Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Result after 6 hours and a bit.

Author: Jeremiah Penery

Date: 10:11:54 11/29/00

Go up one level in this thread


On November 29, 2000 at 10:34:42, Mogens Larsen wrote:

>On November 29, 2000 at 08:10:24, Uri Blass wrote:
>
>>My prediction was correct.
>
>Yes :o).
>
>>You can see from the following posts that I also predicted that this will be the
>>score at depth 13 when the sore at depth 14 will be positive but smaller and
>>Gandalf will reconsider g6 at this depth but without finding Kh2
>
>I tried running the position after Nxh6 c3 Nf5 cxb2 Qg4, since both variations
>had these moves in common. The analysis was stopped after an hour.
>
>Analysis by Gandalf 4.32g:
<SNIP>
>3...g6 4.Kh2 Qd7 5.Nh4 Rd8 6.Nxg6+ Rxg6 7.Qxg6 Qg7 8.Qh5+
>  ±  (1.31)   Depth: 9   00:02:15  20930kN
>3...Bc8
>  ±  (1.29)   Depth: 9   00:05:13  20930kN
>3...Bc8 4.Qh4+ Rh6 5.Nxh6 gxh6 6.Ree3 Qe7 7.Qxe7 Bxe7 8.Nxd6
>  ±  (1.04)   Depth: 9   00:05:26  20930kN
>3...Rf6
>  ±  (1.02)   Depth: 9   00:07:15  20930kN
>3...Bc8 4.Kh2 g6 5.Qh4+ Kg8 6.Ng5 Ra7 7.Nxe6 Bxe6 8.Nh6+
>  +-  (1.74)   Depth: 10   00:12:05  121488kN
>3...g6
>  +-  (1.72)   Depth: 10   00:13:13  121488kN
>3...g6 4.Re2 Qd7 5.Nh4 Bc6 6.Nxg6+ Rxg6 7.Qxg6 Qg7 8.Qh5+
>  ²  (0.57)   Depth: 10   00:16:56  121488kN
>3...g6 4.Kh2 Qd7 5.Nh4 Bc6 6.Nxg6+ Rxg6 7.Qxg6 Qg7 8.Qh5+
>  +-  (1.50)   Depth: 11   00:37:10  354951kN
>
>It doesn't get all the moves right, but it discovers the impact of Kh2 twice
>(depth 9 and 11). I imagine it'll switch between Bc8 and g6 for a while, which
>will take a long time.

Yeah, Bc8 and g6 are the two best defenses, and they're roughly equal I think.
It will be hard for any program to stick with one of those forever.  The really
cool thing is that I think the original position is something like mate in 18 or
less.  I haven't proved this, of course, but the lines in the Informant analysis
lead to it, and it seems to be pretty thorough analysis.



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.