Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: One mate to solve for Champions!

Author: leonid

Date: 08:16:30 10/02/01

Go up one level in this thread


On October 02, 2001 at 10:19:30, Paul wrote:

>On October 02, 2001 at 10:05:25, leonid wrote:
>
>>On October 02, 2001 at 09:52:50, Paul wrote:
>>
>>>On October 02, 2001 at 08:41:20, leonid wrote:
>>>
>>>>[D]kq4BK/pq4QQ/2qq1P2/1NnqBnN1/3PpP2/2P2r2/RQ4qr/RQ4bb w - -
>>>>
>>>>I have said that this position is for "Champions" only having in mind that some
>>>>people will try to solve it by brute force. Brute force is the only way to know
>>>>minimal nomber of moves that lead to mate. But it could be not all the time
>>>>easy. When I tried this position on two professional program, I was impressed.
>>>>On Celeron 600Mhz, and with 1M of hash, one expected to solve position in around
>>>>4000 years. Other, with no hash, after around 11000000000 years of thinking
>>>>expected to recognize shortest mate.
>>>>
>>>>Please indicate your result.
>>>>
>>>>Thanks,
>>>>Leonid.
>>>
>>>Hi Leonid,
>>
>>Hi, Paul!
>>>
>>>I just now started to look at both your positions, the simple one from yesterday
>>>and this one, and I think you have a bit of a twisted nomenclature!
>>>
>>>This one is the simple one, the other one is more difficult. Even if this one
>>>seems to be a mate in 12, and the other in 9. But maybe that's only the case for
>>
>>It is mate in 12 for sure. I solved this mate by brute force in 44 minutes. 11
>>moves took 45 minutes to say that it have no mate.
>>
>>This position is more simple that previos and this is the reason why mine was
>>able to see it all 12 moves deep.
>>
>>
>>>Pretz, I'm curious about LLchess.
>>>
>>>For this one I find a mate in 12 in 3 seconds by applying some gentle force:
>>
>>Excellent! I see that your work great. When program can see so deep without any
>>mistake it is a real sign that it is basically solid.
>>
>>My selective found mate in 12 moves in 0.16 sec.
>
>That's a great time, Leonid! Your use of assembler and choice of move ordering
>and move selection mean a lot when mate solving. I still use a search and move
>order that is usual in other amateur chess programs. I also don't use any
>special ply routines, all plies are handled equal. Will have to try checking
>moves first one of these days.

I don't know exactly how other programs do move ordering, or selective search,
since I never really went to read about this. Only few indications here and
there in this Chess Club and that all. Few short glaces into other programs that
were written in the language that was not mine. Even this was done only after
writing my mate solver. What I could understand, after my short talking to some
people, is that special plys writing make all the difference. This I could see
when I found that Heiner have special plys writing and its ideas are probably
coincide with mine. His branching factor is fantastic. In the same time when I
spoke to Tom Kerrigen I understood that usually all plys done in the same
fashion. If you will look into special plys writing it could be that your mate
solving capacity will jump to the heaven. Now it is only very good.

Cheers,
Leonid.

>Groetjes,
>Paul
>
>>>00:03 WM12 12 Rxa7+ Q8xa7 Rxa7+ Kb8 Rxb7+ Nxb7 Qf8+ Qc8 Qxc8+ Kxc8 Nxd6+ Nfxd6
>>>Be6+ Kd8 Qb6+ Ke8 Bd7+ Kf8 Bxd6+ Nxd6 Qb8+ Nc8 Qxc8#
>>>
>>>With colors switched it's a mate in 2 ... ;)
>>
>>Yes. Since it is very short I even not indicated it. Since your last finding
>>(very good one) I try to not miss to verify for the opposit side, even if I
>>think that other side had no chance or have very short mate.
>>
>>Cheers,
>>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.