Author: Thorsten Czub
Date: 10:31:42 04/04/98
Go up one level in this thread
On April 04, 1998 at 08:27:39, Ernst A. Heinz wrote: >On April 03, 1998 at 13:42:03, Howard Exner wrote: > >>4r3/1b2r1bk/pq1pP3/1p1P1NP1/n1p2pN1/2P5/1P2RQ2/4R1K1 w - - id "CSTal - >>Fritz5" >> >>This position is interesting. Did CSTal ever consider the move Nf6+ here >>instead of Re3? Perhaps other programs could verify this also as it >>looks >>like Nf6+ forces the win for white. One line runs 43. Nf6+ Bxf6 >>44. g5xf6 Rg8+ 45. Kf1 Qxf2+ 46.Kxf2 Rc7 47.f7 Rf8 48.Rg1! > >"DarkThought" instantly thinks you are right: 43. Nf6+! scores +1.94 >after iteration #11. > >PV = 43. Nf6+ Bxf6 44. gxf6 Qxf2+ 45. Rxf2 Kg6 46. Nxe7+ Kxf6 > 47. Rxf4+ Kxe7 ... > >=Ernst= You are right. There is a "bug" in chess system tal commercial cd. Whenever it is in a winning position and is able to win it will not play the winning line but instead play a draw by repetition line. I had these BUG in Aegon tournament when CSTal played 5 draws (!!) and one win or so, many draws in winning positions. I have no idea why chris was never able to fix the bug or to throw it out. It came always again into the program although he claimed that the bug is fixed. In Paris the "fixed" draw by repetition bug was SO GOOD that it was in a WEAKER position and found a draw by repetition against virtual but when it was able to PLAY the draw, it did not played the move and played a losing move instead. I have tried to explain to chris that this bug is important, especially to get better scores and results. But he was unable to fix it accurate.
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.