Underwood.
1.9%
pre-tourney
Our miss. We had Duke winning the East at 73% confidence. Braylon Mullins hit the buzzer-beater. UConn is going for three titles in four years.
Bottom line
The model correctly picked 3 of 4 Final Four teams and identified Michigan as the tournament favorite from February. The miss: we had Duke winning the East at 73% confidence, the most expensive wrong pick of the tournament.Connecticut's buzzer-beater comeback erased a 19-point deficit.
Selection Sunday Report Card
62We projected 41 of 68 tournament teams correctly
Field
60%
Seeds
77%
Top 8
75%
Bubble
43%
Pre-tournament projection accuracy from March 15