Is a Tennis Player's Popularity on Twitter Predictive of Playing Ability?

• Activity Number: 485
• Type: Topic Contributed
• Date/Time: Wednesday, August 6, 2014 : 10:30 AM to 12:20 PM
• Sponsor: Section on Statistics in Sports
• Author(s): Stephanie Kovalchik*+
• Companies: RAND Corporation
• Keywords: Prediction ; Sports ; Social Media

Abstract

World Rankings are the standard for evaluating performance and seeding tournaments in professional tennis. Although rankings are strong predictors of winning probability, they are not updated during tournament play and are insensitive to within-tournament events. Today, social media services like Twitter create a continuous feed of a self-selected fan base for tennis players, allowing capture of within-tournament events, like results or injuries. However, it is unclear whether an athlete's social volume is predictive of future performance. This paper investigated the predictive value of a player's 24-hour social volume and 'share of voice' (SOV) (volume relative to competitor's) for next match win for WTA and ATP singles players at the 2013 Grand Slams. Social volume counts were based on keyword-matched tweets. After adjusting for relative rank between player and opponent, neither social volume nor SOV was significantly associated with match wins for male players; yet women with greater total, positive, or negative SOV were significantly more likely to win their next match. Considering a woman tennis player's relative social volume in addition to rank may improve match predictions.