June 22, 2012
Big data might be finding all sorts of patterns, but for basketball fans, it’s a matter of what makes the ball go in the hoop. Systems developed to track missiles are now being used to track basketballs (and their players). SportUV Systems has cameras hanging from NBA arenas, and they are recording everyone’s every move. Basketball teams are starting to sign on to using the data collected to improve their games.
While it is standard to keep track of a basketball player’s shooting average, everyone knows that circumstance plays a huge role. Big data can bring in conditional probability. What is the shooting average of Kevin Durant when his has a defender 3 feet away? SportVU can tell you. And it can also tell you the likelihood he makes a shot given many other circumstantial pieces of information, like who passed him the ball and how many minutes he’s been playing.
The results are limited by a number of factors. The biggest is lack of data; without data on past performance, no future performance can be predicted. This why SportVU wants to increase its base – currently only 10 of the 30 MBA teams are inputting data. The second relates to something more philosophical: determinism versus random error. While certainly these players have incredible skills, when matched up to each other, their statistical differences recorded as past behavior may only be part of the story. Small amounts of randomness could influence the game more than we know – and it simply cannot be predicted with all the data in the world.
June 13, 2012
Sleep and eating aren’t just connected in the brain.
The L.A. Times headlined something most of us have experienced personally: Sleep well, or face the possibility of eating poorly. But the “news” in the news story was that someone used an fMRI to measure brain patterns of people while deciding what they would like to eat. As the L.A. Times put it, “The study is looking at the neural systems of the brain to see how they affect the decisions made.“
The problem is that the neural systems may not “affect” the decisions made, but rather “reflect” the decisions made. Or, actually, the neural patterns may reflect how a tired versus a non-tired person thinks about food, regardless of which food they choose. From a detailed report on the abstract, one can see that the fMRI research suggested that, among tired people choosing foods, the frontal lobes (long associated with reasoning) reflected differences between people who had slept and those who hadn’t, but the regions traditionally associated with “reward reactivity” did not.
What does it all mean? It could be, as the L.A. Times suggested, that the brain’s functioning is impaired, leading us to make bad decisions. It could also be that sleep deprivation leads to poorer functioning of this region of the brain, but there could be other physical explanations for the desires of people to eat less healthy food when tired other than our failure “to integrate all the different signals that help us normally make wise choices about what we should eat,” (as the lead author, Stephanie Greer, a graduate student at UC Berkeley, put it). These could be changes in our energy needs, which result in specific regions of the brain reacting to those needs.
None of this changes the well-known result that sleeping less leads to eating more, and eating less healthily. It does change how the study (which was on how the brain structure might reflect this fact) is reported in the press.