Trying (and Failing) Perfectly (at Chess)


When I’ve talked and thought about Let’s Play: Ancient Greek Punishment in the past, one of the elements I’ve always enjoyed and dwelled on is its relationship to time and computation. Specifically, I like the way the code itself of those small games enacts the punishments ‘accurately’, in that they last forever. Your computer, in playing them, is willing to let you try to push Sisyphus’ boulder up the hill indefinitely.

Over the last few days I’ve been making something related, but with chess. Specifically, I’m making a game, Best Chess, which is just, well, a game of chess, except that your opponent plays perfectly. Or would.

It goes like this: you are presented with a standard chessboard and invited to make your first move. You make your move (whatever it is, but say ‘1. e4’). Then the computer opponent (playing as black) proceeds to solve chess from that position by exhaustive search. That is, it examines every possible outcome of every possible game ensuing from your first move. Once it has done that it will either make a move of its own (because it has established that with perfect play it will win no matter what), offer you a draw (because it has established that with perfect play there will be a draw), or resign (because it has established that with perfect play you will win no matter what). That’s it.

But of course that’s not it. For one thing, black will never move its piece, because solving chess in this way takes quite a while. There are quite a few games to examine from the starting position – people say it’s about 10^120 (also known as the Shannon Number). That’s so many that it outnumbers the atoms in the atoms in the observable universe, for example. So your if you play Best Chess on your iPhone, say, it’s not realistically going to be able to calculate all those games and then respond. Nobody is.

But the magical tension in there, for me, is that your iPhone will still be doing it nonetheless – your iPhone would be engaged in solving chess in all earnestness. So you would then be holding a computer in your hand that is solving chess. That’s interesting and strange. It seems simultaneously hopeless (and a bit sweet somehow) and extraordinarily powerful. It both is and is not performing a momentous act of computation.

I like that.

28 July 2015
← next words previous words →