Baseball Hacks

Get Recent Play-by-Play Data

Sometimes total statistics like hits, runs, and strikeouts aren’t good enough for your analysis. This hack explains how to build a database of data on every play.

Many baseball books and baseball databases contain only seasonal statistics: the average or total performance of a team or player during the season. But sometimes these statistics aren’t enough. You might want to know what tends to happen in certain situations (like with one man on first and no outs), or you might want a count of events (like the number of at bats with more than five pitchers) that aren’t measured in standard statistics. If you knew everything that happened in a game, you could calculate these statistics. This hack shows you how to get data on every play of every game.

Many sports-oriented web sites give detailed play-by-play descriptions of individual games, including Yahoo! Sports, ESPN.com, CBS SportsLine, and MLB.com. Most of these web sites include detailed test descriptions of each play that you can easily encode and save in a database. We’re going to use more spidering techniques [Hack #25] to find all files on a web site and save copies of them for later analysis. This hack shows you how to use the play-by-play descriptions on MLB.com because that site provides the data in XML format. While you can get the same data from other web sites in HTML format, it’s a little easier to work with the XML files.

Tip

The best source of play-by-play data is to buy it from one of ...