Author: Uri Blass
Date: 21:51:46 11/10/00
Go up one level in this thread
On November 10, 2000 at 17:06:28, Ernst A. Heinz wrote: >Hi Uri, > >>[D]2rqr1k1/pb2bp1p/1p2p1p1/n7/B2P4/P1PQ1NP1/3B1P1P/R3R1K1 b - - 0 1 >> >>Gambittiger has no problem to see the good sacrifice of GM brown against Deep >>thought(Qd5) > >This looks like easy tactics for modern chess programs, even >for those of 1-2 years ago. > >The version of "DarkThought" that became micro vice-champion >at the 16th WMCC in Paderborn, June 1999, locks onto Qd5! in >iteration #9 after barely 10 sec. It scores the move as >roughly +1 for Black in iteration #14 after 2:30 min. > >=Ernst= I think that it is not about tactics but about evaluation. Other programs have scores close to 0.00 or even positive score for white for Qd5 and they may not find it because they prefer another move like Bc6 with slightly positive score. My memory say that I gave it to at least one of my programs one or two years ago and it could not find it. I find today that Fritz5.32 has 0.22 pawns for black with Bc6 at depth 11 and the score at depth 13 is even 0.34 pawns for black. Even if I give it the position after Qd5 Bxe8 Rxe8 Kg2 it cannot find a clear advantage for black if it does not have a long time. Hiarcs7.32's scores after Qd5 Bxe8 Rxe8 Kg2 from white's point of view Rc8 0.32/7 Rc8 0.27/8 Rc8 0.30/9 I gave it only some minutes on p200MMX but the position is not the original position. The scores in the original position from white's point of view are: Bc6 0.33/7 0.17/8 0.18/9 I guess that it cannot find Qd5 in the original position unless it can see positional drop with the score of Bc6. I thought that evaluation of pins can help to find Qd5 because after Qd5 Bxe8 Rxe8 Kg2 the white knight is pinned for a long time but it seems that there are other factors that help programs to find Qd5 by evaluation. Uri
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.