Author: leonid
Date: 16:43:33 10/24/01
Go up one level in this thread
On October 24, 2001 at 16:54:57, Heiner Marxen wrote:
>On October 23, 2001 at 14:23:55, Paul wrote:
>
>>On October 23, 2001 at 12:21:10, leonid wrote:
>>
>>>On October 23, 2001 at 11:03:15, Paul wrote:
>>>
>>>>On October 23, 2001 at 09:49:55, leonid wrote:
>>>>
>>>>>[D]NRqQQqRQ/kqrqnrbn/Nq4qQ/1Qn2qQ1/2QrbQ2/3BB3/8/K6Q w - -
>>>>>
>>>>>Please indicate your result.
>>>>>
>>>>>Thanks,
>>>>>Leonid.
>>>>
>>>>Hi,
>>>
>>>Hi!
>>>
>>>>Pretz (she's a TSCP-clone and proud of it ;)) says this is an easy mate in 10:
>
>:-))
>
>>>What is exactly "TSCP"? You want, probably, say that you looked into some
>>>program by doing your own, but if it is so, what exactly?
>>
>>Sorry Leonid, that remark was for Heiner, see his message from earlier today:
>>
>>http://www.icdchess.com/forums/1/message.shtml?193869
>>
>>TSCP is "Tom Kerrigan's Simple Chess Program". It has a basic board structure,
>>move generation and alpha-beta search, a (very) small evaluation function and is
>>very clearly written and easy to understand (if you can read C).
>>
>>In short, IMHO it is *the* program to read if you want to start programming one
>>yourself. I used the first version of it to start my Pretz project.
>>
>>I wish Heiner's "Chessmaster clone" remark was true, because that would mean I'd
>>know the insides of Chessmaster, boy .. that would be something! I think it's a
>>really really (ok ... one more :) really great program!
>>
>>Except for the fact that my CM6000 doesn't seem to work correctly on Windows XP.
>>Installed it just now and tried your last position, the first try took 6
>>minutes, I thought: wot? that can't be right? ... and on the further tries it
>>just froze. :( Must be doing something wrong, but what?
>>
>>>Mate is really easy but still you almost reached me on this one. Mine took 3.8
>>>sec (can be said 4 sec) to find mate by selective in 10 moves deep.
>>
>>But I used a p3/1000 and a hashtable for this, sooo ... :)
>>
>>>It is mate in 10. Verified only 9 moves by brute force, where branching factor
>>>was mild.
>>
>>I checked my solution with Chest later, it's a mate in 10 alright ... I'll leave
>>the details to Herr Marxen.
>>
>>>>00:04 WM10 07 Rxb7+ Qbxb7 Qcxc5+ Qxc5 Qgxc5+ Kxa8 Nxc7+ Qdxc7 Qa5+ Qga6 Qaxa6+
>>>>Qa7 Qaxa7+ Qxa7+ Qxa7+ Kxa7 Qhb6+ Ka8 Qa6#
>>>
>>>Cheers,
>>>Leonid.
>>
>>Groetjes,
>>Paul
>
>Well, so I have to provide more detail... After 48 minutes (K7/600, 350MB)
>Chest confirms that the above is the shortest and the only solution.
Mine was not bright on this position. I expected that it will find mate in some
30 or 40 minutes (10 moves deep, brute force) but even after 3 hours solution
was not there. I stopped waiting since result was found by selective anyway for
10. It could be that jump in branching factor that you found between 7 and 8
move happened in mine just after 9 move.
>PV: Rxb7+ Qbxb7 Qcxc5+ Qxc5 Qgxc5+ Kxa8 Nxc7+ Qbxc7 Qxc8+ Nxc8 Qa6+ Qxa6+ Qxa6+
>Qa7 Qcxc8+ Qxc8 Qxc8+ Qxc8 Rxc8#
>
>Again, there is an interesting spike in the EBF at depth 8:
>
># 1 0.00s 0kN 0.87 1- 0
># 2 0.01s 0kN 1.00 1- 0
># 3 0.03s [ 3.00] 1kN [ 12.64] 0.95 97- 0
># 4 0.19s [ 6.33] 9kN [ 6.74] 1.03 1048- 0
># 5 0.91s [ 4.79] 39kN [ 4.50] 1.19 5715- 0
># 6 3.66s [ 4.02] 161kN [ 4.11] 1.37 25375- 0
># 7 16.01s [ 4.37] 742kN [ 4.62] 1.60 106210- 0
># 8 266.52s [ 16.65] 15814kN [ 21.32] 1.40 1239680- 1
># 9 1067.19s [ 4.00] 62768kN [ 3.97] 2.16 4669068- 18973
># 10 2872.63s [ 2.69] 167700kN [ 2.67] 2.91 12986453- 4313861
Your program, for sure, see some shortcut in "specialized plys" that mine don't
see. This is why your time for 4 and 5 moves are much better that mine. My time
for brute force was:
4 moves 0.44 sec 48k (NPS)
4.2 branching factor
5 moves 2.08 66k
4.8
6 moves 10.16 sec 81k
4.01
7 moves 40.76 sec 69k
7.65
8 moves 5 min 12 sec 157k
6.9
9 moves 35 min 57 sec 146k
Celeron 600Mhz. No hash.
Cheers,
Leonid.
>I guess that black exhausted its material to check white, got mated,
>and had to search for alternate defensive measures, which helped two
>moves deeper.
>
>More detail? Here is the start of the solution tree, 9 plies deep:
>
> Rxb7+ Qbxb7 Qcxc5+ Qxc5 Qgxc5+ Kxa8 Nxc7+ Qbxc7 Qxc8+ ...+5
> Kb8 Na6+ ...+3
> Qdxc7 Qxb7+ ...+3
> Qgb6 Qhxb6+ Kxa8 Nxc7+ ...+3
> Qxb6 Qcxb6+ ...+2
> Qbb6 Qcxb6+ Qxb6 Qhxb6+ ...+2
> Kxa8 Qb8+ ...+1
> Rxc5 Qb8+ =*= Qxb8+ ...+1
> Kxa8 Nxc7+ Qbxc7 Qa5+ Qa7 Qb8+ ...+4
> Qa6 Qhxa6+ ...+2
> Qxa5+ Qdxa5+ ...+2
> Kb7 Qaa7#
> Kb8 Qca7#
> Qdxc7 Qxb7+ Bxb7 Qxb7+ ...+3
> Qxb7 Qda5+ ...+2
> Kxb7 Qfxc7+ ...+1
> Kb8 Na6+ Ka8 Qb8+ ...+2
> Qgb6 Qhxb6+ Kxa8 Nxc7+ Qbxc7 Q6a6+ ...+2
> Qdxc7 Qxb7+ ...+2
> Kb8 Qa7#
> Qxb6 Qcxb6+ Kxa8 Qb8+ ...+1
> Qbb6 Qcxb6+ Qxb6 Qhxb6+ Kxa8 Qb8+ ...+1
> Kxa8 Qb8+ Qxb8 Qdxb8#
> Rxc5 Qb8+ Qbxb8 Qxb8+ Qxb8 Qxb8#
> Qcxb8 Qxb8+ Qxb8 Qbxb8#
> Qcxb7 Qxb6+ Qgxb6 Qxb6+ Qxb6 Qb8+ Qxb8 Qxb8+ ...+3
> Qbxb6 Qb8+ Qxb8 Qxc5+ Qxc5 Qxc5+ ...+2
> Qbb6 Qxb6+ ...+1
> Qgb6 Qhxb6+ ...+1
> Rxc5 Qfxb8#
> -*- Qxb8+ ...+1
> Bxb7 Qxb6+ Qxb6 Qxb6+ Kxa8 Nxc7+ Kb8 Na6+ ...+2
> Qxc7 Qba6+ ...+2
> Kxa8 Nxc7+ Kb8 Na6+ Ka8 Qb8+ ...+1
> Qxc7 Qea4+ Na6 Qaxa6+ ...+1
> Nxa4 Qca6+ ...+1
> Kb8 Qfxc7#
> Nxb7 Qxb6+ Qxb6 Qxb6+ Kxa8 Nxc7+ Kb8 Na6+ ...+2
> Qxc7 Qba6+ ...+1
> Kxa8 Nxc7+ Kb8 Na6+ Ka8 Qb8+ ...+1
> Qxc7 Qca6+ Kb8 Qfxc7#
> Rxb7 Qbxb6+ Qxb6 Qhxb6+ Kxa8 Qb8+ =*= Qxb8+ ...+1
> Rxb6 Qxb6+ Kxa8 Qfb8+ ...+1
> Rxb6 Qxb6+ Qxb6 Qxb6+ Kxa8 Qbb8+ ...+1
> Kxa8 Qfb8+ Qxb8 Qbxb8#
> Kxa8 Qb8+ Qxb8 Qxb8+ Rxb8 Qbxb8#
> Rxb8 Qxb8+ Qxb8 Qxb8#
> Kxb7 Qxb6+ Qxb6 Qxb6+ Kxa8 Qb8+ Qxb8 Qxb8#
> Kxa8 Qb8+ Qxb8 Qxb8#
> Kxa8 Rb8+ Ka7 Qxb6+ Qxb6 Qxb6#
> Qbxb8 Qxb8+ Qxb8 Qxb8#
> Qcxb8 Qxb8+ Qxb8 Qbxb8#
>
>Except at the top level, Chest did not search for duals, i.e. alternate
>solutions for white.
>
>Cheers,
>Heiner
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.