Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: One mate to solve.

Author: leonid

Date: 18:11:34 08/10/01

Go up one level in this thread


On August 10, 2001 at 20:09:40, Heiner Marxen wrote:

>On August 10, 2001 at 11:44:53, leonid wrote:
>
>>Hi!
>>
>>If you like, you can solve this mate:
>>
>>[D]8/nqq1QQQ1/k1qpN2R/r1qpQ2R/bqq1N3/b3Q2Q/n3Q2K/r3QBB1 w - -
>>
>>Tthis position is actually very interesting from my point of view. On  it my
>>initial speed (for brute force) is very bad and branching factor is worst that
>>usual. At least, position is not very deep to go.
>>
>>Please indicate your result.
>>
>>Thanks,
>>Leonid.
>
>Hello Leonid!
>
>According to Chest this is a mate in 7 moves.  But for this depth it is
>comparatively hard to compute: 3.1 hours on a K7/600 with 30 MB hash.
>There are two solution moves:

Will look later 7 moves. Since you have found that this is move in 7, I now see
that mate is not that far. Mine solved this position by selective in 9 moves and
it searched only 6 moves by brute force. Initial speed is low.

Moves        Time            Branching factor        NPS

3            0.1648 sec
                             54
4            8.9 sec                                 103k
                             28.1
5            4 min 10 sec                            108k
                             23.0
6            1 hour 36 min 5 sec                     145k


In this position is possible to see NPS goes higher with the depth of the
search. This is probably due to the fact that closer is the search to the
"special plys", more efficent is the search and as result NPS is much low. Since
on this position my "special plys" are unefficent, NPS stays almost twice of
what it could be expected for similar position.



>PV: Nxc7+ Qbxc7 Qxc7 Qxc7 Qxc7 Bc6 Nxc5+ Qxc5 Qxa7+ Qxa7 Qgxa7+ Kb5 Qaxa5#
>PV: Qxc7 Nc3 N6xc5+ Rxc5 Nxc5+ Qbxc5 Qxc4+ Qxc4 Bxc4+ Qcb5 Qcb6+ Qxb6 Qxb6#
>
>The EBF is rather high (>14):
>
>#  1      0.01s                 0kN           0.87          1-         0
>#  2      0.01s [  1.00]        0kN           1.00          1-         0
>#  3      0.09s [  9.00]        4kN [ 18.80]  0.99        107-         0
>#  4      2.29s [ 25.44]      127kN [ 34.76]  1.22       5672-         0
>#  5     32.63s [ 14.25]     1758kN [ 13.86]  1.55      94813-         0
>#  6    525.09s [ 16.09]    28850kN [ 16.41]  1.92    1475199-    689537
>#  7  11112.36s [ 21.16]   628820kN [ 21.80]  1.87   33822787-  33036356
>
>More hash would help the last depth, but not that much.
>Up to now you have posted only one position which has been nearly equally
>expensive at depth 7:
>Posted by leonid on February 09, 2001 at 06:51:15:
>[D]1bqQBnRn/3N2Qb/Q1QN2np/1Q1Q1qpk/4Qqbn/2B1Qrnn/2Q5/K1R5 w - -
>
>PV: Qxg5+ Nxg5 Qdxf5 Nhxf5 Qh2+ Bh3 Nf6+ Kh4 Qxh6+ Nxh6 Q4xf4+ Ng4 Qfxg5#
>
>Again, the PV contains a non-checking move.
>Also, in both cases, black has not much active counter play, and so the tree
>has a not so small EBF.
>
>That a mate in 4 costs more than 2 seconds is quite exceptional, also.
>
>That the board looks like a large "PD" is intended, I suspect :-)
>Does that mean something?

No! Reason for "PC" is that today I thought what kind of position I must compose
before I started. Then I glanced to one PC magazine for August and did my
position of "PC" shape.

Cheers,
Leonid.



>Cheers,
>Heiner



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.