Author: Roger Brown
Date: 08:50:03 05/02/05
Go up one level in this thread
On May 02, 2005 at 08:45:00, John Brownlee wrote: >Friends, > >Is there a convention in general use for engine tournaments concerning the use >of resignation settings? > >Some engines have bugs mating lone kings, and when they have advantage they will >usually win if lone king resigns, but if lone king refuses to resign the game >will end in a draw. (Specifically List512 and SmarThink0.17a have similar but >different bugs in mating lone kings.) > >So it comes to a question of tournament philosophy. Should the bug be "repaired" >by the tournament master by applying common resignation settings (say -900 for 5 >moves) so the engines with bugs get the win? or should tournament require all >mates to be played to completion, so that the engines with bugs are made to pay >the price by suffering a 50 moves draw against a lone king? > >What difference in rating points might be expected between the two settings? > >Thank you in advance for your interesting comments. > >Regards, >John Hello John, There will others far more knowledgable than I but as a user I will state that if an engine cannot mate a lone king with (say) king and rook then that engine is not even a novice and I should know because I could eventually execute such a mate and I am a novice (alright, so maybe a touch lower, sue me!). The chess engine is an entity and if it cannot mate with the bare essentials then why should it win because the gui can provide it with a way out? Having said that, please yourself. It is your machine, software and time. Do whatever makes you happy. Somebody is always going to disagree anyway! :-) Later.
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.