LEARNING YOUR CHESS ENGINE'S LINGO
It sure would be cool if you could feed one of your games to a chess
computer, let it analyze for a while, and have it spit out a multi-page plain
language commentary about your chess strengths and weaknesses,
including several Kmochian paragraphs about your pawn structures,
and a Dvoretsky-like breakdown of your endgame play. Yep, that would
definitely be cool.
But it won't happen. At least not anytime soon.
Many computers can understand voice commands (as long as they're
programmed to do so), but they have the devil of a time using natural
language to talk back to us. Some chess programs claim to offer game
analysis in such natural language – and they do – but it's mostly smoke
and mirrors. The responses are “canned”, that is, pre-programmed and
often not very informative.
A chess computer can give us a whole lot of input as to what we dis
wrong in a game and how we could have played better. Over the course
of many such analyzed games we can also spot tendencies in how we
play, categories of mistakes that we make over and over again, which will
show us what problems we need to correct in order to improve. But to get
that information, we'll have to meet our computer chess coach halfway.
It can't learn to use our language but, believe it or not, it's actually pretty
easy for us to learn its language.
All you need to remember is this: 1.00 equals a pawn, and positive
numbers mean that White's ahead.
I'm serious. At its core, it's that easy.
87
chessking.com