Author: Jesus de la Villa
Date: 11:20:57 06/22/99
Go up one level in this thread
On June 22, 1999 at 13:11:39, Dann Corbit wrote: >On June 22, 1999 at 12:13:20, Jesus de la Villa wrote: > >>Hi, >> >>The WCCC show us again that "Big Irons" are not necesary to make good >>chess. what is the point ?, big hardware ?, such harware can calculate >>than at 80 ply it can win a pawn, but hardware, not software. Any >>one of us can write a brute force, put it on big hardware, and win the >>game from the first move, is it the point?. >I doubt very much if that is the case. The only "Big Iron" to score well was >Cilkchess. The other two monster hardware entries got the stuffings knocked out >of them. Do you imagine that you could do better? It is a very difficult >project to separate chess into independent threads of execution. Those >programmers like Vincent Diepeveen, Robert Hyatt, and Don Dailey (et.al.) have >done something remarkable in providing that separation. Your question was, >"What is the point?" > >The point is that we should try as many things as we possibly can to find out >what works best. If a given approach does not work as well as some other >approach, then we modify it. If it still does not work as well, perhaps we will >try something else. > >>What programs has to improve are strategics, the programs must target to >>a long term strategic supported by small tactical combinations, what a >>beauty, isn't it the way we play ?, do we need to see 20 ply to make a >>single move ?. >None of the chess programs in the contest plotted strategies. Human written >chess programs have a lot of tactical knowlege, a little positional knowlege and >zero strategic knowlege. > >>All the time we are targeting to a estrategic point, sometimes appears >>tactics with and end other than our estrategic target, but that is >>circunstantial, we don't played the game for this. >That is the biggest difference between humans and computers in chess playing. >We do very well at forming a long range goal and trying to reach it. However, >short range details can trip us up. Computers are the opposite. > >>excuseme if i say WE when i must say I. >> >>IMO speed is not the main point, improve knowledge. >I agree with this assessment, if we are to have some kind of revolutionary >advance. It seems like if putting estrategic in chess progs is a paradigm, don't askme to break it, i just want to encourage some genius out there to make some effort on this way, putting more and more processors will not make great advances at all. I know the difference between man-machine but, if it is not an excuse than actually any program has long term estrategics to avoid the challenge. Regards
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.