Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Try solve this mate.

Author: Heiner Marxen

Date: 21:05:57 10/22/01

Go up one level in this thread


On October 21, 2001 at 20:40:05, leonid wrote:

>On October 21, 2001 at 17:27:51, Heiner Marxen wrote:
>
>>On October 21, 2001 at 13:03:27, Paul wrote:
>>
>>>On October 21, 2001 at 01:43:54, leonid wrote:
>>>
>>>>[D]4r3/1KPQrNBR/1PQ1b1QR/1QQBbQQN/3n1n2/2q3q1/1qq3qq/1q5k w - -
>>>>
>>>>Please indicate your result.
>>>>
>>>>Thanks,
>>>>Leonid.
>>>
>>>Hi Leonid,
>>>
>>>Pretz finds a mate in 11 reasonably quick, but it takes a long time on different
>>>settings to lower that to a mate in 10 ... will have to conjure up something to
>>>speed that up:
>>>
>>>01:01 WM11 12 Nxg3+ Kg1 Ne2+ Qxe2 Qxb1+ Qce1 Qxg2+ Nxg2 Rxh2 Rb8+ cxb8=Q Rxd7+
>>>Qxd7 Bxd5+ Qdxd5 Qexb1 Qcxd4+ Bxd4 Rh1+ Kf2 Qgg3#
>>>
>>>09:36 WM10 10 Nxg3+ Kg1 Rxh2 Rb8+ Kxb8 Bxc7+ Qcxc7 Re8+ Qxe8 Qcf3 Bxd4+ Qxd4
>>>Qxd4+ Qe3 Rxg2+ Qxg2 Qxe3+ Qf2 Qbxb1#
>>>
>>>Groetjes,
>>>Paul
>>
>>Hi Leonid, Hi Paul,
>
>Hi, Heiner!

Hi Leonid,


>>Chest on my K7/600 (350 MB hash) used already nearly 9000 seconds (2.5 hours)
>>to proove "no mate in 9", so may be Pretz does not so bad, really.
>>Also, the search for Chest does appear to be somewhat unstable:
>
>Pretz do well, I even not see that many other programs to try this position.

Yes, indeed.  Pretz quite consistently is fast and mostly finds shortest
mates.  Must be a Chessmaster clone.. :-))

>Mine spent even on 8 moves already 1 hour and 38 min. Your went much better.
>
>
>>#  1      0.00s                 0kN           0.87          1-         0
>>#  2      0.00s                 0kN           1.00          1-         0
>>#  3      0.01s                 1kN [ 12.12]  0.95         68-         0
>>#  4      0.16s [ 16.00]        7kN [  8.32]  1.04        701-         0
>>#  5      1.05s [  6.56]       53kN [  7.99]  1.18       4912-         0
>>#  6     23.45s [ 22.33]     1150kN [ 21.65]  1.50     139214-         0
>>#  7    175.82s [  7.50]     8700kN [  7.56]  1.93    1111052-         0
>>#  8   1181.87s [  6.72]    58998kN [  6.78]  2.32    7380574-    367907
>>#  9   8966.72s [  7.59]   445492kN [  7.55]  2.67   54287666-  45539765
  # 10 100550.62s [ 11.21]  5020655kN [ 11.27]  2.91  592253071- 583505170

i.e. nearly 28 hours.

>>That spike at depth=6 is quite uncommon.  That the EBF at the last iteration
>>increased (most of the time it decreases), is most probably due to the fact
>>that the mate is near.  Currently Chest searches depth=10, but that will take
>>around a day.  More to come...
>
>Good! Now it is only 10 that is under the question.
>
>Your spark with depth 6 is probably more related to your special writing. Mine
>was very stable and had branching factor that went around 10. I speak for sure
>about brute force.
>
>4 moves 0549 sec
>                  10.4
>5 moves 5.71 sec
>                  10.08
>6 moves 57.63 sec
>                  10.7
>7 moves 10 min 17 sec
>                  9.58
>8 moves 1 hour 38 min 35 sec
>
>Celeron 600Mhz. No hash.

I see.  Yes, most probably you are right.  Sometimes Chest is so busy using
the defenders material to check the attacker, that after some moves all
active defenders have disappeared... and still some moves to go... uuh.
But overall the agressive style seems to pay off.

Well, Chest found 5 (!) key moves for "mate in 10":

PV: Nxg3+ Kg1 Rxh2 Rb8+ cxb8=N Rxd7+ Nxd7 Bxd5 Ne2+ Qxe2 Rh1+ Kf2 Qgxf4+ Bxf4
Qxf4+ Qcf3 Qxg2+ Kxg2 R7h2#
PV: Nxf4 Rb8+ cxb8=N Rxd7+ Nxd7 Bxd5 Rxh2+ Q2xh2 Q6xd5+ Qcf3 Rxh2+ Qcxh2 Qxb1+
Qhg1 Qxg3 Qxd5+ Qxd5+ Qg2 Qh5#
PV: Qxg3 Rb8+ cxb8=N Rxd7+ Nxd7 Nxd5 Qxg2+ Qcxg2 Ng3+ Qcxg3 Rxh2+ Kg1 Qxb1+ Qxb1
Qgxb1+ Qe1 Qxe1+ Qf1 Qexf1#
PV: Qgxf4 Rb8+ cxb8=N Rxd7+ Nxd7 Bxf4 Nxg3+ Kg1 Rxh2 Bxd5 Ne2+ Qxe2 Rh1+ Kf2
Qxf4+ Qcf3 Qxg2+ Kxg2 R7h2#
PV: Qxc2 Rb8+ cxb8=N Rxd7+ Nxd7 Bxd5 Nxg3+ Kg1 Ne2+ Kf1 Nxf4+ Qe2 Qg1+ Qxg1
Qbxe2+ Nxe2 Qgxg1+ Nxg1 Q5f2#

Most notably four of the key moves are non-checking moves (at least they are
capturing moves).  Huh, well, you didn't expect that, did you?
Looks sort of interesting!

Cheers,
Heiner


>Had tried to look into other engin "Yace". Found that it is done also for very
>patient and those that have special knowledge. Otherwise, you are out of luck.
>Never tried its mate solving capacity. I hope that somebody will come here later
>with result for that engin.
>
>Cheers,
>Leonid.
>
>
>>Cheers,
>>Heiner



This page took 0.02 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.