Wednesday, April 10, 2013

NCAA Tournament Bracket Entries: How The Statisticians Did?

Having not followed college basketball as closely as years past, I relied more on "statistical analysis" to fill out my bracket using two sources: Georgia Tech Professor Joel Sokol's LRMC method and New York Times' Nate Silver (guy who predicted Presidential Election)

If you need a quick brush up on LRMC you can find it here.

Nate Silver's prediction performed the better of the two, especially when you factor in points which many bracket challenges do.

I did not count points from the 1st round (i.e. the First Four) as very few bracket challenges do. If you were like me you were filling your bracket out at 11:30 on Thursday morning, the day the Tournament started. 

Georgia Tech Industrial Engineering Professor Joel Sokol bracket using LRMC
1st rd: 1-3
2nd rd: 21-11
3rd rd: 11-5
Sweet 16: 4-4
Elite 8: 1-3
Final 4: 0-2
Champ Game: 0-1
Total: 38-29 (57%)
Points:  67

Sokol gained fame in the Atlanta community after being one of the few (Jessica Simpson being another) to pick his employer to make it to the Championship Game in 2004.  This year was a down year for his method only picking 57% of games correct and only getting one Final Four team.  A major flaw of this bracket was having Gonzaga advance all the way to the Championship Game (vs Florida).

Nate Silver New York Times

1st rd: 3-1
2nd rd: 22-10
3rd rd: 11-5
Sweet 16: 4-4
Elite 8: 1-3
Final 4: 1-1
Champ: 1-0
Total: 43-24 (64%)
Points: 116

Silver did considerably better nailing the Louisville pick to win it all although his numbers suggested the Cardinals were only a modest favorite.  

Silver would have placed in the 96th percentile of ESPN's Bracket Challenge, just behind Dick Vitale but chances are he would have not won your office pool. 

Me?  I finished a respectable 85th in a 564 member pool with Louisville winning but way too many 2nd and 3rd losses did me in.  

1 comment:

  1. Once month past and I already miss college bball!! come backkkkkkk

    ReplyDelete