Author: Dave Gomboc
Date: 00:16:58 11/13/99
Go up one level in this thread
On November 12, 1999 at 12:42:04, Bruce Moreland wrote: >On November 12, 1999 at 11:55:18, guy haworth wrote: > > >>I'd also like to see the 50-move rule suspended where an engine can prove that >>it is working effectively with the appropriate EG tables. > >I am in disagreement with this. The games should be played in accord with FIDE >rules. FIDE rules at this point don't allow for 100 move mates, so the >programmers have a task ahead of them to make their programs work. I don't agree with your interpretation. FIDE rules stipulate that a tournament director may announce in advance of the tournament that players who reach certain endings may be given more than 50 moves to attempt to win them. For each such ending, the TD has to indicate how many moves will be allowed. >Someone who doesn't use the tables shouldn't have to modify their program so >that it continues to play on after a game that is drawn in accord with FIDE >rules. If mine gets to move 50 it declares a draw and stops, and I shouldn't >have to change this, since it is in accord with FIDE rules. 50 moves is a good number for humans. That doesn't make it a holy grail for machines. If a program cannot be configured to take into account TD-specified n-moves-without-pawn-push-or-capture-until-draw values in certain endings, it is that program's author's fault only. This doesn't sound terribly difficult to implement, but with the ready availabilty of code to read tablebases already, it does seem rather pointless. >Actually, probably Eugene or someone else will change the table generator >eventually, and everyone will take this change for free. > >It is interesting that at some point in one of these tournaments it will >probably be Eugene's code playing against Eugene's code, whene neither of the >competing programs was written by Eugene. > >If I am wrong, and these super-long wins are allowed, this rule change should be >given to participants well in advance of the time of the tournament. This >should not be a casual last-minute rule change. Yes, advance notice of what the situation would be helpful to all concerned. >This hasn't come up in the last few players' meetings, but I'm sure it will >eventually. > >bruce Dave
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.