Mapping Batter Ability in Baseball by Using Spatial Statistics Techniques

• Activity Number: 482
• Type: Contributed
• Date/Time: Wednesday, August 4, 2010 : 8:30 AM to 10:20 AM
• Sponsor: Section on Statistical Graphics
• Author(s): Dana Draghicescu*+ and Benjamin Strong Baumer
• Companies: Hunter College, CUNY and CUNY Graduate Center
• Keywords: baseball ; hot zone ; kriging ; pitch f/x data ; kernel smoothing ; spatial prediction

Abstract

In baseball, an area in or around the strike zone in which batters are more likely to hit the ball is called a hot zone. Typically, scouting reports are based on maps displaying a player's batting average in a discretized area of the strike zone. These reports are then used by both batting and pitching coaches to devise game strategies. This paper is motivated by the Sportvision Pitch f/x data, which provides accurate continuous location coordinates for individual pitches using high-speed cameras. Extended exploratory analyses show a number of interesting and challenging spatial features that we employ in order to produce improved hot zone maps based on both parametric (kriging) and nonparametric (kernel smoothing) techniques.