Author: Paul Richards
Date: 12:27:41 06/07/99
Go up one level in this thread
On June 07, 1999 at 14:03:29, Dann Corbit wrote:
>Here are two checkmates seen from 3 ply for 4 ply {2 moves}:
>rnbqkbnr/pppp1ppp/4p3/8/5PP1/8/PPPPP2P/RNBQKBNR b KQkq g3
>rnbqkbnr/pppp1ppp/8/4p3/5PP1/8/PPPPP2P/RNBQKBNR b KQkq g3
>
>And here are some more for seen from 3 ply for 6 ply {3 moves}:
>rnbqkbnr/pppp1ppp/4p3/8/5P2/7P/PPPPP1P1/RNBQKBNR b KQkq -
>rnbqkbnr/pppp1ppp/4p3/8/5P2/7P/PPPPP1P1/RNBQKBNR b KQkq f3
>rnbqkbnr/pppp1ppp/4p3/8/8/5P1P/PPPPP1P1/RNBQKBNR b KQkq -
>rnbqkbnr/pppp1ppp/8/4p3/5P2/7P/PPPPP1P1/RNBQKBNR b KQkq -
>rnbqkbnr/pppp1ppp/8/4p3/5P2/7P/PPPPP1P1/RNBQKBNR b KQkq f3
>rnbqkbnr/pppp1ppp/8/4p3/8/5P1P/PPPPP1P1/RNBQKBNR b KQkq -
>
>It is apparent that these particular formations need no further exploration,
>even though white could wiggle out if black were to play stupidly enough.
>
>It seems interesting that this dead brances appear so early. So I am >wondering, if we keep tracing down branches with perfect play {from a given >position forward the play is perfect -- not necessarily so to get there},
>how many moves will be in the full tree? Will it contain the vaunted 10^42, >something fairly close to that number, or some tiny subset of that number?
I vote for a tiny subset, though of course voting is not meaningful. :)
I think the amount of non-lopsided positions is a small fraction of
the amount of legal positions. The nature of the game is such that
a bad position has effects far into the future of the game, and with
perfect play a disadvantage/weakness will ultimately lead to a loss.
As play becomes more perfect the window of what constitutes a
winning or losing position will shrink, as will the number of viable
lines. Today giving DB a position with a 200 centipawn advantage
might be hopeless. In the future a 100 centipawn advantage might
be insurmountable, etc.
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.