Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: can a program solve this mate in 17.5?

Author: Jeremiah Penery

Date: 15:21:16 02/12/99

Go up one level in this thread


On February 12, 1999 at 17:27:10, Paulo Soares wrote:

>On February 12, 1999 at 12:33:53, blass uri wrote:
>
>>8/p2q4/4pQ2/p2p4/p1pQ1K2/2p2b2/1N1N1n1p/2B4k b - - 0 1
>>
>>This is a study I composed a few years ago.
>>
>>black say mate in 17.5 against itself.
>>I used Genius3 for many hours to help me to compose and to check the endgame.
>>
>>The main line is 1...e5+ 2.Kxf3 Qh3+ 3.Kxf2 Qg2+ 4.Ke1 exd4 5.Qf1+ Qxf1 6.Kxf1
>>cxb2 7.Bxb2 c3 8.Nb1 cxb2 9.Kf2 a3 10.Nxa3 a4 11.Nb1 a3 12.Nxa3 d3 13.Nb1 a5
>>14.Nd2 b1Q 15.Nxb1 a4 16.Nd2 a3 17.Nf1 a2 18.Ng3#
>>
>>Uri
>
>Uri,
>
>	Fritz5.32  has problems with the promotions and it does not
>obtain.
>	Genius6 has more knowledge in that it says respect to the
>promotions and goes more far, look:
>
> PII-300 Ht=64Mb
>
>	After  8... cxb2
>	       9.Kf2 #10 depth16    14sec.
>
>	After  7... c3
>	       8.Nb1 #11 depth18   133sec.
>
>	After  6... cxb2
>	       7.Bxb2 	 depth20 >1200sec.
>
>	I do not believe a program can solve this problem. Repairs
>that I gave up after 20 minutes when Genius6 probably would
>find #12 but, which would be the time necessary to find mate in
>17,5?  probably many months (years?).  It is a problem that requires
>much depth of analysis, and still has possible promotions of
>pions on the part of the blacks.

I think Uri is right when he says 5. Qf1+ is the hardest move for the computer
to find.  After making this move for it, Crafty finds a -Mat13 (Mated in 13)
with the exact PV that Uri gives above.  It did take a while (Maybe an hour or
so, but I don't remember).  I'm running it now from the beginning position with
a Crafty I modified specifically to try and solve this position.  Currently it
is on ply 10, resolving a fail low.  Here's some output:

10     1:35     --   1. ... e5+
10   181:15   1/31   1. ... e5+

It got to this ply in about 1 minute, and has been resolving the fail low for
the last 3 hours. :)  Maybe it'll solve it by Monday.

Jeremiah



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.