Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Endgame tablebases

Author: jefkaan

Date: 08:46:33 05/21/05

Go up one level in this thread


On May 21, 2005 at 09:45:32, James T. Walker wrote:
>defined as perfect chess.  In my opinion, the only way to play perfect chess is
>to solve the game so that when anyone plays 1.e4 , the "perfect chess" player
>announces mate in 73.

it would be more like 173 or more; and then only as a rought estimate,
being able to 'outbook' the other player, GM, or engine in a
more or less 'fundamental' way; such as being almost certain
to be able to arrive at a winning endgame, whatever black's
*conventional* defence would be.

In case of unconventional defences, like the Pablo or Nemeth
cases, well that would be much more difficult, and not possible
with such an approach, at least not in the foreseeable future.
On the other hand such 'imperfect' chess might well be refutated
by a strong engine with anti-human setting on in such cases,
eg. preferring open positions, or whatever. The best programmers
certainly would know how to do that, also against P&N,
although it probably will need some more work to be done.

a direct connection from the opening book into endgame tablebases ?
forget it, maybe for some queen exchange opening lines 9and even
that is doubtful) , but certainly not in other cases.
Some time ago i think Tony Werten (Xinix) believed in it
and tried it (he know has another approach i think),
and Hyatt would be interested but knows it is impossible.

Yet wat *is* theoretical possible, is that after the opening,
a extremely strong engine takes over, and then will be
able to crush black, even when playing against itself.
That imho would pretty much already approach 'perfect chess',
Jef




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.