Author: Thorsten Czub
Date: 12:03:10 02/20/00
Go up one level in this thread
On February 20, 2000 at 14:28:35, Fernando Villegas wrote: >T: >What you call "guessing" OR "thesis" is not but calculation, uncomplete, >aproximative,etc, BUt calculation. i see it different. what you call calculation works in a position e.g. a test-suite like bednorz-toennissen or bratko-kopec-test or whatever. there you have key-moves. but in a normal chess game you often do not have those key-moves. so the move you play is a kind of guessing depending on your experiences (evaluation-function, hash-tables) and you power to try out (search-tree). it is a thesis ! > To do that, you does not need a total solving >of the problem, but just the best aproximative solution in a narrowed front. The >game is not played in one shot from mone 1 o move 89, but it is a serie of >positions or situations where , with limits, you can grasp some features and do >some calculations that are mesnsurable in terms of better o worst grasping of >the problem. right. each position is an event. and if there is not an exact key move situation, you have to build up a thesis in your main-line. when the programs are similar, their main line is similar. when the programs are very different, there thesis is very different. > Inside these limits, inside the structure of a position, you can >and in fact you must calculate. Let me insist: is not the entire game what you >must calculate or not, but a position. but you don't know exactly if a position is good for white or good for black. you can only guess within a horizont. and this horizont is limited. you make a thesis within a range of plies. but you don't know. it is not backtracking , it is forward-trying-out. you order your so called best moves into a main-line. but the moves are only best moves within this range. thats all. you don't know if your main line is senseless out of this range. so - in fact - you don't know much. > Call it guess if you want.Call it >imperfect calculation. No matter what, the program or player with the best >aproximate calculation to what should be best inside he position, wll do the >best practical moves. no. if the programs are very similar, the one with the deepest calculation is the best. if the programs are very different, this does not work anymore. today many programs are very similar. so the fastest is the best. i call this incest. >Perhaps our problem is sinaxtic: perhaps you end to conceive calculation just as >those exhaustive calculation you do when yo add a limited nuber of numbers. For >me is just to take into account the features of a problem. Is that a guess? i do not understand this sentence fernando. please translate it in a more easier to undrstand structure. otherwise i cannot answer it. thanks. >Fernando
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.