Author: leonid
Date: 20:04:12 10/02/01
Go up one level in this thread
On October 02, 2001 at 22:02:56, Paul wrote: >On October 02, 2001 at 20:00:24, leonid wrote: > >>On October 02, 2001 at 10:07:56, Paul wrote: >> >>>On October 01, 2001 at 12:56:43, leonid wrote: >>> >>>>Hello! >>>> >>>>Today I wanted to put some simple mate position. This one, at least, have very >>>>unsophisticated shape. >>>> >>>>[D]8/8/3QQQQK/3nbrpn/3pQQBQ/3bQrqq/3qRqNB/3qRnNk w - - >>>> >>>>Please indicate your result. >>>> >>>>If you happened to have some version of Chess Master that do brute force search, >>>>please find for me time that it take for him to find mate in 4 and 5 moves for >>>>this position. I need this in order to have some idea what is his branching >>>>factor. I tried to do this myself with CM5000, that I have, but found no >>>>possibility to indicate my position. Since CM is the most performant and the >>>>most frequently mentioned program that solve the mate, I would like to know his >>>>basic mate solving capability. >>>> >>>>Thanks, >>>>Leonid. >>> >>>Hi! >> >>Hi! >> >>>Haven't got CM6000 installed at the moment, and don't remember if it's got a >>>mate level search, but John Merlino is the best person to answer that. >>> >>>For Pretz this is a difficult problem, for now it finds this: >>> >>>06:18 WM9 11 Nxh3 Ndxf6 Bxf3 Ng4+ Qfxg4 Bg7+ Qxg7 Rf6+ Qgxf6 Bxe4 Nxf2+ Qxf2 >>>Nf4+ Bxf3 Qgxf3+ Qxf3 Qxf3# >>>20:52 WM8 11 Bxf3 Ndxf6 Bxg3 Ng4+ Qfxg4 Bg7+ Qxg7 Rf6+ Qgxf6 Nh2 Nf4+ Nxf3 >>>Qgxh3+ Qh2 Qxh2# >> >>I am not sure what you found but mine looked only by selective. I don't know >>what is the minimal number of moves for this position. Mate by selective was >>found 10 move deep. It took pretty long time - 66 seconds. > >Oh ... I didn't realize you couldn't read my output format, I'll explain: > ><minutes:seconds> <evaluation> <ply> <principal variation> > 06:18 WM9 11 Nxh3 Ndxf6 Bxf3 Ng4+ Qfxg4 ... Qxf3# > 20:52 WM8 11 Bxf3 Ndxf6 Bxg3 Ng4+ Qfxg4 ... Qxh2# It is good that you explained. I was lost in 'W', 'M' and only after your indication could find what it is. >The evaluation 'WM9' means 'W'hite 'M'ates in '9' moves, and WM8 ... you got it! >So Pretz first found a mate in 9, and after that on the same ply a mate in 8. > >I checked later today without hash, just to be sure, and I think the mate in 8 >is correct. Also tried some different settings that made everything quite a bit >faster (just over 2 minutes for the mate in 9), but found nothing shorter than a For this position I actually went only 6 moves deep that took already 13 min. Branching factor was 13 and look like that it have tendency to diminish. Between four and five moves it was almost 16. My guess is that if the mate is there in 8 moves then it could take around 3 hours to reach first solution. For now I am not that sure what to do first, look seven moves deep, or go directly 8 moves for mate. >mate in 8. > >Maybe you can check with LLChess's brute force search? I started our Champion! >solver Chest on this problem a few minutes ago, but don't see a way to set the >size of the hash table(s) yet. The documentation mentions an -M switch, but >Chest doesn't seem to accept that in a DOS box under XP. It is good that you, at least, can use it. I am not at all knowledgeable of all abreviations and expressions that are in its manual. So for me the best strategy for now is to wait for Chest future edition. Then it will be easy for me to recognize all his strange expressions by looking stright into his graphics interface. >Hmmmm ... Heiner, hope you read this? It's now taking about 42MB of memory ... >started it like this: 'chest319 -b -Z8 x.epd'. Also tried -M350, but to no >avail. > >Oh ... and I did not find a mate for black in this position, and neither did >Pretz. :) Cheers, Leonid. >Groetjes, >Paul
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.