Author: Jonas Cohonas
Date: 11:40:12 09/26/05
Go up one level in this thread
>Jonas I agree with you 100%, however Uri does raise good points as well. A win >is a win, although the method and fashion of winning seems cheap to me. Trust me >when I say and 1800+ player can win in this fashion(I can, so can anyone with >enough pratice),although what does that prove? Yes engines are weak to >anti-chess, who cares? As long as they can beat top Gm's in classical chess, >this is the true measure of success. I would love to see these obviously talents >players actually play or learn some theory and then play for the win. However, >perhaps they do not want to do this and that is also quite fine. It just gets >old seeing the same stonewall, then shuffle the pieces for the win on time. Lame >and wasted talent if you ask me. Cheers The main point that i hear Uri make is that this proves that low rated players can draw a position and win on time against the best programs in the world, fair enough. But he fails to take into account, as do many others, that there is server lag involved (which imo is the main cause of the computers losses) and that the games are _very_ similar and are based on a few simple priciples where the main one is; to lock the pawn structure and win on time. Now let's say that a programmer was even bothered to pay this any attention he could simply program the engine to bust open the position regardless of it being the best move or not and i am sure that the programs would make up for it quick and this "anti-computer" style would die out just as fast as the human would go mate. Or the programmer could make an oening book with gambits where you accept all gambits and or trade pawns quick, no need for code change even, i bet. In addition it takes many lost games (on the human part) to find out how to lock the pawn structure from engine to engine and even many and even many lost games in the process because of human screwups. So the strategy is to find out how to lock the pawns and then win on time in a drawn position... in very fast games... yaaaaawn. If anything is to be learned (of value) from this, it is that computers could be better at handling pawns in general and to reckognize attempts to block the center. An interesting code addition could be to tell the engine that if in 5 consecitive moves (out of book that is) the opponent does not play the expected move then bust open the center regardless of a 0.50-1.00 pawn drop from a material point of view. There are many ways to meet this, but it could hurt playing strength in serious games played under serious classical conditions not to mention all the time it would take to impliment, and for what, to stop a couple of guys playing gimmick chess? When people have seen this game repeated enough times i am sure the "book" won't sell :) Regards Jonas
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.