Oct

28

Stats Monkey generates game recap articles

Filed Under Sports Media 

The New York Times’ blog Media Decoder points out an interesting development: the Intelligent Information Lab at Northwestern University has created Stats Monkey, a system that uses an algorithm to generate a game recap story by simply using available stats and play-by-play.

Given information commonly available online about many games—the box score and the play-by-play—the system automatically generates the text of a story about that game that captures the overall dynamic of the game and highlights the key plays and key players.  The story includes an appropriate headline and a photo of the most important player in the game.

The blog post includes an example and I was quite surprised to see how well a computer could describe a game. As the blogger pointed out, the lede was buried in the piece and it could always miss out an important development in the game because it just didn’t make the stat sheet.

Rich at the Idea Lab does not believe that machine generated news will act as a threat to journalists, but not all of his commenters agreed with him. I thought he did bring up a couple of interesting points. This system could be used by journalists to help flesh out story angles because it can dig up historical trends for players and teams that a journalist may not be aware of. Think of sport organizations that may get very little media coverage, if any, like youth teams. This system could allow a Little League manager to input the stats and with the click of the button have a quick recap of the game for the team or league’s website.

What do you guys think?

Comments

Leave a Reply