Author: Howard Exner
Date: 15:13:48 06/13/99
Go up one level in this thread
On June 13, 1999 at 14:13:31, Eran wrote: >I received a new Hiarcs 7.32 CD-ROM by mail today, and I happy with it. > >I did try the position shown below with Hiarcs 7.32. It successfully found the >correct move g4 in a few minutes! I am impressed :) I've always liked this position as a test for computers. It reveals how different programs choose to extend certain positions that have threats to the king in order to search a bit deeper. I ran this on Rebel 10c(AMD K6-233 40MB hash) and its initial choice was Bxa8 (+6 something eval). Then the eval dropped (+4 something) and I thought that it would soon dismiss the move as the black reply Ng3+ was in its search line. However it just chose to stop extending. CM6000 on the other hand sorts out the flaw in Bxa8 rather quickly and plays g4 in 21 seconds (AMD K6-233 16MB hash). This position is not really that taxing as the move choices after Ng3+ (given that white plays Bxa8?) are very limited. CM6000 must be extending the line as its knowledge of king safety in my view is quite good. I forget how other programs did on this position as it was covered here in CCC awhile back. There were more then a few programs that had trouble with this one. > >r1bq1rk1/8/6p1/1PN2n2/p2pB2p/3Nb3/P5PP/1R1Q1R1K w - - 0 1 > >wKh1,Qd1,Nc5,d3,Be4,Rb1,f1,Pa2,b5,g2,h2/bKg8,Qd8,Nf5,Bc8,e3,Ra8,f8,Pa4,d4,g6,h4 > >Congratulations to Mark Uniake. > >Eran
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.