Skip navigation

Chess is said to be a game of complete information. This means that all information about the game worth having is on the board. You don’t have to guess where the other player has his white bishop, for example. It just is where it is.

Some other games of complete information display another characteristic that is somewhat correlated with the former. Tic-tac-toe, for example. It is predictable. And boring for that. What makes chess unpredictable is that despite being information-complete, it is computation-incomplete.

What this means is that, even if you can’t really hide anything in chess, you do not know where things are going either. You cannot predict it.

There are many many forms of unpredictability, but lately i took a fancy for the kind of unpredictability seen in chess. It is unpredictable just because you can’t think of each possible thing to happen. With 16 pieces, many different moving patterns, some diverse blocking patterns, etc, there are simply too many possible moves to be considered. This makes chess fun, despite being “information complete”.

But this fact, at face value so straightforward, is now being challenged. Capablanca created a chess variant with a bigger board and some bizarre hybrid pieces, supposedly for “when chess became overanalysed”. And i’ve seen someone scoff at that claim, as if it was impossible to overanalyse chess, and this was after Deep Blue!

Computers today crack chess through brute force. When you are very little, tic-tac-toe is fun, and it is a struggle. Soon you see that there are actually very few ways to play the game. Then all games end the same way, stalemate. And there is nothing to be done about that either.

Chess is overanalysed today, as a matter of fact, because computers can brute force it. Instead of figuring out abstract patterns out of the game, they simply try out every possible combination. Even if there are quite some complexities to chess-AI design, basically the raw mips of the computer at hand is the sine-qua-non condition for strong AI play.

I am not remotely good enough at chess to see it, but there are some plays that are “computer-like”. There are occasions where people say “a human player would be more likely to play N or M, but a computer would find F more atractive”.

So, let’s cut to the chase. This all exemplifies a fairly common (and usually deeply ingrained) bias in self-concept. That intelligence equals computation. Intelligence is not that, but rather adaptation capability.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: