It's that time of year - the NCAA tournament, called "March Madness", when office pools all across the United States have people researching teams and reading predictions to try and optimize their chances of winning money by predicting basketball games.

They are, sadly, doomed to fail.

Last year, Warren Buffett offered $1 billion for a perfect winning bracket, but the highest scoring bracket among subscribers was still 18 games off - and those people pay to know sports.
Clinical trials rely on statistics to show whether drugs are more effective than placebo pills. But how can we be certain?

How should scientists interpret their data?

Emerging from their labs after days, weeks, months, even years spent measuring and recording, how do researchers draw conclusions about the results of their experiments?

The concept of randomness has intrigued great minds for as long as history of great thinkers has been recorded.  From the nature of wind to drunks running into lampposts on their way home, scholars have been intrigued with making random predictable. 

A group of chemists say they have challenged traditional interpretations of randomness by computationally generating random and mechanically rigid arrangements of two-dimensional hard disks, think pennies with no thickness, for the first time.
Researchers focusing on a new way to discuss statistical mechanics have modeled "World War Z", an oral history of the first zombie war, to explore how an "actual" zombie outbreak might play out in the U.S.  

By Ker Than,  Inside Science

(Inside Science) -- Opinions rarely form in a vacuum. People are heavily influenced by the opinions of others in their social networks, whether they be real or virtual. Some people are not open to new ideas.

These are the zealots, who proselytize an opinion -- the superiority of Apple products, for example, or skepticism about climate change -- in the hopes of convincing others, while stubbornly resisting being influenced themselves.

Optimization algorithms seek to find the minimum values of mathematical functions in engineering. It's how engineers evaluate design trade-offs, assess control systems and how people find patterns in data.

One practical way to solve a difficult optimization problem is to first reduce it to a related but much simpler problem, then gradually add complexity back in, solving each new problem in turn and using its solution as a guide to solving the next one.

If only someone could find a way to characterize optimization theoretically. You know, optimize it.

Considering environmental effects such as a gravitational tidal force spread over a scale much larger than a galaxy cluster will be indispensable to explain the distribution and evolution of dark matter halos around galaxies, according to a comparison between theory and simulations by researchers at Kavli IPMU and collaborators.

The sex ratio in your community may affect what you're looking for in a relationship. Shutterstock

By Ryan Schacht, University of Utah and Monique Borgerhoff Mulder, University of California, Davis

Games have been test beds for new ideas in Artificial Intelligence (AI) since computers came on the scene and there have been significant milestones - Deep Blue sort of defeated Kasparov in chess and Watson sort of defeated Jennings and Rutter on Jeopardy! 

But solving a game is a lot tougher than defeating a player, though researchers in the Computer Poker Research Group at the University of Alberta in Canada say they have essentially solved heads-up limit hold'em poker