Might basketball be similar? In a quest to find out, de Sáa Guerra and his team did the rational thing — they watched 6,130 NBA games. Their work, which was recently published in the International Journal of Heat and Technology, shows that there’s more to basketball than March Madness.
As they analyzed seemingly endless passes, plays and scores, the team looked for a Poisson distribution. This statistical concept helps predict the probability of a given number of events happening in a fixed period of time. Poisson distributions help turn seemingly random and rare events into predictions, but they can also work the other way around, telling people whether a seemingly random set of data is really random after all.
Rather than following a classic Poisson distribution, basketball ended up acting more like a self-organized system. In these systems, order comes out of what looks like chaos — even when the stimuli that organizes them is random. For example, sand dunes could be considered self-organizing. They consist of millions of tiny grains of sand acted upon by random winds, but they come together into a physical pattern of ripples. Flocks of birds are self-organizing; so, too, are bees that come together to build a hive that has its own logic and structure.
It turns out that basketball players act similarly. Despite the seemingly endless number of options available to an individual player, the researchers observed, teams acted as a unit that kept the game flowing forward. When games stopped flowing, teams simply flowed toward another solution, taking advantage of even the smallest movement, loss of control or slip to push forward their group goal.
That might sound like the beauty of teamwork, but it also sounds suspiciously like an evolutionary hypothesis called the Red Queen Effect. In Lewis Carroll’s "Through the Looking-Glass," the Red Queen tells Alice that in this strange world, people have to run just to stay in place. Based on the theory named after that confusing queen, all organisms must adapt and evolve at all times just to keep from disappearing.
That theory was at play on the court, too. The research team’s analysis revealed that during the last 60 seconds of play, everything that came before it stopped mattering as both teams snatched any advantage they could. Defensive intensity ratcheted up during that last minute, and teams tended to risk more, force more shots and make more fouls.
All sorts of patterns emerged from the data: Not only did the vast number of games end with a difference of between one and 11 points, but games in which nobody scored for 100 seconds had higher probabilities of remaining scoreless within the next few seconds than their counterparts. Ninety-four percent of the most contentious games were decided in the last minute based on fouls.
“Self-organization is even more important in that last minute,” says de Sáa Guerra. “The me has to serve the us.”
The existence of a self-organizing, Red Queen-style group that must continually evolve in response to changing conditions could give researchers an opportunity to study micro-evolution in action, much as they watch schools of fish or herds of zebras to figure out how a group adapts and improves. But learning that basketball isn’t that random after all could also help coaches and teams up their game on the court, de Sáa Guerra notes. He should know — he’s a basketball coach himself. He says that his new insights into the predictable disorder of the game has made him a better player and coach. His tip for would-be NCAA, NBA and WNBA coaches? Force your systems to self-organize. “Introduce uncertainty into your drill,” he says. “Only by introducing uncertainty can you improve your players.”