Tag Archives: investigative crowdourcing

How crowdsourcing could aid ESPN’s sign-stealing investigation

ESPN hit a double with its investigation of alleged sign-stealing by the Toronto Blue Jays. A little bit of crowdsourcing could now drive it home.

A short summary of the investigation: ESPN talked to four visiting players who claimed to see a man relaying information about upcoming pitches to Blue Jays batters, which is a big ethical no-no in baseball. The man, the players claimed, sat in the outfield seats in the players’ direct eyesight, wore white, and raised his hands above his head whenever an off-speed pitch was coming. This knowledge would give the batters a significant advantage.

Considering four players claimed to see the same thing, ESPN should be able to pinpoint which section the man-in-white sat in, maybe even the approximate row. But they don’t have any other evidence the man exists. Putting aside some statistical shortcomings of the allegations, proving that this man exists would be the most significant step in the investigation.

This is where investigations can become social. The proof, if it exists, is out there in the crowd. A man who constantly raises his hands above his head – and only when the home team is batting – would be highly unusual and noticeable behavior; surely someone in his section would have noticed if he were doing it all year. And people take a lot of pictures of themselves and their surroundings at baseball games; surely one would have the man-in-white in the background.

ESPN could have ended its story with a call to action for its millions of readers. They could say: We’re looking for people who sat in Section XXX. Help us find them.

There are natural ways to verify these accounts: The scoreboard in the background aligning with the game the reader claimed to attend. Pictures of ticket stubs. Checking the metadata of the photos. If a picture turns up, you ask your readers again: Help us identify this man. Once the man is identified, investigative reporters will do what investigative reporters do.

All this would require is a willingness to expand your source pool by asking readers for help. Let’s make it part of the automatic process.