Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Mate in ...?

Author: leonid

Date: 03:54:14 02/16/00

Go up one level in this thread


On February 16, 2000 at 04:11:43, Andreas Stabel wrote:

>On February 15, 2000 at 17:22:40, leonid wrote:
>
>>On February 15, 2000 at 07:40:32, Andreas Stabel wrote:
>>
>>>On February 14, 2000 at 17:39:42, leonid wrote:
>>>
>>>>On February 14, 2000 at 11:29:23, Heiner Marxen wrote:
>>>>
>>>>>On February 14, 2000 at 05:08:22, Andreas Stabel wrote:
>>>>>
>>>>>>On February 13, 2000 at 18:02:20, James T. Walker wrote:
>>>>>>
>>>>>>>On February 12, 2000 at 10:18:44, leonid wrote:
>>>>>>>
>>>>>>
>>>>>>>
>>>>>>>This is from George Koltanowski's book "TV Chess".  It is a mate in 10 with 4
>>>>>>>underpromotions to a Knight.  It is one of my favorites.  I was surprised to
>>>>>>>find in the late 80's that the Fidelity Champion Sensory Challenger would solve
>>>>>>>this problem in about 42 seconds.  Todays programs will solve this in less than
>>>>>>>1 second.
>>>>>>>
>>>>>>>[D]7K/P1p1p1p1/2P1P1Pk/6pP/3p2P1/1P6/3P4/8 w - - 0 1
>>>>>>>
>>>>>>>Jim Walker
>>>>>>
>>>>>>This is a nice variation on the same position the I made from another
>>>>>>variation on this position :)
>>>>>>Note the nice queen moves at the end, and that promoting to a knight the
>>>>>>third time will not work.
>>>>>>
>>>>>>[D]8/1p1p1p1p/3P1P2/pp5P/kp6/1p4P1/1P4P1/2K5 w - - 0 1
>>>>>>
>>>>>>Regards
>>>>>>Andreas Stabel
>>>>>
>>>>>Ah, I like it!  It is a mate in 16.  Chest on P/133 with 10 MB hash table
>>>>>needs 1.15 seconds to find the main variation:
>>>>>
>>>>> g4 h6 g5 hxg5 h6 b6   h7 g4   h8=N g3 Ng6  fxg6 f7    g5 f8=N
>>>>> g4 Ne6 dxe6 d7 e5 d8=Q e4 Qb8 e3 Qxg3 e2 Kd2 =*= Qxe1 g3 Qa1#
>>>>>
>>>>>Complete solution tree in 1.67 seconds.  Wow.
>>>>>
>>>>>Heiner
>>>>
>>>>Now I see why my logic could not touch it. It is too deep to see. My mate logic
>>>>goes only 13 moves deep. But compose position as deep as 16 moves is a real
>>>>challenge. Well done!
>>>>
>>>>Leonid.
>>>
>>>Thanks, my only regret is that this is an illegal position. The configuration
>>>of black pawns can never be reached in an actual game.
>>>Notice the nice mate in 6 for white if black answers 1. g4 with 1... b6 !
>>>
>>>The position I made this problem from is the following with white to move
>>>and win. Note that it is NOT a mate in n problem.
>>>
>>>[D]8/pp1p1p2/1p1P1P2/k7/8/1p4p1/1P4PP/2K5 w - - 0 1
>>>
>>>Here it will work to promote to a knight the last time also, but if you
>>>choose to promote to a queen, the mate will be the same as in my mate in
>>>16 problem. This is of course asuming that plays for a stalemate, and close
>>>up the king position completely.
>>>
>>>Best regards
>>>Andreas Stabel
>>
>>
>>Sometime very impossible position is good to have when you write your logic. It
>>permit to find, generally speaking, if logic work well. Position with 9 queens,
>>6 rooks... I created many positions when I wrote my initial logic that generate
>>legal moves. To see if everything goes well, positions that contain inevitable
>>mate were ideal for debugging.
>>
>>Just by curiosity, how you reached your quit sofisticated positions, by "hand"
>>or with usage of some program? Did you came to the creation of the position my
>>way? I want to say that you started writing some logic and solving the positions
>>was your way to find some responses. Or you just created your positions for fun
>>since the beginning?
>>
>>Regards,
>>Leonid.
>
>The last of the positions above:
>8/pp1p1p2/1p1P1P2/k7/8/1p4p1/1P4PP/2K5 w - - 0 1
>was a problem I was shown some years ago. I liked it, but discovered that
>it was not a forced mate because black doesn't have to play for a
>stalemate. In addition I really love the version where you choose a
>queen on the last promotion, so I started fiddling with the position, first
>to make it a forced mate in as many moves as possible. When I had done
>this I found that adding the pawn at b7 giving the position:
>8/1p1p1p1p/3P1P2/pp5P/kp6/1p4P1/1P4P1/2K5 w - - 0 1
>which is shown above, would force white to choose a queen on the last
>promotion. This was just a lucky accident, but perhaps what makes this
>mate in 16 a real beauty to me. The draw-back is of course that this
>position could never be reached in a real game and that the version where
>you choose a knight at the end, which is also beautiful, was lost.
>Perhaps a small consolation is that the variation where black answers
>1. g4 with 1... b6 gives a mate in 6 with a similar theeme.
>
>I think the whole process of thinking about this position and fiddling
>with it perhaps took 5-10 years :)
>
>Regards
>Andreas Stabel

As I see now it was much more profound that I expected. I was really amazed with
the 10 moves mate and puzzled when found that the second position goes even far.
Only Marxen was able to go down to the solution. I know by experience that
finding position for more that 7 moves ahead is really tough. Was never ready
(and maybe able) to create the position by observing and thinking about it.
Everything that I did was realized through specially dedicated logic, for the
sake of making it work.

Regards,
Leonid.



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.