Author: Gregor Overney
Date: 17:04:55 05/25/99
Go up one level in this thread
On May 24, 1999 at 18:12:26, Chuck wrote: >In light of the Rebel 10 - GM Rhode game, it would seem that a new >emphasis has been placed on positional understanding as opposed to >tactical efficiency in chess programs. > >I wonder what the opinon is of the CCC programmers on this. To create >a grandmaster-level program, do you need to sacrifice more speed for >knowledge than has been done in the past? Can a pure tactical searcher >like Fritz make it over the hump to grandmaster-level? > >Chuck Fritz, Junior 5, etc., they all are already playing at grandmaster level. They all have earned ratings above 2600 USCF. As soon as multiprocessor support is available for Junior 5, things will become even more difficult for humans to win in Chess. - In Chess, the "brute force" approach seems to succeed. Only Whei-Chi (also called Go) is beyond this "brute force approach", at least so far. Shogi, although more complex than Chess, could also be a good candidate for high speed search engines, as others have already pointed out. Starting from a well designed Chess program, add a super-computer to it and you will probably not lose many games. It appears that "human-based" grandmasters make more "modest" errors than expected. Just analyze their games using Junior 5, Crafty, etc. Gregor
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.