Sadly, the chances of winning $1 billion from Warren Buffett’s Berkshire Hathaway for predicting the winner of every one of the 63 games in the National Collegiate Athletic Association men’s basketball tournament are zero this year.

Some complex legal wrangling reported in the financial press is to blame for the fact that the $1 billion prize offered by Quicken Loans and insured by Berkshire last year isn’t on the table again in 2015. (See, for example, Mar. 12, 2015 article, “Buffett’s $1 billion NCAA bet goes bust” on the CNN Money website for details of the legal battle.)

A businessman giving thumbs up looking at his March Madness bracket
A businessman giving thumbs up looking at his March Madness bracket

But what if the prize were still up for grabs again this year? What shot would you have at early retirement?

Not 1 billion-to-1, which was a popular estimate put forth by media outlets last year.

Try 1.64 trillion-to-1, according to insurance researchers William Wilt and Alan Zimmermann of Assured Research. Or more precisely, 1,644,978,976,149-to-1 instead, the pair wrote in a research note published earlier this week.

Noting that lower seeded teams have not historically had a 50/50 chance of beating the highest seeded teams, Wilt, the president of the New Jersey-based firm, and Zimmermann, the managing director, assigned historic probabilities to the first two rounds in each region in order to arrive at the estimate. From that point on, they simply assumed a 50/50 shot of picking the rest of the winners correctly.

The research note provides more detail, moving from the odds of correctly calling one entire bracket, all four brackets and then the entire tournament.

And just for fun, they give the tournament odds without using historic probabilities—just assuming that you have an even chance of picking all 63 games. The $1 billion answer using that approach is 9.2 quintillion-to-1, they say.

Is Uber Cheaper?

Using a computer instead of a calculator for a different kind of analytical exercise—this one using big data—computer scientists in the U.K. answered another burning question last week: When is Uber cheaper than a Yellow cab in New York City?

An article published in the MIT Technology Review on Mar. 16 (Data Mining Reveals When a Yellow Taxi Is Cheaper Than Uber) reveals the answer from Cecilia Mascolo of the University of Cambridge, who worked on the analysis. “Uber appears more expensive for prices below 35 dollars and begins to become cheaper only after that threshold,” the article says quoting Mascolo and providing a link to her research (arxiv.org/abs/1503.03021 “OpenStreetCab: Exploiting Taxi Mobility Patterns in New York City to Reduce Commuter Costs”).

What caught Carrier Management’s attention wasn’t the result. Instead, it was the details of the big data exercise described in the MIT Technology Review article. According to the article, the Yellow Cab data set used for the analysis covered “hundreds of millions” of trips, “tipping the scales at 50GB” with information including the location of every pickup, drop off and fares.

Using the coordinates of each journey, the researchers asked Uber the rate it would charge (using the cheapest version of its service) for the same trip, the article says.

The article also notes that Mascolo and her research colleagues from the ComputerLab at University of Cambridge, and the University of Namur in Belgium have developed a comparison pricing app called OpenStreetCab based on the analysis.

More here: http://www.openstreetcab.com/openstreetcab.pdf