Author: Robert Hyatt
Date: 05:31:12 09/16/04
Go up one level in this thread
On September 16, 2004 at 07:30:26, Ricardo Gibert wrote: >On September 15, 2004 at 23:16:05, Jon Dart wrote: > >>> Qd1 is an inferior -- >> >>> after Qd1+ Kg2 Qg1+ Kf3 -- you are back to the starting position and Qf1+ is >>> clearly best...looks like a bug in Arasan >> >>There is no penalty here for repeating the position once. If a mate score was >>found things are different, since the score encodes the distance to mate. But >>without that, Arasan doesn't know that a farther away gain of material is any >>worse than a nearer gain. >> >>--Jon > >A "horizon effect" possibility would indicate this to be a bug unless you have >some other way of dealing with them. Wouldn't such repetitions have the effect >of diluting the strength of the engine by inflating the depth analyzed by 4 ply >or more in many of the lines arasan analyzed? I would think such a thing would >be worth avoiding. 4 plies into the tree 2-fold repetition should be scored the same as a 3-fold repetition to avoid this. This also points out that the program would have a problem with repetition detection if it can only see a 2-fold repetition but not the 3rd, letting it overlook a forced draw it might want to avoid or accept depending on game circumstances...
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.