On Wednesday IBM (s ibm) detailed early results of a new sentiment-analysis experience to determine how Twitter users feel about Major League Baseball’s World Series, which begins on Wednesday. The findings won’t tell you much about who might win on any given night, but perhaps they will set the stage for a new way of thinking about who to sign, trade and cut.
Analytics already have a home in baseball
As much as many of us hate to admit, sports is a business, which is why many teams operate under a “what have you done for me lately?” mantra when it comes to re-signing, trading and cutting players. Next week at IBM’s very own Information On Demand conference, for example, Oakland A’s general manager Billy Beane (Brad Pitt plays him in Moneyball) will discuss how he changed the game by analyzing somewhat-obscure data to determine which players represented the best cost-benefit ratio. That’s the type of analysis that any businessperson can get behind: Is this guy worth the money, or is there someone else who gives me more bang for the buck?
But in most sports, business is not all about winning and losing; it’s also about pleasing the fans that keep ticket and merchandising revenue coming in. In a recession that isn’t easing, one might argue that keeping cash-strapped fans happy is more important than ever. Sometimes that might mean keeping a fan favorite even if it means paying him more than his recent statistics suggest he’s worth (see, e.g., Derek Jeter). Sometimes that might mean not overpaying for someone who the fans never wanted in the first place (pick your favorite outlandish contract).
Don’t my feelings matter?
A few years ago, my beloved Green Bay Packers had a chance to sign all-star wide receiver Randy Moss for a song. Assuming the fans wouldn’t welcome the move (Moss did pretend to moon Packer fans a year earlier), they didn’t sign him. The results were (arguably) the failure to win a Super Bowl, even though the Packers got close, and a very public falling out with Hall of Fame quarterback Brett Favre, who had wanted the team to be more aggressive in bringing in new talent.
Had social sentiment analysis been the talk of the town then, and had I suspected someone was listening, I couldn’t have tweeted hard enough to express my feelings: Sign Randy Moss! I might have been in the minority, but knowing the decision was based on hard evidence at least would have soothed the blow.
Enter the “T” score
Already, IBM’s research indicates some interesting trends that one World Series team, the St. Louis Cardinals, might take into account. Jonathan Taplin of USC’s Annenberg Innovation Lab (IBM’s partner on this project and some similar ones) highlighted the following in a blog post:
So far the students have mined an initial test of more 1.5 million public baseball-related tweets during the National League Championship Series (NLCS), gauging positive and negative nuances and establishing overall sentiment rankings among a sampling of NLCS players. They determined that the Cardinals’ Chris Carpenter garnered the highest number of tweets indicating sentiment at 1,573 — 61.4 percent positive and 21.6 percent negative. While fellow Cardinal David Freese, a fan favorite and official NLCS MVP of the pennant race, garnered fewer tweets than Carpenter (768 tweets total), he had a more uniformly positive twitter following — a ‘T’ score of 89.3 percent positive tweets compared with only 15.4 percent negative.
As the project expands over the course of the World Series, these trends will likely change somewhat and others will emerge. Done over the course of a season or two, the analysis of fans’ feelings about particular players could be telling, especially if analyzed against seemingly disparate data such as merchandise sales or attendance. Winning is great, but if slightly less production from one player means slightly more revenue from fans happy to come watch him play, or if paying big bucks to bring in a controversial player will likely cause fan backlash, it might be worth listening to what the fans have to say.
Image courtesy of Flickr user ewan and donabel