How Computational Review of Chess Games Revealed Narrative Errors of Critics

Paradoxically, when other top players wrote about games in magazines and newspaper columns they often made more mistakes in their commentary than the players had made at the board. Even when the players themselves published analyses of their own games they were often less accurate than when they were playing the game. Strong moves were called errors, weak moves were praised. It was not only a few cases of journalists who were lousy players failing to comprehend the genius of the champions, or everyone missing a spectacular move that I could easily find with the help of an engine, although that did happen regularly. The biggest problem was that even the players would fall into the trap of seeing each game of chess as a story, a coherent narrative with a beginning and a middle and a finish, with a few twists and turns along the way. And, of course, a moral at the end of the story.

[...]

Computer analysis exploded this lazy tradition of analyzing chess games as if they were fairy tales. Engines don't care about story. They expose the reality that the only story in a chess game is each Individual move, weak or strong. This isn't nearly as fun or interesting as the narrative method, but it's the truth, and not just in chess. The human need to understand things as a story instead of as a series of discrete events can lead to many flawed conclusions. We are easily drawn away from the data by a nice anecdote that fits our preconceived notions or that fulfills one of the popular tropes. This is how urban legends propagate so efficiently; the best ones tell us something we really want to believe is true. I'm certainly not immune to this tendency myself. and it's impossible to overcome all our intellectual biases. But becoming aware of them is a good first step, and one of the many benefits of human-machine collaboration is helping us overcome lazy cognitive habits.