National Sports Journalism Center

Before Christmas Adam Hochberg contacted me for an article for the Poynter Institute about StatSheet, a service put out by a North Carolina company that creates sites for college basketball, college football and high-school basketball. StatSheet uses proprietary software to take stats and box scores and turn them into a bare-bones game narrative.

I was the one defending Team Robot. Which leads to your next question: What gives, traitor?

My first reason was an unhappy acknowledgment of our times. As I told Hochberg, “services like this are interesting if you’re looking at finite, shrinking, disappearing resources in newsrooms. If it’s going to satisfy part of an audience and let you use precious resources for other things you really need people for, that’s worth having a conversation about.”

If you run a newsroom that’s flush with cash and looking to expand sports beats, StatSheet probably isn’t for you. (Though maybe it is — more about that in a minute.) But such newsrooms are few and far between — all of us everywhere are being asked to do more with less. Handing off routine work to software would let newsrooms make better use of the resources they still have. And few newsrooms are going to be able to cover everything — if the choice is between a robo-service’s recap of a high-school hoops game or no recap, why not go with the service?

Robot game stories could be a stopgap while reporters are busy elsewhere. Reading recent accounts of sportswriters’ days by my colleague Dave Kindred and Lindsay Jones at Nieman Reports, I was struck by how often they did the same work over and over again for different editions and/or different media. If some of those stories are going out to a limited audience (say, for an early edition) or will only be online briefly, why not turn them over to the service and give the reporter an extra hour or so to spend on analysis, one-on-one interviews or something offering more value to readers? (Reading Nieman, I also wondered why reporters were so busy satisfying their bosses with mayfly-lived scoops that readers no longer notice or care about. But that’s a future column.)

For another reason, look at my column about reinventing the game story, which included automation as one possible approach. I don’t find StatSheet’s game stories illuminating or interesting reading, but I did spend time with the graphics they used for game flow and player impact. (See examples of both here.) A service I love is Fangraphs’ graphical charts of win probability for baseball games. Look at these over a few weeks and the unique fingerprint of a barn-burner leaps out from the run-of-the-mill caterpillar crawls of 4-1 contests. This is the win-probability graph of a June 2000 game in which the Mets came from seven runs down to shock the Braves with a 10-run inning. I was at that game, thought I would have a heart attack and decided I was so happy that I didn’t care. Looking at the doom-to-ecstacy levitation near the end of the graph makes me remember and smile.

And we should remember that it’s early. Efforts from Stats Sheet and its ilk will get more complex and interesting with more development. Students and professors at Northwestern developed Stats Monkey (more here) for baseball, which uses advanced stats to identify key plays and construct stories around them. That’s a tool for better robo-stories. Now, imagine joining robo-writing with tweets sent in by fans at the game offering crowd reactions and other color, and with postgame quotes supplied by teams.

I’m not saying that would make robo-stories the equal of gamers written by a real live writer. Of course not: As Hochberg and I discussed, a robo-story might well render the disastrous Steve Bartman play as a simple foul ball, missing a few seconds that will be remembered forever by anguished Cubs fans. On the other hand, Alex Gonzalez might then get his just portion of blame for booting the double-play ball that soon followed. Human observation can distort things, too.

If nothing else, as they stand right now robo-stories might show us where we’ve gotten lazy. For color, Stats Sheet’s reports fall back on cliches — “hostile territory” and “edge of our seats” — and easy talk about momentum. So, sometimes, do gamers written by real live reporters. Perhaps one day the editor’s lexicon will include the warning that “the robot could do that.”

Jason Fry is a freelance writer and media consultant in Brooklyn, N.Y. He spent more than 12 years at The Wall Street Journal Online, serving as a writer, columnist, editor and projects guy. While at WSJ.com he edited and co-wrote The Daily Fix, a daily roundup of the best sportswriting online. He blogs about the Mets at Faith and Fear in Flushing, and about the newspaper industry at Reinventing the Newsroom. Write to him at  jason.fry@gmail.com, visit him on Facebook, or follow him on Twitter.on Fry is a freelance writer and media consultant in Brooklyn, N.Y. He spent more than 12 years at The Wall Street Journal Online, serving as a writer, columnist, editor and projects guy. While at WSJ.com he edited and co-wrote The Daily Fix, a daily roundup of the best sportswriting online. He blogs about the Mets at Faith and Fear in Flushing, and about the newspaper industry at Reinventing the Newsroom. Write to him at  jason.fry@gmail.com, visit him on Facebook, or follow him on Twitter.