Author: Sune Fischer
Date: 03:56:46 11/03/02
Go up one level in this thread
On November 03, 2002 at 05:20:31, Omid David wrote:
To solve it the search algorithm would have to be very different. There would be
no need for an evaluation other than {-mate,draw,mate}. There would be no need
for iterative deepening since each branch should be searched to the end anyway.
Although using a transposition table would increase speed, it wouldn't be a
requirement, information about window size (alpha-beta) is all that is needed
for "storage".
Secondly, even though there may be around 10^128 positions, proabably only
around sqrt(10^128)=10^64 would need to be checked.
Still a huge number of course.
-S.
>The game of chess can never ever be solved:
>
>There are about 10^128 potential chess positions. If we start searching with a
>supercomputer with the speed of 100 million nodes per second (10^8 NPS), it will
>take about 10^113 years to process all possible positions! What is the speed you
>can imagine in the next 100 years? Let's say 100 million million nodes per
>second (10^14 NPS); then it will take "only" 10^107 years to solve the game of
>chess!
>
>And even if we process all 10^128 possible positions, we will have one little
>problem: where to store the data?! Even if we manage to store a position in an
>atom, there won't be enough atoms for that, since there are "only" 10^80 atoms
>in the entire universe...!
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.