Author: Jeroen van Dorp
Date: 14:34:42 01/14/01
Go up one level in this thread
Constructed positions are regularly used for assessment of engine strenght, so there's no real difference here in validating strenghtor capability. But it's not about the position, it's about the assessment of the position. The position - and a lot similar- shows that chess engines just do what everyone claims them to do: calculate and give a position a value. It's clear that sheer calculating is not the only way to play chess. Van der Wiel said it himself: the beast can calculate more than a million positions per second. I can't calculate that amount in my lifetime. Yet he is able to beat a chess computer. How come? Well, the old and weird position I showed is an illustration of the dynamics behind that. How far we are at this moment you can see in matches chess engines compete in and their impressive results. But until now a chess engine needs a lot of preconfigured -static- chess knowledge (a human interface), or a high NPS figure, or a combination of both. Yet it doesn't assess a lot of chess positions in an "intelligent" way. It will come. No doubt, and soon. But until then the chess engine is too vulnerable to stand up against specialized top chess players. And about your remark - I'm not sure, but you could be wrong with that observation. A lot of things a computer should be able to handle are still, after decades of computer science, too difficult to be handled. I don't know if positional thinking and general strategy in chess games is one of them, but looking at speech recognition or medical diagnostics and the fuzzy logics involved, it might be quite a bigger job that we think. Jeroen ;-}
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.