|
|
Tags:
NCAA BB
Starting on 3/19 we began using data from our "smart" crowd picks to identify potential upsets for the March Madness games. Now it's time to go back, look at the predictions, and see how we did. It's tough to determine what a good benchmark is for predicting upsets - it's not the same as predicting a winner or even who will cover the spread since an "upset" is an unlikely event. But here's a list of all our upset predictions, both low and high confidence, and whether we were right ("Y") or not ("N").
3/19: W. Kentucky over Illinois - Y
3/19: Maryland over Cal St. - Y
3/20: Boston over USC - N
3/20: Sienna over Ohio - Y
3/20: Utah over Marquette - N
3/21: W Kentucky over Gonzaga - N
3/21: Maryland over Memphis - N
3/21: Texas over Duke - N
3/22: Cleveland over Arizona - N
3/22: Dayton over Kansas - N
3/26: Villanova over Duke - Y
3/27: Oklahoma over Syracuse - Y
3/28: Villanova over Pitt - Y
3/29: Oklahoma over NC - N
4/4: Michigan St. over UConn - Y
4/6: Michigan St. over NC - N
We made a total of 16 upset predictions and were correct on 7 of them. If you only count the final games since 3/26 we hit 4 out of 6 upsets which is outstanding no matter which benchmark is used. Additionally since 3/26, the "smart" crowd managed a 67% against the spread - not too shabby!
The CrowdPicks Team
|
|
|