Instrumented Interconnecteds Intelligent

By Chris Nay, IBM Research Communications

ASPcollege_hoopsEvery year, top U.S. men’s college basketball teams enter a month-long tournament for a chance to be crowned champion. And it always stirs up fan and pundit predictions to pick potential winners of all 63 games.

To really give the pot a stir this year, Berkshire Hathaway, Quicken Loans and Yahoo Sports teamed up to create the Billion Dollar Bracket Challenge. Choose every match up correctly and win $1B.

While the odds of picking a perfect bracket in the challenge were steep, it didn’t put off millions of basketball fans everywhere from filling out brackets. However, it only took 48 hours, 25 games (and several upsets) for every bracket to be eliminated. (See an infographic that breaks down the bracket data.)

What the data told us

IBM Research worked with Berkshire Hathaway, who provided the bilion dollar prize, and Quicken Loans, who provided the data, to set up a private SoftLayer cloud to protect and ensure the integrity of the bracket’s data. But as data and analytics geeks, the team couldn’t help wonder about what insights they could glean from people’s picks. Was there was any logic to the madness?

As you can see in the infographic, participants who hunted for Cinderella didn’t necessarily have any more success than those who went chalk. As we head toward the last three games this weekend, the data predicts a Florida vs. Wisconsin final … with a win for Florida!

Predictions by Percentage

(download high resolution version)

IBM Research Team: Suresh Chari, Ken Goldman, Ted Habeck, Charanjit Jutla, Hugo Krawczyk, Ian Molloy, Elaine Palmer, Tal Rabin, Josyula Rao, Wilfried Teiken and Chandra Venkataraman

Bookmark and Share

Previous post

Next post

23 Comments
 
September 18, 2014
2:13 pm

She has a charm that is slightly reminiscent of the childhood favourite Rosie and Jim’s Ragdoll, with a modernized interior to suit the 21st century traveller.
Methods for baiting a hook – Step 5: Use them for fishing.
You can then take your armored vehicle and cash it in at
a chop shop, creating a nice little criminal spree combo.


Posted by: dunnage box braids pictures
 
September 1, 2014
4:31 pm

Usually, the repayment period is very long and it takes nearly ten years.

Consult with an accountant to find out if this could work in your situation. Using a borrowing power calculator, you can determine your borrowing power.


Posted by: home loan bank
 
August 27, 2014
3:35 pm

If this is not done timely, the expat will NOT be
able to use the exclusion. No matter how much we dislike it still taxes are the fact of life.
Then you narrow down to your area of expertise (in this case accounting) and place and put in your keywords and press
enter.


Posted by: boca raton classifieds rentals
 
August 24, 2014
12:57 pm

Staying inside a tanning bed for lengthy hours still have a high risk for the development of
skin cancer. Your friends and family will notice, but the tan is still natural looking and definitely not orange.
This is to ensure you are protected should anything untoward happen in the salon.


Posted by: chloe sims fake norvell neutrogena spray Tan reviews
 
August 19, 2014
11:55 pm

It could be a excellent in addition to valuable little bit of data.. dufresne We are fulfilled that you simply discussed this helpful data about. Remember to stop us up to par this way. Thank you for giving.


Posted by: hochzeitsforum.wordpress.com
 
July 26, 2014
2:18 pm

Besides of being the field of fashion, you may also prepare the costly expenses of
the profession you may want to take on. Instead try a
nice pair of wool or cotton chino dress trousers with
a button down shirt, sweater vest, tie, and sport jacket.
Being a model requires more than just being a pretty
face.


Posted by: bandage dress aliexpress wholesale
 
July 15, 2014
1:05 am

Great article.


Posted by: Lawyer Referral Service In Peoria Il
 
July 14, 2014
3:08 pm

I really like what you guys are usually up too. This kind of clever work and reporting!
Keep up the terrific works guys I’ve added you guys to our blogroll.


Posted by: Lawyer Talk In Peoria Il
 
April 30, 2014
3:56 am

one ball , ten man , big game ! Like It !


Posted by: akumuliatoriu supirkimas
 
April 17, 2014
11:15 am

Hi! Do you know if they make any plugins to safeguard against hackers?
I’m kinda paranoid about losing everything I’ve worked hard on.

Any tips?


Posted by: fitness gear
 
April 17, 2014
10:56 am

basketball more then the game !


Posted by: katalizatoriu supirkimas kaune
 
April 11, 2014
1:19 pm

This is interesting article. My 4 year old daughter picked 3 out of 4 final teams, Wisconsin, Kentucky and Florida. She won the the house tournament. Next year, I need to talk to my daughter about 2015 picks.


Posted by: Rafal Konik
 
April 11, 2014
12:24 pm

I ended up with 48 out of 65 games picked correctly. At one point, I was ranked #89 in the Billion Dollar bracket. Unfortunately, I picked so many games correctly early on, that I got screwed deeper in with fewer correct picks which were worth more points. Had the contest simply been correct number of picks, well who knows, I might be $100k richer today. :)

Interesting data in this article. I too had a Florida/Wisconsin final picked, with Wisconsin taking it all. You can imagine how pumped I was when Bucky made it into the final four!


Posted by: jameos
 
April 10, 2014
2:32 pm

this is Great. This is the core of how Analytics can be used, benefit the user and get the user to think. Broadcast this and we should help to get the message of High Value Big Data to the world !


Posted by: Jason Horner
 
April 10, 2014
5:45 am

More than I could have dreamt of! :)


Posted by: Eric FERON
 
April 10, 2014
3:27 am

I participated in this but got eliminated after consecutive upsets. It was fun though.


Posted by: Ludo
 
April 9, 2014
2:31 pm

To build on comments from Jadi and Philip I would love to see Watson “help” users pick the perfect bracket. After all Watson does not make decisions, it helps people make decisions. Have an interface that people can use to alter the results based on criteria they consider most important. This could help illustrate the convergence of Watson/Analytics/Mobile/Social. What a marketing impact!


Posted by: John Astarita
 
April 9, 2014
9:22 am

Next year, I plan to use a coin toss to select my bracket entries. I think it may be just as effective as analyzing the stats.


Posted by: Suzanne McHugh
 
April 9, 2014
6:07 am

It should be interesting to know what kind of data should have been analyzed to generate the bracket prediction.

To better explain the Big Data potential, for basketball lovers, this Goldsberry post on Grantland (http://grantland.com/features/expected-value-possession-nba-analytics/) shows how it is possible to predict the player efficiency for almost all the game decision he made in a single play.


Posted by: Davide
 
April 8, 2014
4:58 pm

I’d like to see Watson try to pick a perfect bracket next year. That’s a project I’d LOVE to work on.


Posted by: Jadi Mwendo
 
April 7, 2014
2:08 pm

Warren Buffet indicates he’d like to sponsor the competition again next year. With 1 year to prepare, how about using Watson to predict to provide the top 100K most likely brackets out of the quintillion possible combinations. Who will win the NCAA Tournament is a Big Data question with a non-trivial splash of intangibles. Spend a year preparing to download the relevant data points for the 64 entries: Wins, Losses, strength of schedule, injuries, coaching, individual and team statistics, Home/Away performance, conference and non-conference, winning streaks, trends, venue altitude, etc Its all data and the question is whether Watson can convert the data to information! Imagine the Marketing impact of having a bracket prediction that lasted to the Sweet 16 or the Elite 8 or the Final 4 or…..


Posted by: Philip Poetzinger
 
April 7, 2014
1:55 pm

Re: Buzz words.

Not really. When you consider how many people signed up and how many teams per person and all the possibilities that they “could have” chosen.

The info graphic is only 0.5% of all that was analyzed. This was only for us… the rest of what was analyzed is for the businesses.


Posted by: Frank
 
April 7, 2014
1:14 pm

This is “Big” data? Someone with access to the DB and basic query skills could produce the same analysis..


Posted by: Buzz Words
 
Post a Comment