Good Luck With That 'Perfect' March Madness Bracket. You'll Need It Millions of basketball fans will fill out NCAA tournament brackets this week and try to correctly predict the outcomes of every game. The chances of succeeding are about 1 in 150 quintillion. A group of computer scientists are trying to beat those odds by writing programs that learn to pick winners.
NPR logo

Good Luck With That 'Perfect' March Madness Bracket. You'll Need It

  • Download
  • <iframe src="" width="100%" height="290" frameborder="0" scrolling="no" title="NPR embedded audio player">
  • Transcript
Good Luck With That 'Perfect' March Madness Bracket. You'll Need It

Good Luck With That 'Perfect' March Madness Bracket. You'll Need It

  • Download
  • <iframe src="" width="100%" height="290" frameborder="0" scrolling="no" title="NPR embedded audio player">
  • Transcript


Basketball fans have one more day before the NCAA tournament to complete a spring ritual - they finalize their March Madness brackets. People try to pick the winners round after round. And by necessity, they also try to pick the losers, which turns out to be a nearly impossible task to do perfectly.

As NPR's Adam Cole reports, neither man nor computer has been able to do it.

ADAM COLE, BYLINE: If anyone could fill out a perfect bracket, you'd think it would be Andy Dieckhoff. He's a senior at Portland State University studying applied linguistics - but in his spare time he studies basketball.

ANDY DIECKHOFF: I tend to think of Bracket Day as more enjoyable than my birthday, most years.

COLE: He started watching games with his dad when he was five or six, and he ran his first bracket pool in middle school.

DIECKHOFF: But I wanted to give myself kind of an advantage.

COLE: So he poured a bunch of win-loss stats into a spreadsheet, and used some basic equations to try and predict the winners.

DIECKHOFF: It was nothing sophisticated or anything like that when it started.

COLE: But since that first year - when Dieckhoff beat all the 11-year-olds in his pool - his system has become a lot more complex. The spreadsheet has grown to include not just scores, but rebounds, assists, three-point shooting percentages; and numerical values for some qualities that are harder to define, like hustle, discipline, even luck. The system is all about using intuition to interpret the stats.

DIECKHOFF: The weighting of everything is done on completely gut feeling.

COLE: Dieckhoff's system correctly predicted 66 of the 68 teams that made it into this years tournament. But he doubts it will help him that much when he's filling out his bracket.

DIECKHOFF: There's no making sense of it. I'm trying.


DIECKHOFF: I'm trying pretty hard to make sense of it. But it's just such a crapshoot that even when you know everything, you really don't know anything.

COLE: OK, so what if you know nothing? What are the odds of randomly predicting the outcome of every tournament game?

MIKE WEIMERSKIRCH: One hundred and forty-seven quintillion to one.

COLE: That's Mike Weimerskirch, a math professor and sports fan at the University of Minnesota.

WEIMERSKIRCH: You're far more likely to win the lottery than you are to hit up on the perfect bracket.

COLE: But what if you take a slightly smarter approach and always pick the team that is ranked higher?

WEIMERSKIRCH: You bring it down to about 150 billion to one.

COLE: But in the age of Google, shouldn't we be able to get closer than that? Shouldn't we be able to use all our computational power to create the perfect bracket? Daniel Tarlow is a computer scientist, and he's trying to do just that. He runs a very special March Madness pool.

DANIEL TARLOW: So really, there's only one serious rule.

COLE: No humans allowed. In this pool, computer programs complete the brackets. Programmers write a few lines of code and the programs must learn about basketball by chewing through the stats from thousands of games. It's similar to voice recognition software learning speech. The programs guess the outcome of a game, then refine their algorithm based on the actual results. Then they make a new guess on a new game and refine again - thousands of times.

So are these programs any good at filling out brackets?

TARLOW: Looking at the group of algorithms, it's probably not that much different than you would expect to see out of your group of friends.

COLE: The cold statistical precision of a computer program is just as unsuccessful as human intuition. Weimerskirch there just isn't enough data to overcome the randomness of college basketball.

WEIMERSKIRCH: No matter how much computer analysis you do, you're still stuck with the way the ball bounces.

COLE: So you might as well try a creative strategy.

WEIMERSKIRCH: Somebody apparently won their office pool by basing their picks on who would win the game if the mascots fought.

COLE: In some years, that strategy could bring up tricky questions.

WEIMERSKIRCH: If, say, Ohio State plays Stanford, you've got a nut going against a tree - I'm not quite sure how I would call that one.

COLE: And that kind of unpredictable match-up, that's what March Madness is all about.

Adam Cole, NPR News.


INSKEEP: I predict that we are about to say: It's MORNING EDITION, from NPR News. I'm Steve Inskeep.


And I'm Renee Montagne.

Copyright © 2013 NPR. All rights reserved. Visit our website terms of use and permissions pages at for further information.

NPR transcripts are created on a rush deadline by Verb8tm, Inc., an NPR contractor, and produced using a proprietary transcription process developed with NPR. This text may not be in its final form and may be updated or revised in the future. Accuracy and availability may vary. The authoritative record of NPR’s programming is the audio record.