Author: Tord Romstad
Date: 07:39:47 11/21/05
Go up one level in this thread
On November 21, 2005 at 10:14:48, James T. Walker wrote: >Hello Tord, >It seems to me that the opposite of what Uri is trying to prove is true here. >Any program with 6 man tablebases will play Kxb3 instantly with no real search >needed. For a program with 6 man tablebases, yes. But how many users do you think have the complete 6 man tablebases installed? Hardly anyone, I would wager. And even with the complete 6 man tablebases installed, the problem remains: There will be endgames with 7 pieces where the program refuses to make some obvious capture because there is a very deep tablebase mate in the resulting 6 man endgame. I am actually fairly sure this program gets much *more* serious when going from the complete N man EGTBs to the complete (N+1) man EGTBs (assuming that N is far below 32, of course). The frequency of very long and difficult wins increase with the number of pieces. Tord
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.