Kaggle Excercise

 

 

With eleven games to play in the NCAA basketball tournament, we (me and @statsbylopez) find our team in second place in the March Madness Machine learning contest, a mere 0.00365 behind the leader Grimp Whelkin.  So, we’re 0.00365 away from $15K.

 

Screen Shot 2014-03-28 at 11.37.28 AM

 

What’s really impressive though, is that both of our entries are doing so well.  Our best entry is currently at 0.47589 and good for second place, but our other entry is at 0.48081, which would STILL BE GOOD FOR SECOND PLACE.  (Maybe we’re on to something here?)  This is good news as we have TWO realistic ways to win this $15K.  One of our models is big on Virginia and the other is big on Michigan State.  It’s possible that our other submission actually moves ahead of our other submission and becomes our scoring submission.

Screen Shot 2014-03-28 at 11.35.18 AMI’m not sure I can handle finishing second for $15K.  I’d rather have finished 100th just so I don’t have to worry so much about this (that’s not true of course.  I’m gonna brag about this forever.)

 

I do hope Kaggle will release winning scenarios for the top 10 prior to the Final Four so that we can at least use that info to hedge our bets.  Though I’d guess they aren’t going to do this.

Cheers.

 

 

Advertisements

Posted on March 28, 2014, in Uncategorized. Bookmark the permalink. 1 Comment.

  1. Congratulations! That’s a great achievement. I don’t think they plan to release details before the final games, so no hedging for you. It would be great if you would do a posting at some point about your approach and what you think it says about predicting college basketball in general.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: