Author: Robert Hyatt
Date: 06:28:30 10/23/00
Go up one level in this thread
On October 23, 2000 at 03:13:56, Thorsten Czub wrote: >On October 23, 2000 at 02:33:25, Mike Adams wrote: > >>Sometimes, quite often actually, mate is not there at the limits of what the >>computer can search in a reasonable amount of time and probably also an >>unreasonable amount of time. > >the "limit" of what the computer can search does not have to see a mate. >actually that exactly the main-point i am talking about. >it does not need to SEE the mate, if it knows what a mate is, >and knows about how to do it, it can plan it out. >but many todays chess programs do not plan this within their search-limit. >they do in fact nothing else than playing checkers. eating pieces. >this works, if the opponent makes mistakes within this limits, >but it is not the goal of chess. > >>Human chess players even at the top, dont assume >>they can win a game with checkmate before the endgame. > I hate to throw water on your parade, but you are attempting to create a second "urban legend" without knowing what you are talking about. GT and CSTal have exactly _nothing_ in common. GT has larger-than-life king safety scores. That is all. No different search paradigm or anything else. Let's keep this out of the world of science fiction and in the range of programming fact. It isn't the second-coming of CSTal, with a more accurate search. It is a program with an evaluation that is tuned differently. It is _still_ the same "bean counter" has always been, just some of the 'beans' are now larger than others. Back to reality... >humans have different problems playing chess. they have a main problem that is >that they make mistakes. blunder. > >> Also an important >>element of the king attack in GM chess as well as computer chess is you put >>pressure on the opponents King first. > >and how does fritz know where and how to do it ? or junior ? or nimzo? >(to only name a few bean-counters...) > >> He may be able to succesfully defend but >>in doing so unbalances his position or drops material. The human and computer >>then needs to know that the King attack has been exausted and its time to reap >>the benefits of the matrial or postional gain the opponent is offering in >>exchange for saftey. So king attacks are not so simple. The computer still >>needs to weigh other postional and material considerations to fully take >>advantage of the benefits of a strong attack against the king. > >yes yes. >gambit tiger is doing this too. but in opposite to the others it >looks for the mate. and not only within the search limits. >rc6 had nothing within the search limits , or ?!
This page took 0.01 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.