Out of the 28 Major League Baseball teams competing today on “Opening Day,” there are a handful of squads using data and analytics to gain an advantage on the diamond. However, only one is utilizing a high-powered Cray supercomputer to run a complicated statistical analysis on the numbers.
Cray, the Seattle-based company that reeled in $561.6 million in revenue last year, is typically known for selling its supercomputers to clients in the military, financial, or earth science industries, among other non-sports arenas.
But with the rise in sports analytics — also known as sabermetrics in baseball, popularized by Oakland Athletics in the early 2000s — Cray has found new customers in the athletic world.
The 43-year-old company won’t reveal the specific MLB team using its Urika-GD appliance — which costs anywhere from $100,000 to more than $1 million — due to privacy and competitive reasons. But in an interview with GeekWire, Cray VP of Marketing Barry Bolding explained why his company’s products are now attractive to sports franchises and baseball clubs in particular.
For starters, Bolding notes that 95 percent of baseball stats have been created over the last five years thanks to the growing amount of data sensors and innovative methods of analyzing players.
“They are gathering so much data that a single person with an Excel spreadsheet can no longer analyze, in a sophisticated way, all the data they have,” Bolding said. “They need bigger and bigger computers to be able to analyze the data.”
As popularized by Michael Lewis’ Moneyball and the subsequent movie, using baseball data to drive decisions about player personnel — and ultimately win more games — was a strategy first used successfully by the Oakland in 2003.
After seeing the A’s apply sabermetrics, more and more teams in various leagues have hired data analytics experts and invested in better computing technology; others, meanwhile, still aren’t convinced that data can help teams win.
But for those that are “all-in” with the analytics, a more powerful computer is appealing when it comes to crunching huge amounts of data in an efficient manner.
“When you need thousands of processing cores to be able to understand how the data is related to each other, you begin to need a larger computational device — not just a big disc drive,” Bolding said.
The MLB has strict rules that keep players and managers from using technology to make changes during a game. But Bolding explained that the analytics are really applied before and after the live action, when teams use data to set lineups or recruit minor league talent.
Bolding also made it clear that relying on analytics in sports does not replace a coach’s gut feeling or a certain player’s ability to make a clutch play.
“This is not going to ever make a prediction that a particular player will do a particular thing at a particular time,” he said. “It will only tell you that there’s a probability that something can happen, or that a particular matchup is favorable.”
However, Bolding added that baseball is one sport where the data is more useful given the length of a season.
“Baseball is a game of averages, and at the end of the day the data can get a game wrong or make a bad choice,” he said. “But if you can improve your averages — maybe your batting average goes up 10 points — that can have a huge impact on a 162-game season.”
While Cray has worked with the mystery MLB team for about two seasons, it has had other sports-related customers over the past decade or so. For example, golf manufacturer PING used Cray supercomputers to help beef up its research and development. The company also works with Formula-1 race car drivers looking to optimize their vehicles.
Bolding said that “not every sports team wants to invest in a large scale big data or computing infrastructure” and that each franchise has a certain philosophy that dictates how analytics are used. But he’s certain that more and more front offices will start realizing the value of using data to help drive decisions about specific players or overall strategy.
“You’ll need to be using it to be able to argue that you are at the top of your game,” Bolding said. “In 10 years, I can’t imagine that there will be a professional baseball team that doesn’t have some level of analytics.”
