Computer Chess Club Archives




Subject: Re: Mate in 46

Author: Heiner Marxen

Date: 15:10:40 01/31/01

Go up one level in this thread

On January 31, 2001 at 16:08:26, Paul wrote:

>On January 31, 2001 at 10:25:50, Heiner Marxen wrote:
>>On January 30, 2001 at 17:15:45, Olaf Jenkner wrote:
>>>Perhaps I had in mind this problem what is easily to prove that there is
>>>no shorter mate.
>>>Pervakow/Sum.  #46
>>>[D]2N5/p4P2/p4RqP/P2krb2/5RNK/2PPBnP1/3P1pP1/5Q2 w - - 1 0
>>Yes, that can be.  Much to my own surprise Chest really proves this
>>as a mate in 46 moves, and needs just 8.2 seconds (K7/600) for it.
>Hey, congrats!! How does that feel? ... great doesn't it :)

Yes, great!  :-)  :-)
I'm still a bit puzzeled, though.  When I find the time, I will look at
the search tree in more detail.

>>It gives the following PV:
>> 1. gxf3  Qh5+
>> 2. Kxh5  Bg6+
>> 3. Kxg6  Rg5+
>> 4. Kh7   Rg7+
>> 5. Kh8   Rh7+
>> 6. Kg8   Rh8+
>> 7. Kg7   Rh7+
>> 8. Kf8   Rh8+
>> 9. Ke7   Re8+
>>10. Kd7   Rd8+
>>11. Kc7   Rd7+
>>12. Kb8   Rb7+
>>13. Ka8   Rb8+
>>14. Kxa7  Rb7+
>>15. Kxa6  Ra7+
>>16. Kb5   Rxa5+
>>17. Kb4   Rb5+
>>18. Ka3   Rb3+
>>19. Ka2   Rb2+
>>20. Ka1   Ra2+
>>21. Kb1   Rb2+
>>22. Kc1   Rb1+
>>23. Kc2   Rb2+
>>24. Kd1   Rb1+
>>25. Ke2   Re1+
>>26. Kxf2  Re2+
>>27. Kg1   Rg2+
>>28. Kh1   Rg1+
>>29. Kh2   Rh1+
>>30. Kg2   Rg1+
>>31. Kh3   Rxg3+
>>32. Kh4   Rh3+
>>33. Kg5   Rh5+
>>34. Kg6   Rg5+
>>35. Kh7   Rg7+
>>36. Kh8   Rh7+
>>37. Kg8   Rh8+
>>38. Kg7   Rh7+
>>39. Kg6   Rg7+
>>40. Kh5   Rg5+
>>41. Kh4   Rh5+
>>42. Kg3   Rh3+
>>43. Kf2   Rxf3+
>>44. Ke2   Rf2+
>>45. Nxf2  Ke5
>>46. R4f5#
>Very funny line!
>>I even had to increase the maximal allowed depth in Chest (formerly 30).
>>Quite an exceptional position!
>I also increased my maxdepth to 256 plies before trying the Mate in 120 one :)
>But alas ... it didn't work. You can solve this one probably because it's a
>short mate everytime the Rook stops checking?

Yes: c4# or Nb6# normally.

>>Can the playing programs solve this, also?  I suspect so.
>I don't think so ... mine doesn't, anyway. Not that that says anything :)
>Maybe a singular extension program like Ferret could succeed; I think full ply
>se's might just do it. Would give it a try myself, but don't have them (yet)!

Hmm, when there is a direct mate in one, there is not much sense in searching
any other move deeper than one ply, I would think.
If Chest can span that tree, others should be able to do it, also.
I suspect that a 1-ply mate does not stop them from searching alternatives.
But it could, right?

Ah, yes, Chest always does internal iterative deepening.  That helps,
of course.  That is not very expensive for Chest, since its TT cuts most
of it off.

>Also maybe the position after the 3rd White or Black move, when the Crazy Rook
>starts checking is doable within a week or so ... I see the first 23 ply of the
>pv on my screen and that looks promising ;)

Hmm, that could be fun  ;-)

Regards, Heiner

This page took 0 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.