If you didn’t watch Jeopardy! tonight, you missed out. Ken Jennings and Brad Rutter, two of Jeopardy’s most well-known contestants, went up against Watson. Watson isn’t any old new contestant. He’s a machine created by IBM to take on Jeopardy! as a regular contestant would, and he’s designed to understand language and process answers (or questions) just as a regular contestant would, then buzz in with questions (or answers). In deciding whether to buzz in, Watson chooses the top three answers from all the possible answers, then decides how confident he is about the top three. This confidence level of the top three are shown at the bottom of the screen. If the top answer’s confidence level is below Watson’s buzz level, he doesn’t buzz in.
A lot of tonight’s episode was devoted to explaining how Watson worked; after all, no contestant like Watson had ever shown up on Jeopardy before. Because of this only the first round of Jeopardy aired. Double Jeopardy and Final Jeopardy will air tomorrow night, followed by a full match on Wednesday.
Watson himself wasn’t in the room because he’s too big. He was next door to the Jeopardy! room, but his avatar was there, which lit up in different colors to show his confidence level or how hard he was thinking. When Alex Trebek introduced Watson, the audience applauded, and Alex said that Watson would have appreciated the applause, but he couldn’t see or hear, then went on to explain how Watson worked.
Then Watson went on to tie with Brad for the lead by the end of the first round. Welcome to the future. He got off to a slow start, getting incorrect answers before getting ahead Ken and Brad, before taking a huge lead by the first commercial break. Watson also had his moments. He buzzed in with an incorrect answer that Ken had already said. Alex said, “Ken said that.”
Overall, Watson’s a huge leap in technology, and I look forward to seeing what happens in the next two days and beyond.
(If you missed the episode, the questions and answers are already up on the unofficial Jeopardy! archive. They’re fast.)