Author: Robert Hyatt
Date: 11:58:57 02/20/01
Go up one level in this thread
On February 20, 2001 at 12:21:37, Carl Cox wrote: >On February 19, 2001 at 07:44:30, John Wentworth wrote: > >>This is just an observance and may be completely wrong, but it seems that when a >>program castles Queen side his chances of losing go up by a lot. Every time I >>see a program do this, I say to myself he's going to lose and I bet more than >>60% of the time he does. This may be a problem with humans vs humans as well, I >>don't know. Anyone else notice this? > > >I have noticed this too. I always figured that the reason the program castles on >the Qside in the first place is because he is unable to do so on the Kside...... >viz: the game is not going ideally for him and he is in a slightly weak spot. > >I'm just curious - assuming that a program can castle on either side on the same >move, and assuming that all tactical and other strategic considerations evaluate >the same for both moves, which side will the program choose to castle on? >Is it a random choice between the two moves or is Kside castling preferred? > >What about other positions where 2 or more moves eveluate exactly the same (rare >as that may be!) how does the modern chess playing program handle this? > >CC. There are lots of answers to your question. 1. (assuming Crafty here). I castle to the side that appears to be the safest. IE if you have played Bxf6 and I had to reply gxf6, then I won't castle kingside unless it is absolutely forced. I would choose to castle queen-side if it is safer, or stay in the center if it is safer. 2. castling opposite is dangerous. if you want to play wild chess, this is the way to go, but your program had better understand attacking and not simply checkmating, because just pushing pawns won't do the trick. You have to create open lines, but more importantly, open lines that are _useful_. Most programs consider _all_ open lines bad. Many consider this even if there are no rooks on the board at all. How they plan on using open files is a mystery there, of course. 3. Many programs castle as soon as they can. Which is considered to be an elementary mistake by strong players. Some well-known GM once said "Castle if you want to, or castle if you must, but do _not_ castle just because you can." IE there is no point in telling your opponent where the king is going to live too soon in the game. This is _still_ the weakest part of computer play. If computers can solve the king safety problem, they will very likely become unbeatable most of the time... and will only lose occasionally.
This page took 0 seconds to execute
Last modified: Thu, 15 Apr 21 08:11:13 -0700
Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.