Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Find a mate.

Author: leonid

Date: 18:40:28 08/03/02

Go up one level in this thread


On August 03, 2002 at 16:39:51, Uri Blass wrote:

>On August 03, 2002 at 11:42:03, leonid wrote:
>
>>On August 03, 2002 at 10:42:17, Manfred Rosenboom wrote:
>>
>>>On August 03, 2002 at 08:03:46, leonid wrote:
>>>
>>>>...
>>>>If above position is too easy for You then try other one. In it my program was
>>>>very slow and I don't know shortest mate. By selective mate was found in 10.
>>>
>>>
>>>>[D]2bqkb2/1r4rR/nPPPPPPn/1pNppNp1/bNp2pNb/8/8/7K w - -
>>>
>>>Yace 0.99.56 indicates at least mate in 9:
>>>
>>>D1  00:00  -16.22   Rh7xg7 Rb7xg7 f6xg7
>>>D1+  00:00  -16.15   Nf5xg7 Bf8xg7 c6xb7 Bc8xe6 b7b8Q Na6xb8 f6xg7
>>>D1  00:00  -14.62   Nf5xg7 Bf8xg7 c6xb7
>>>D1+  00:00  -14.59   d6d7 Rb7xd7 e6xd7 Rg7xd7 c6xd7 Bc8xd7
>>>D1  00:00  -11.04   d6d7 Rb7xd7 e6xd7 Rg7xd7 c6xd7 Bc8xd7 Rh7xd7
>>>D2+  00:00  -10.63   d6d7 Rb7xd7 c6xd7 Bc8xd7 f6xg7 Qd8b8 g7xf8Q Ke8xf8 e6xd7
>>>Na6xc5 d7d8Q Qb8xd8 Ng4xh6 Qd8xb6
>>>D2+  00:00  -9.62   d6d7 Qd8xd7 e6xd7 Bc8xd7 f6xg7 Bd7e6
>>>D2  00:00  -9.11   d6d7 Rb7xd7 c6xd7 Bc8xd7 f6f7 Nh6xf7 e6xd7 Qd8xd7 Nc5xd7
>>>Rg7xh7
>>>D3+  00:00  -8.68   d6d7 Qd8xd7 e6xd7 Bc8xd7 Nf5xg7 Ke8d8
>>>D3  00:00  -7.88   d6d7 Rg7xd7 e6xd7 Rb7xd7 c6xd7 Qd8xd7 f6f7 Ke8d8
>>>D3+  00:00  -7.87   Nf5xg7 Rb7xg7 d6d7 Qd8xd7 c6xd7 Ke8d8 d7xc8Q Kd8xc8 f6xg7
>>>Na6xb4
>>>D3+  00:00  -7.68   Nf5xg7 Bf8xg7 d6d7 Bc8xd7 e6xd7 Ke8f8 Nc5e6 Kf8g8 Ne6xd8
>>>Rb7xb6 f6xg7 Nh6xg4 Nb4xd5
>>>D3  00:00  -5.75   Nf5xg7 Bf8xg7 d6d7 Bc8xd7 e6xd7 Rb7xd7 c6xd7 Qd8xd7
>>>D4  00:00  -5.75   Nf5xg7 Bf8xg7 d6d7 Bc8xd7 e6xd7 Rb7xd7 c6xd7 Qd8xd7 Nc5xd7
>>>Nh6xg4 Nb4xa6
>>>D4+  00:00  -5.74   d6d7 Rb7xd7 c6xd7 Qd8xd7 e6xd7 Ke8d8 d7xc8Q Kd8xc8 f6xg7
>>>Bf8xg7
>>>D4+  00:00  -5.35   d6d7 Rb7xd7 c6xd7 Qd8xd7 e6xd7 Ke8d8 d7xc8Q Kd8xc8 f6xg7
>>>Bf8xg7
>>>D4  00:00  -5.04   d6d7 Rb7xd7 c6xd7 Qd8xd7 e6xd7 Rg7xd7 f6f7 Rd7xf7 g6xf7
>>>Nh6xf7 Ng4f6 Ke8d8
>>>D4+  00:00  -5.02   f6f7 Nh6xf7 d6d7 Rb7xd7 c6xd7 Qd8xd7 e6xd7 Bc8xd7 Nf5xg7
>>>Bf8xg7 g6xf7 Ke8d8 f7f8Q Bg7xf8 Nc5xa4
>>>D4  00:00  -4.56   f6f7 Rg7xf7 g6xf7 Nh6xf7 c6c7 Rb7xc7 b6xc7 Qd8xc7 d6xc7
>>>Na6xc7 e6xf7 Ke8d8
>>>D5  00:01  -4.27   f6f7 Rg7xf7 g6xf7 Nh6xf7 c6xb7 Bc8xe6 Nc5xe6 Nf7xd6 Ne6xd8
>>>Nd6xf5
>>>D5+  00:01  -4.26   d6d7 Rb7xd7 c6xd7 Qd8xd7 e6xd7 Bc8xd7 f6xg7 Bf8xg7 Nf5xg7
>>>Ke8f8
>>>D5+  00:01  -4.16   d6d7 Rb7xd7 c6xd7 Qd8xd7 e6xd7 Bc8xd7 f6xg7 Bf8xg7 Nf5xg7
>>>Ke8f8
>>>D5+  00:01  -3.16   d6d7 Rb7xd7 c6xd7 Qd8xd7 e6xd7 Bc8xd7 f6xg7 Ke8d8 g7xf8Q
>>>Bd7e8
>>>D5  00:01  -2.15   d6d7 Rb7xd7 c6xd7 Qd8xd7 e6xd7 Bc8xd7 f6xg7 Bf8xg7 Nf5xg7
>>>Ke8d8 Rh7h8 Kd8e7 Nb4xd5 Ke7d6
>>>D5+  00:02  -2.14   Nc5xb7 Bc8xe6 f6xg7 Qd8xd6 Nf5xd6 Bf8xd6 g7g8Q Bd6f8
>>>D5  00:02  -1.06   Nc5xb7 Bc8xe6 f6xg7 Nh6g8 d6d7 Qd8xd7 c6xd7 Be6xd7 g7xf8Q
>>>Ke8xf8 Rh7xd7 Na6xb4 Nf5xh4 g5xh4 Ng4xe5
>>>D6+  00:03  -0.66   Nc5xb7 Bc8xe6 f6xg7 Nh6g8 d6d7 Qd8xd7 c6xd7 Ke8xd7 g7xf8Q
>>>Be6f7
>>>D6+  00:03  0.34   Nc5xb7 Bc8xe6 f6xg7 Nh6g8 d6d7 Qd8xd7 c6xd7 Ke8xd7 g7xf8Q
>>>Be6f7
>>>D6  00:04  2.80   Nc5xb7 Bc8xe6 f6xg7 Nh6xf5 d6d7 Be6xd7 Ng4f6 Qd8xf6 c6xd7
>>>Ke8xd7 g7xf8Q Nf5g7 Qf8xf6 Na6xb4 Qf6xg7 Kd7e6
>>>D6+  00:04  2.81   d6d7 Rb7xd7 c6xd7 Qd8xd7 e6xd7 Bc8xd7 f6xg7 Bd7xf5 Ng4f6
>>>Ke8e7 g7g8Q Nh6f7 Nc5xa4 Ke7xf6 Rh7xh4
>>>D6+  00:05  5.34   d6d7 Rb7xd7 c6xd7 Qd8xd7 e6xd7 Bc8xd7 f6xg7 Bf8e7 Ng4f6 Ke8d8
>>>Rh7h8 Be7f8 g7xf8Q Bd7e8
>>>D6  00:13  M9  d6d7 Rb7xd7 c6xd7 Qd8xd7 e6xd7 Bc8xd7 f6xg7 Bf8xg7 Nf5xg7 Ke8d8
>>>Ng7e6 Bd7xe6 Nb4c6 Kd8e8 Ng4f6 Ke8f8
>>>D7  00:57  M9  d6d7 Rb7xd7 c6xd7 Qd8xd7 e6xd7 Bc8xd7 f6xg7 Bf8xg7 Nf5xg7 Ke8d8
>>>Ng7e6 Bd7xe6 Nb4c6 Kd8e8 Ng4f6 Ke8f8 Nc5xe6
>>>D8  01:17  M9  d6d7 Rb7xd7 c6xd7 Qd8xd7 e6xd7 Bc8xd7 f6xg7 Bf8xg7 Nf5xg7 Ke8d8
>>>Ng7e6 Bd7xe6 Nb4c6 Kd8e8 Ng4f6 Ke8f8 Nc5xe6
>>
>>Much better in time and shorter that my solution. Mine found mate in 10 only in
>>3 min and 31 sec. Celeron 600Mhz. No hash.
>>
>>Cheers,
>>Leonid.
>
>still better than Movei

Your program is still better that mine since it have no specilized mate solver.
I doubt that mine will go even close with material echange to Your time.

When mate must be found 10 or 11 moves deep then difference in time even 5 fold
is completely negligeable.

>Movei found only mate in 11 and needed almost 5 minutes on p800 that is faster
>than your celeron.
>
>only after almost 10 minutes it could see mate in 10
>The pv may be wrong in the last plies because movei has some bugs about the pv
>in the qsearch that I still did not investigate.
>
>9 9979 28368 22216746 d6d7 g7d7 c6d7 d8d7 f6f7 h6f7 g4f6 e8d8 g6g7 b7b6 g7f8q
>d8c7 e6d7 b6d6 d7d8q d6d8 h7f7 c8d7 f8d6 c7c8
>9 9980 54206 42318757 c5b7
>9 9981 55899 43686927 c5b7 c8e6 d6d7 d8d7 c6d7 e8d7 g4e5 d7e8 b7d6 f8d6 f5g7
>e8f8 g7e6 f8g8 f6f7 h6f7 e5f7 a6b4
>9 9981 71121 55590196 c5b7 c8e6 d6d7 d8d7 c6d7 e8d7 g4e5 d7e8 b7d6 f8d6 f5g7
>e8f8 g7e6 f8g8 f6f7 h6f7 e5f7 a6b4
>
>The first position that you posted is not too easy and it is too hard for Movei
>that cannot find even material advantage for white in 20 minutes.

This is slight surprise for me since frist position contain many pawns and could
be considered as "safe". But it is true that nobody came with solution for it,
anyway.

For my program some positions, that are very simple for other programs, are slow
as well.

First position is mate in 10 moves. It was solved by selective in 5 seconds. By
brute force my program verified position only 9 moves deep to be sure that it is
mate in 10. Nine moves said "No" in 12 min and 41 seconds.

Leonid.

>Uri



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.