Computer Chess Club Archives




Subject: Re: here it is.

Author: leonid

Date: 19:58:10 01/17/01

Go up one level in this thread

On January 17, 2001 at 22:27:36, Heiner Marxen wrote:

>On January 17, 2001 at 19:41:34, leonid wrote:
>>On January 17, 2001 at 18:43:43, Paul wrote:
>>>On January 17, 2001 at 18:18:34, leonid wrote:
>>>>On January 17, 2001 at 18:10:38, Paul wrote:
>>>>>My program says mate in 9 for black ... here's a line:
>>>>>1... Qxb4 2. Rg8+ Ne8 3. Nxe4 Ncxb3+ 4. cxb3 Nxb3+ 5. axb3 Qxa3+
>>>>>6. bxa3 Qcc3+ 7. Nxc3 Qxc3+ 8. Ka2 Rd2+ 9. Bc2 Qexb3x
>>>>Can you indicate by what program you solved and all basic data about solution?
>>>>Selective or brute force search, hash table, special mate solver or usual chess
>>>>game. Speed of your computer.
>>>Hi Leonid ...
>>>It's my own program 'Pretz', selective nullmove search, 48MB hash, usual
>>>chess program, 400MHz PII. So it's possible there's a shorter mate.
>>>Don't know the exact solving time, cause I had more things running
>>>at the same time, but it was a couple of minutes.
>>Pretty new for me to see that nullmove could be used as some kind of selective
>>search for mate.
>>It is possible that 9 moves is the shortest move. I only found that mate is
>>between 11 moves and 7. Six moves I looked through brute force. In six mate not
>>Brute force, even 6 moves took on mine around 6 minutes. Branching factor looked
>>very bad for this position. With similar branching factor and having no hash in
>>the program, I thought that looking for exact solution will be too long.
>Chest says there is no mate in 7.  214 seconds K7/600 with 20 MB hash table.
>Since I have already running a long job, it will take some time to
>complete depth=8.

Pretty good!

Heiner, and how much you think that you can save the time on your mate solver
with hash while looking 7 moves deep? I speek about brute force.

Recently I was puzzled and disappointed when I found, in one initial trial of
hash table (still few questions to solve there), that hash can speed my program
(not mate solver) only inside of few initial plys. More deeply plys have too
efficent branching factor to be happy with hash. I probably miss there something
very simple.


>>Please indicate address of your program, if it is complet chess program (not
>>engin) and if it free. Curious to see if I already know it.

This page took 0.01 seconds to execute

Last modified: Thu, 07 Jul 11 08:48:38 -0700

Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.