Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: "A fantastic mate in seven"

Author: leonid

Date: 15:28:37 03/12/00

Go up one level in this thread


On March 12, 2000 at 08:36:34, Heiner Marxen wrote:

>On March 11, 2000 at 12:13:44, leonid wrote:
>
>>On March 10, 2000 at 16:34:35, José Antônio Fabiano Mendes wrote:
>>
>>>Source:
>>>http://www.correspondencechess.com/games/kopecg0.htm
>>>Kopec-Sowrey,analysis,"North Athantic Team Tournament",1990
>>>[D]2r1k2r/Q1P1n1pp/1pq5/5b2/2B2n2/8/PP3PPP/K2RR3 w - -
>>>35.Rd8+ Rxd8 36.Rxe7+!! Kxe7 37.c8N+!! Kf6 38.Qe7+ Kg6
>>>39.Bf7+ Kh6 40.Qh4+ Nh5 41.Qxh5#
>>>Recently people have been discussing here the underpromotion
>>>theme,I think this position is a worthy addition to the files.
>>>        Enjoy,JAFM
>>
>>
>>Interesting position. It is curious since the "quick logic" take it and response
>>is instant. On LLCHESS on AMD 400Mhz time should be well beyond 0.055 sec.,
>>since could see only zeros in time indicator. Brute force solution could take
>>much longer time.
>>
>>Leonid.
>
>Yes, much longer, but still fast enough for me:  on a PII/550 chest 3.19
>needs 12.5 seconds to verify there is no mate in 6, and 117 seconds to
>verify the unique key Rd8+ for the mate in 7.
>
>Curious question, leonid: is your "quick logic" just restricting the
>attacking side (normally white) to always say check (putting the black
>king "under fire") ?  If so, that is what I would call "Schachfolge-Matt"
>in German.  Can someone suggest what would be the proper English term?
>Many years ago I did experiment with this restriction, and found it
>to be very fast, but also rather limited in use.  Allowing at least one
>quiet move would be much better.  Sometime I'll try to implement that
>in chest.
>
>Heiner

Hello, Heiner!

Actually I wrote at least some 40 or fifty different version of "quick logic"
for solving the mate. All of them worked on different bases. Some started their
search by checking moves, that are all the time at the head of the moves chain.
Some did its recognition of the best moves by "affection" of the king's lignes,
and so on. Finally, I left in my "quick mate solving logic" only four
variations. One goes after checking moves first and the three others are some
kind of mixture of "quick logic" and "complet logic".

About official name it is better not to ask. Only when I met you here that I
found that you name your logic, that do something very close to mine, as "mate
solver". My logic I did completely alone and named it after the name that sound
to me as the most descriptive.

What is good with the "quick logic" is that it take the most part of the mate
that existe in normal game. It never lie to you as "quick positional logic".
Quick positional logic will indicate you some move even if good move this logic
was not able to find. Quick logic for mate will indicate you the move only when
it found for sure inevitable mate. When not is will stay silent. It is very
quick. 13 moves revision through the one "quick mate logic" is done usually in
less that 0.055 sec.

You must for sure try your own "quick logic". It permit to drop many pieces on
the board and find almost instantly mate in it. Put in your quick logic checking
moves at the head and try them first. Limite also the number of those first
moves that your logic see.

Description in German name sound to me as quit accurate.

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.