Author: Mark Young
Date: 13:20:56 07/07/01
Go up one level in this thread
On July 07, 2001 at 12:31:02, Peter Kappler wrote: >On July 07, 2001 at 09:50:39, Mark Young wrote: > >>In an incredible achievement ChessTiger defeated all 4 Grandmasters in the IV >>República Argentina International Chess Tournament. ChessTiger also chalked up >>the unbelievable TPR of 2788 and going undefeated in the tournament. >> >>So much for computers not being able to cope with the positional understanding >>of Grandmasters. Either tactics are more important in chess then positional >>understanding, or computers understand positional play better then we give them >>credit. > > >Tactics are much more important. We humans tend to over-estimate the importance >of positional play, because it's our strength. :) > >Until recently, I was a member of the "computers aren't GM strength" camp, but >it seems painfully obvious that that line has now been crossed. I've followed >the debates here, and I find it amusing how much faith the nay-sayers place in >the power of anti-computer playing style, as if an average GM can just spend a >few weeks working on this skill, and suddenly his strength against computers >will increase 200 or 300 points. I think it's more like several months of >practice, with a potential gain of perhaps 100 points. (And perhaps even this >is nothing special - a GM can realize a significant relative strength gain >against ANY opponent (computer or human) if he dedicates a large amount of time >to studying their particular weaknesses, pet openings, etc.) > >Given the recent 2700+ performances by Tiger and Junior, "I think the machines will still be GM strength even against proper anti-computer play." The problem I have with the anti-computer play argument is the word “proper”. Now that computers can win even against anti-computer play, the definition of anti-computer play has changed. It now means anti-computer style of play that is only won by the human player. So in many minds since they use this new definition, anti-computer play is bullet proof, and no computer can win against that kind of proper play. The problem is it is much harder to drive the computer into positions that the computer can’t handle well. The programmers have plugged many shortcomings in the computers playing style. It was not to many years ago that computers would commit positional suicide without much if any effort by the human players. Those days have now past, and it is not easy to just knock the computer of book, hope for a suicidal move, or lock up the center and squeeze the computer to death. There is still hope for the humans; computers are still vulnerable to the slow build up of material behind their pawns with the idea of an all out king attack. Even this is not bullet proof. And with CPU >speeds doubling every 12-18 months, this battle looks hopeless to me. Remember >that Tiger's recent result was on relatively slow hardware (866 MHz). > >Everybody seems to agree that machines are World Champion strength at blitz, and >Super-GM strength at rapid (30 min), so why all the controversy over GM strength >at standard time controls? > >-Peter
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.