Computer Chess Club Archives




Subject: Random play

Author: Russell Reagan

Date: 21:23:24 04/08/03

In my latest attempt at a chess engine ("engine of the week") I have gotten as
far as implementing random play and winboard support. So for fun I played 1/0
bullet matches all day long on ICC :)

Final score: 57/514 (+46 =22 -446)

A result of greater than 10% isn't that bad for completely random play. Of
course all of the wins came either by time forfeits, resignations, or
disconnections. Here's a breakdown.

446 losses by checkmate

23 wins by time forfeits
18 wins by resignation
5  wins by disconnection forfeits

15 stalemates
5  draws by insufficient material
2  draws by repetition

It was interesting that the result of each game always came down to whether or
not the opponent figured out my program was playing completely random. Once that
secret was divulged, they just took all of my pieces. Every few games my program
would play "acceptable" moves during the opening, and the opponent would try and
figure things out and over analyze the situation, and lose on time.

So next time you're playing bullet chess, focus more on playing fast and you'll
get 10% of your wins for free :)

This page took 0.17 seconds to execute

Last modified: Thu, 07 Jul 11 08:48:38 -0700

Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.