[Retros] 50-moves draw revisited

[Rol, Guus]

Tom, my views are not so dramatic. I would not wish to include "the
significances" of a positions into this particular rule. By that I mean
that I would not consider to include anything that would require
positional analysis or playing strength. In chess, there are three
different rules that concern themselves with the time/change-management
aspect of the game (a) repetition - no progress (b) 50 moves rule - no
essential progress (c) dead positions - no future progress. Only the
third one is designed to let in some understanding of the game with
regard to what might happen. Whatever elevated concepts they derive
from, rules should preferably be simple in an intellectual game like
chess. The message that I wanted to convey is that most rules are
inevitably imperfect projections of the concepts they represent, and
that some rules - like the current 50-moves rule - score even below par
on that scale of imperfection. Songs of praise for the "(ir)reversible
moves" concept are inappropriate, whereas the same concept applied to
positions is an improvement. I would include in this evaluation only the
static properties of the positions involved, not the dynamic conversion
process from initial to end state. I mentioned the impossibility of some
position conversion only to illustrate the limitations of the
"reversibility" concept.  The power of the positional approach is not as
much in its outcome as in its procedural superiority. Before the
castling query, the 50-moves law might have been the same under both
approaches, but never ever could the mistake have been made of
overlooking the essential (and irreversible) change of the loss of
castling rights when examining position properties! And the afterburning
opportunity for an initial e.p. move would at least have been under
consideration.
 
I have considered the interaction of the rules you mention. Instead of
helping you I will make things a little worse. What if the stipulation
reads "white to mate in N moves after 3 consequtive black rook moves" or
"white to win after giving black 12 capturing opportunities"...  Surely
a bookkeeping progress indicator could detect tangible progress in some
positions otherwise rated as "not essentially different" (similar). The
problem we are facing here is essentially the same as in the Russell
paradox, the intertwining of objects and containing objects. Here it
manifests in the interaction between a "system" and "its evaluation
functions". We are asked to evaluate a system which behaves depending on
parameters in the process of its evaluation. An impossible knot. Scoring
evaluators should not be allowed to generate intermediate input for the
behaviour of the system. Chances are it will start to oscillate or
generate a paradox. The solution, in my view, is the same as in the
previous paragraph. Stick to simplicity in the rule-making. Only allow
into consideration the variables that contribute effectively to the
purpose of the evaluation.
 
To show you the destructive nature of some stipulations consider a
position in which:
(a) mate in 3 is possible if white castles
(b) mate in 2 is possible if he plays e.p. (not allowed since
unprovable)
(c) it is provable that if white has the right to play 0-0-0, he can
also play e.p.
(d) the question is: "what is the minimum amount of moves for white to
mate?"
You say "3" and I say "yeah, by castling but then faster by e.p.!"
Or you say "2" and I say "yeah by e.p. but that is illegal for you can't
prove it!"
 
Guus Rol.

	-----Oorspronkelijk bericht-----
	Van: retros-bounces at janko.at [mailto:retros-bounces at janko.at]
Namens pastmaker at aol.com
	Verzonden: vrijdag 5 januari 2007 17:01
	Aan: retros at janko.at
	CC: pastmaker at aol.com
	Onderwerp: Re: [Retros] 50-moves draw revisited
	
	
	An interesting idea.  Is it your view, Guus, that any move that
has the potential to deprive the players of the opportunity to reach any
particular position should interrupt the counting?  For example, an
early Ke1-d1 might later prevent the WKR from getting to the other side
of the WK.
	 
	That seems reasonable enough to me.  After all, although a
capture prevents the players from thereafter reaching any positions
having the total number of units that were on the board immediately
prior to the capture, it nevertheless allows them to reach quite a few
other postions. 
	 
	What would we think about a move that reaches a given position
for the second time?  Would we incorporate here the notion that the
third repetition would terminate the game as a draw, and that,
therefore, a move that causes a position to occur twice is a
sequence-interrupting move for purposes of the 50-move draw rule?  (Of
course every move causes a position to occur the first time, so
presumably we would factor that out altogether even if we would accept
the second-occurrence as an interruption.)  By the way, as I am not at
all familiar with the 3-repitition rule, does it matter who is on the
move when a position recurs?
	 
	Regards,
	Tom Volet
	 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.pairlist.net/pipermail/retros/attachments/20070108/0e8137af/attachment.htm>