Computer Chess Club Archives


Search

Terms

Messages

Subject: Re the FIDE 50-move rule

Author: guy haworth

Date: 06:47:32 04/10/00

Go up one level in this thread


Bruce,

I certainly have some sympathy for what you say.

I agree that the rules should be the same for both players, that there should
not be a rule for 'an EGTB owner' and another for 'someone who does not have the
EGTB'.  I agree that you should not be able to turn up with a new EGTB just
before a tournament and get the rules changed, thus surprising your opponents.

Basically, if a k-move rule is in force, computers must take that into account.
My ICGA J paper (about to appear) covers this.

A)  it proposes SM* (and similar) as strategies, i.e.
    "minimise DTM subject to not elongating the current phase beyond k moves"

B)  it argues for a DTR ("Depth By the Rule") metric.  Against this metric, a
computer would minimise the 'k' of 'k-move rule' that it needs to avoid a
draw-claim.

Frankly, it would be surprising to see a human without EGTBs holding out against
a computer with EGTBs.  Even in KRNKNN, there is amply opportunity for the
defender to 'lose depth' by the shedload in the early stages of a deep position.
 However, it is worth exploring the theory of 'constrained optimisation' and
I've taken a first step in this direction.

Guy




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.