Author: Amir Ban
Date: 04:30:35 01/22/98
Go up one level in this thread
On January 22, 1998 at 05:04:43, Howard Exner wrote: >On January 21, 1998 at 14:24:23, Ernst A. Heinz wrote: > >>Hi all, >> >>I am currently running the original ECM through "DarkThought" >>with 90 sec per position. In addition to the inconsistencies >>already pointed out by Amir, "DarkThought" begs to differ on >>some more positions among the first 99 positions (even after >>20 min analyses each). >> >>ECM #016 >>original solution = [ Rxc5 ] >>"DarkThought" says = [ dxe3 ] Rxd8+ Rxd8 Bb2 Qd6 ... is equally good >> (score = +0.4 after iteration #13) Right >>ECM #031 >>original solution = [ Nxd4 ] >>"DarkThought" says = [ Rd6, Qd5, Qa6, h6, h5, Kh8 ] are equally good >> (score = 0.8 after iteration #13) Disagree. Nxd4 takes a pawn safely and simplifies into an endgame with a pawn up. You are evaluating the position as already +1 for black but I don't know if everyone will agree with that. >>ECM #070 >>original solution = [ e4 ] >>"DarkThought" says = [ Rf3 ] Kh1 g5 d4 exd4 Re6 Qg7 ... >> (score = +1 after iteration #13, >> g5 directly seems to be a transposition) I tend to disagree. e4 is positional, and the point is that it's possible. It prepares a pawn storm through d5, g5 and eventually e3, but white is not dead yet. Rf3 doesn't seem to do much. >>ECM #075 >>original solution = [ Nh6 ] >>"DarkThought" says = [ Qh4, g6 ] are also good >> (score = 1.2 after iteration #13) >> Nh6 takes a pawn and opens up things. I am not able to look so deep at your alternatives, but seems to me that black keeps material balance and holds with Rd7 for both. >>ECM #099 >>original solution = [ Qxd5 ] >>"DarkThought" says = [ b4 ] Rb5 Qxd5 Be5 Qb3 Bc6 ... is equally good >> (changes between both with scores around +0.3) >> b4 just delays things for one move. Sort of horizon effect. I'm OK with adding it as a side solution. >>What do others think about these alternatives? >> >>=Ernst= > Amir
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.