March Madness: Creating the Perfect Bracket

Today marks the beginning of March Madness, the annual college basketball tournament, and for those who fill out their brackets each year, it's a day filled with promise. Until the tournament actually starts, your bracket is still perfect. You feel really good about it. Everything is potential.

At least that's how it was described to me by a family member who does it every year. I filled out a bracket for the first time ever last night, drawing on no data or knowledge whatsoever. According to my bracket sherpas, my chances of winning this in this particular pool of family and friends (based on past winners and their systems) is the same or better without knowing anything about the teams, even though many of the people who participate study the statistics around the teams and the tournament every year.

They are like many bracket veterans out there who put a great deal of time and effort into their picks. Filling out a bracket would seem like a perfect union of enjoying college basketball and having fun with statistics and math in an effort to predict outcomes.

Everybody has an angle. Everybody has a system.

My family pool of experts sent me over to Nate Silver's FiveThirtyEight for the ultimate rundown of the statistics and predictions around the tournament. Among the many articles about the tournament is this one that points out the odds of filling out a perfect bracket are one in 2 billion (not the oft-quoted one in 9.2 quintillion, which comes from treating all 63 games as coin flips. At AllAnalytics, we all know they're not.)

Over at Forbes, this article throws a bit of shade at FiveThirtyEight for not predicting the Trump election, and runs down how to use social media analytics to fill out your bracket. Forbes points to this video from a few years back from mathematics professor, Jeff Bergen, of DePaul University, again calculating odds of a perfect bracket. Spoilers: your odds are terrible.

"While social analytics might give you an edge, you are going to have to accept the fact that you have as much chance of winning as the cat lady from accounting who picks teams based on the color of their jerseys. Good luck," the article concludes.

(Darn, there goes my bracket strategy!)

The unlikelihood of a perfect bracket is perhaps what is behind investment icon Warren Buffet's March Madness contest for employees. Buffet has announced his second annual March Madness bracket contest for employees. Anyone who picks the Sweet 16 correctly wins $1 million per year for life.

A more promising prediction method, maybe, comes from computer science UCLA undergrad Adit Deshpande in this post about applying machine learning to March Madness. He has Villanova to win with a 64% probability. That's not as strong a probability as Nate Silver's 71% for Clinton winning the election.

Meanwhile, now that you've filled out your bracket, you probably want to follow the games or the results of the games at least. This article offers some guidance about how to watch the tournament if you aren't in front of your home television.

After reading all these recommendations about bracket systems, I'm pretty confident that my system of uninformed choices gives me nearly as good a chance as many other bracket veterans of hitting the perfect bracket. And I think I'm winning in terms of time invested vs. odds of getting that perfect bracket.

But I really want to hear about your system. Did you apply analytics to your bracket choices? What data did you include? Tell us in the comments, and then come back after the tournament and tell us how it worked for you.

Jessica Davis, Senior Editor, Enterprise Apps, Informationweek

Jessica Davis has spent a career covering the intersection of business and technology at titles including IDG's Infoworld, Ziff Davis Enterprise's eWeek and Channel Insider, and Penton Technology's MSPmentor. She's passionate about the practical use of business intelligence, predictive analytics, and big data for smarter business and a better world. In her spare time she enjoys playing Minecraft and other video games with her sons. She's also a student and performer of improvisational comedy. Follow her on Twitter: @jessicadavis.

Success Secrets of Top Omnichannel Retailers

It's a tough and changing environment for retailers. Yet some are enjoying continued success during turbulent times. We take a closer look at how they do it.

A2 Radio: Lean Analytics for 2018

Lean Analytics author Alistair Croll joins AllAnalytics radio to talk about how to apply the process for 2018.

Re: Unpredictable
  • 3/30/2017 4:49:18 PM

I tend to believe that a great majority of participants are in to be socially inclusive. It's pressure not to be outside of the conversation. A win means more than money, the limelight.

Re: Unpredictable
  • 3/30/2017 3:52:27 PM

I am sure there are some dollars being exchanged in neighborhood and office pools, and someone does get close enough to win at some point. I think the fun for many lies not in knowing all the answers but beating their friends and relatives for bragging rights, so it is some analytics, some guess and a lot of the variable called LUCK! Working on the regression for that one now!

Re: Unpredictable
  • 3/24/2017 10:32:57 AM

I'm not much of a sports fan or betting afficiando either. So, I've never even contemplated this annual guessing game. But, I have wondered what percentage of the folks going through this exercise are putting money on the table hoping to win? I would suspect that might be a bit of the attraction if not the media coverage at this time of the year? But it does look like a coin toss if it true that the chances "of filling out a perfect bracket are one in 2 billion."

Re: Unpredictable
  • 3/21/2017 8:51:28 AM

I was going to make some similar statements after reading the blog post this morning.


"which comes from treating all 63 games as coin flips. At AllAnalytics, we all know they're not.)"

"A more promising prediction method, maybe, comes from computer science UCLA undergrad Adit Deshpande in this post about applying machine learning to March Madness. He has Villanova to win with a 64% probability."


I'm not saying it's right to think games are simple coin tosses but this year's tournament isn't doing anything to discredit that theory.  Villanova and Duke coming into the tournamanet as the number 1 and number 2 seed and making quick exits  broke a lot of brackets.  The outcome of those games and a few notable others sure feels like somone is going to win this year by coin toss.

Re: Unpredictable
  • 3/20/2017 7:20:04 PM

The dartboard method works for stock-picking, so why not use the same for putting together a March Madness bracket?

The other data I would consult is the betting odds in Vegas.

There is a difference between the teams people 'know' will win an the teams people are willing to put money on. It's better (though still not certain) if people are willing to risk real money on your team.

Re: Unpredictable
  • 3/18/2017 7:14:53 PM

@Tomsg   I agree.  It is pure chance, so why not just leave it to it ?   

Re: Unpredictable
  • 3/18/2017 7:12:18 PM

" science UCLA undergrad Adit Deshpande in this post about applying machine learning to March Madness. He has Villanova to win with a 64% probability."

So much for machine learning when it comes to picking teams correctly....Villanova was just sent home.

Re: Unpredictable
  • 3/18/2017 7:09:31 PM

Good Luck Jessica !   I don't think your method was much different from what I employed. Since I have never heard of anyone getting a perfect bracket, I go in with that realization that I will (probably) not get it either.

So the pressure is off,   I am suddenly Northerwestern with little to lose.   Which allows me to play loose and let some past history experience guide me where it will.   Like for instance, in years past,  I have always though highly of Kansas and every year they let me down - so that was  an outlier that I factored in but for the most part I took the team with the higher seating, understanding there is always an upset during this round so my chances of a perfect bracket would evaporate quickly.

But this year I wanted to test my theory and as of this post I was 25 out of 32. Not bad for merely using the approach described which means that there were seven constestswhere my method did not hold true. So I will have to figure out how to correctly quess these seven next year and hope my theory continues to hold true.

And of course this is just the first round, there are still four more to go....

I know.  Good Luck with That !

Re: Unpredictable
  • 3/17/2017 1:47:04 PM

I know so little that I have to search March Madness just to make sure I have the right sport.

I wonder how any type of drama or infighting within the team affects their performance.  My logic being a team that has infighting will not work as closely together and perform worse. 



Re: Unpredictable
  • 3/16/2017 9:10:23 PM

When all else fails....

Page 1 / 2   >   >>