Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: ==> time to start solving chess / Chinchalkar????

Author: S. Loinjak

Date: 07:43:33 01/08/03

Go up one level in this thread


First of all - I'm only a hobby correspondance player (not organized) and I'm
playing against a friend, an international correspondence player > 2300 corr.
ELO.

The long term attack I spoke about in my last posting was detected with the help
of several engines (Ruffian, C+GTiger, Fritz ) and could have been started
against me if I hadn't prevented it. I check all my moves in correspondence
chess very thoroughfully with at least C+GTiger 14, Ruffian and Fritz 6 (my box:
2 GHz, 256 MB). As I know how bad computers are in long term (over 30 ply)
strategy I often need to show them some ideas (this usually means several hours
of interactive analysis with computer help) and to let them prove tactically
that they're good.

Maybe interesting for all of you:
------------------------------------
My feeling is that the evaluation of a certain position by an engine is highly
correlated with the (positional) understanding of this special position. Thus I
favor in the analysis of a certain position the engine which thinks it has the
most advantage (I'm always analyzing from the side of the stronger color
-independent of what color I have in this game).



To come back to my original question:
------------------------------------
==> Who knows anything exact about the article of Chinchalkar???


Why do we need a good (i.e. low) upper bound for the legal postions in chess?
------------------------------------
To solve chess perfectly we might need a search from above (game tree search
from init position) combined with endgame tables (maybe up to 10-men). Assuming
a hash table of 10^23 positions (molecular/atomic/quantum  memory - maybe
available in 10-30 years) even the tree based search wouldn't need to explore
more than the root of the legal positions from 32-men down to 11-men. When we
know the exact number of those positions we could gain a good estimation of the
computational complexity and memory complexity of chess and estimate how long we
need to solve it (under the assumption that game tree search combined with
retrograde analysis is the optimal way to solve chess).

I'm working on a calculation for a sharp upper limit (in the sense of
optimization theory) of the legal chess positions. Therefor I'd like to know
what results are known so far and how they were computed.

I see this as a first step towards the solution of chess.
As I've mentioned already in another forum:
  "I want to live to see chess solved"

(Please don't ask how old I am - I've got no good answer for this. :) )


Sini



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.