Wrong on Watson

Well, that Jeopardy playing computer Watson sure crushed Ken Jennings and Brad Rutter. Why was I so wrong?


This should have been so obvious, but the computer had a clear advantage on the buzzer. If all three players knew the answer by the time the buzzers were unlocked, Watson won nearly every time. Not to say that was the only reason Watson won. IBM’s people definitely did a good job with Watson, otherwise the buzzer advantage would have been meaningless.

So I suppose I was wrong for two reasons–neglecting the computer’s timing advantage, and underestimating IBM’s team.

Well, there was also a third reason. I am still unhappy about the way IBM handled Deep Blue vs. Kasparov. IBM retired Deep Blue after winning the 1997 match despite being a game behind Kasparov overall. (In 1996 and 1997 combined, the score was Kasparov 4, Deep Blue 3, and 5 draws.) They clearly were only looking for the publicity, and they made their coup and then rejected Kasparov’s request (or demand, or offer, depending on whom you ask). Watson is more of the same, so I was less willing than I should have been to give them credit in my preview.

Congratulations, IBM–everyone thinks your hot stuff again. And you really did a good job programming your machine. I only wonder, though, if Jennings had picked up one or both of the Daily Doubles in Double Jeopardy of the second game, and had won the match, whether you would have gone Deep Blue on Sony.