21
kiki
2y

One of the first attempts to use machine learning in chess computing yielded interesting results. The AI that was fed hundreds of thousands of historic grandmaster matches immediately sacrificed the queen, smashing it over a random pawn. The reason? When a grandmaster sacrifices the queen, it is usually a hard, calculated decision that leads to winning. So, AI assumed that the best way to win is to sacrifice the queen as quick as possible.

Comments
  • 2
    This wouldn’t sound as dumb if someone actually tried it in a tournament and won because the opponent got totally confused by the queen sacrifice. 🙂
  • 5
    @LensFlare This would be a tournament with all competitors required to drink a half bottle of whisky an hour before play, right? Chess tournaments aren't generally for the easily confused...
  • 3
    @spongegeoff considering the number of Russians in chess tournaments, I’d rather suggest Vodka.
Add Comment