Wednesday, April 09, 2014

Picking Winners

By Gary Berg-Cross

A form of March Madness is leaving the bracket-littered scene, but it provides lingering thoughts about what is given emphasis and effort in our culture and the limits of knowledge and prediction.  It’s all humbling and a bit of a mirror both on our values and also human limitations.  Basketball brackets and the associated discussion are highly structured and “professional.” They are laced with expert observations, data of all kinds and statistical analysis.  There is a run up to the tournament with lots of predictions:

NOTABLE TEAMS RISING: Saint Louis, Florida

NOTABLE TEAMS FALLING: Creighton, Michigan State, Baylor

And on the day after the last regular game newspapers have special sections on all 68 contestants.  The main players on each team are discussed and strengths and weaknesses noted.  You can see a range of predictions. This year one heard of the analysis we had statistical NCAA bracket “optimization” and lots of simulations. You can read about how off or not off predictions were explained as Bracket VooDoo which is way of gesturing that things turned out a bit different than experts believed. 

Of the 11 million people who filled out a bracket on ESPN.com, only 1,780 predicted a Kentucky-UConn final. That’s 0.016% of entrants. It was pointed out that statistically a high school basketball player filling out a bracket on ESPN has twice as much chance of getting drafted by an NBA team (0.03%) as he would have at accurately predicting the final.  Pretty humbling. Our cognitive biases have once again beaten our critical thinking, but there are psychological strategies to consider and use next year.  In America we are optimistic (well not about climate change). 
Bracketology: Somewhere between Art and Science lies perfection


So when we turn to other, important areas of American life, like political campaigns one wonders why predictions seem so much more accurate.  It seems strange that the dynamics are simpler.  Well, of course, there aren’t 68 teams.  But still if we think of an election day there is the notion of predicting the future and winning, which every 4 years involves hundreds of different races at the national, state & local level.  The results of those contests are of much greater importance for the direction of the country than the NCAA tourney, but the culture devotes less consistent, if frustrating, effort to covering it for the citizens and our evolvement seems less focused.  Perhaps this reflects the fact that unlike basketball professionals and experts, pols and their advisors want less intellectual power devoted to help citizens figure out the contests.


So perhaps it is time for the citizenry to bone up on the process of selecting good candidates and what better way than leveraging some ideas from that big elimination contest – March Madness.  Given the disappointments with most of our March brackets we’d probably be far from the ideals of our founders, but I sure would feel better if, as a culture we tried a bit harder and smarter.  

Oh, and of course the bracket metaphor is used widely including for rating religious things as below where people get to vote for their favorite eJsus movie:

No comments: