No chance

Discussion of anything and everything relating to chess playing software and machines.

Moderators: hgm, Harvey Williamson, bob

Forum rules
This textbox is used to restore diagrams posted with the [d] tag before the upgrade.
zullil
Posts: 5028
Joined: Mon Jan 08, 2007 11:31 pm
Location: PA USA
Full name: Louis Zulli

Re: No chance

Post by zullil » Wed Sep 05, 2018 10:28 pm

zullil wrote:
Tue Sep 04, 2018 8:42 pm
zullil wrote:
Tue Sep 04, 2018 10:04 am
Eelco de Groot wrote:
Tue Sep 04, 2018 12:21 am

I was hoping to get confirmation of a M38 at depth 96, but with 512 Mb it was not entirely stable and there was a 800 min delay between depth 95 and 96. This is not exactly the official Kaissa Iv in case anyone wants to verify :) (it is a one thread result, so deterministic. There is a slight difference in the code from Kaissa IV though). But the difference is not very large.
I think this is one of the positions being discussed in this thread. So far the best Cfish can do (direct search) is mate-in-37.



info depth 94 seldepth 74 multipv 1 score mate 37 nodes 1776137957024 nps 52408124 hashfull 999 tbhits 0 time 33890508 pv f2g1 a5c3 b2c3 b3b2 e2d3 b2b1b d3b1 b8b1 d1d2 a3c2 d2c2 a4a3 c2f2 f8b8 h6h7 b8b7 h7g7 g8f8 f2a2 b1b3 g5h7 f8e8 g7g8 e8d7 g8f8 c6a4 f8f7 d7c6 f7b7 b3b7 h7g5 a4d1 a2a3 b7b8 a3a6 c6b5 a6e6 b8f8 e6e7 b5a4 f6f7 d1g4 e5e6 a4b3 e7e8 g4e6 e8f8 e6f7 f8f7 b3c3 f7d7 c3d3 d7d5 c4c3 d5c5 d3d2 h1g2 d2c2 d4d5 c2b2 g5e4 c3c2 c5b5 b2c1 g1e3 c1d1 e4c3 d1e1 e3c1 g6g5 b5b8 g5g4 b8e8
Now at mate-in-34:

info depth 99 seldepth 68 multipv 1 score mate 34 nodes 3827947493918 nps 53390202 hashfull 999 tbhits 0 time 71697564 pv f2g1 a5c3 b2c3 b3b2 e2d3 b2b1b d3b1 b8b1 d1d2 a3c2 d2c2 a4a3 c2f2 c6a4 h6h7 a4d1 h7g7 g8h8 g7f7 f8f7 g5f7 h8g8 f7g5 b1b8 f6f7 g8f8 g5e6 f8e7 f7f8q b8f8 e6f8 g6g5 h1g2 d1h5 f8h7 h5g6 h7g5 g6b1 f2f7 e7d8 f7a7 a3a2 g5e6 d8e8 g1e3 a2a1q a7a1 b1e4 g2g3 e4g6 a1a7 g6f7 e6c5 e8f8 e3h6 f8g8 e5e6 f7e8 a7e7 g8h8 c5d7 e8h5 d7f6 h5f7 e7a7 f7e6 a7h7

Will let this go a bit more, to see if Cfish can find a shorter line.
In case anyone cares, Cfish indeed finds mate-in-33:

info depth 106 seldepth 66 multipv 1 score mate 33 nodes 8196236881556 nps 53770709 hashfull 999 tbhits 0 time 152429399 pv f2g1 a5c3 b2c3 b3b2 e2d3 b2b1b d3b1 b8b1 d1d2 a3c2 d2c2 a4a3 c2f2 f8b8 h6h7 b8b7 f2a2 b1f1 a2a3 b7b1 h7g7 g8f8 g7f7 f8e8 f7e7 e8d8 g5h3 d8c8 f6f7 f1f5 e7e8 c8b7 f7f8q f5f8 e8f8 b1b3 a3a2 b3c3 h3g5 c6d7 g5f7 c3b3 f7d6 b7c6 a2a7 c6b6 f8a8 b6c6 a8d8 c4c3 d8d7 b3b8 d7e7 c6b6 a7a3 b6c6 a3c3 c6b6 c3b3 b6c6 b3b8 g6g5 e7b7 g5g4 b8c8

User avatar
Eelco de Groot
Posts: 3951
Joined: Sun Mar 12, 2006 1:40 am
Location: Groningen

Re: No chance

Post by Eelco de Groot » Wed Sep 05, 2018 11:05 pm

Thanks very much Louis! A lot of nodes that. More than a million million. There must be a way to get it down without compromising search.
Debugging is twice as hard as writing the code in the first
place. Therefore, if you write the code as cleverly as possible, you
are, by definition, not smart enough to debug it.
-- Brian W. Kernighan

Spliffjiffer
Posts: 149
Joined: Thu Aug 02, 2012 5:48 pm
Location: Germany

Re: No chance

Post by Spliffjiffer » Thu Sep 06, 2018 12:16 pm

this thread has already 16 sites and we still have no conclusion about the correctness of LT's queensac...the positions inbetween (with mate in 30 or whatever) are very interesting and there are a lot of testpositions in there were engines are blind for very long time but the main-question remains imo...is the queensac winning or a hoax...in my last post i gave my idea which i think might proove that LT's assumption was correct even when his claims of mate in 8 were...lests say...funny :-)
to me it seems the most important questions remaing are:

1. is there no win after 9...Ne3!! ?
[D]1r3rk1/5p2/1qb1pPpR/3pP1N1/p1pP4/PpP1n3/1P2B2P/3RB2K w - - 1 10
draw? (i for myself was not able to find anything for white)

2.is black lost after 8. Bd2!? (instead of LT's 8.Ng5) ?
[D]1r3rk1/5p1p/1qb1pPp1/3pP3/p1pP4/PpP2N1R/1PnBB2P/3R3K b - - 2 8

a deep engine think at this point would be helpful...not wether its mate in 33 or 35 after 9..Nxa3 ??
Wahrheiten sind Illusionen von denen wir aber vergessen haben dass sie welche sind.

User avatar
George Tsavdaris
Posts: 1530
Joined: Thu Mar 09, 2006 11:35 am

Re: No chance

Post by George Tsavdaris » Thu Sep 06, 2018 1:28 pm

Spliffjiffer wrote:
Thu Sep 06, 2018 12:16 pm
this thread has already 16 sites and we still have no conclusion about the correctness of LT's queensac...the positions inbetween (with mate in 30 or whatever) are very interesting and there are a lot of testpositions in there were engines are blind for very long time but the main-question remains imo...is the queensac winning or a hoax...in my last post i gave my idea which i think might proove that LT's assumption was correct even when his claims of mate in 8 were...lests say...funny :-)
to me it seems the most important questions remaing are:

1. is there no win after 9...Ne3!! ?
[D]1r3rk1/5p2/1qb1pPpR/3pP1N1/p1pP4/PpP1n3/1P2B2P/3RB2K w - - 1 10
draw? (i for myself was not able to find anything for white)
No draw. 10. Nf3!! wins.
I have analyzed all variations and white wins no matter what.
2.is black lost after 8. Bd2!? (instead of LT's 8.Ng5) ?
[D]1r3rk1/5p1p/1qb1pPp1/3pP3/p1pP4/PpP2N1R/1PnBB2P/3R3K b - - 2 8

a deep engine think at this point would be helpful...not wether its mate in 33 or 35 after 9..Nxa3 ??
This question is not interesting to me since white wins with Ng5, so..... :D
After his son's birth they've asked him:
"Is it a boy or girl?"
YES! He replied.....

Spliffjiffer
Posts: 149
Joined: Thu Aug 02, 2012 5:48 pm
Location: Germany

Re: No chance

Post by Spliffjiffer » Thu Sep 06, 2018 1:41 pm

its unbelievable...i just installed MB's McCain-engine and it finds 10. Rh3 (instead of your Nf3!!?) with climbing score....so yes, it seems LT's 8.Ng5 is winning...im still amazed by this McCain :-)...and am still analysing...very impressive stuff

so the queensac seems correct and Lyudmil Tsvedkov was right....congratualtions...hurray :-)
thx George for the fast replay...nice coincident with new McCain and thx to MikeB for this release !
Wahrheiten sind Illusionen von denen wir aber vergessen haben dass sie welche sind.

Spliffjiffer
Posts: 149
Joined: Thu Aug 02, 2012 5:48 pm
Location: Germany

Re: No chance

Post by Spliffjiffer » Thu Sep 06, 2018 2:51 pm

BTW....@George...i remember i analysed 9....Ne3 10. Nf3...and i gave up after 10...Ng4!?....can u give a winning line maybe...im still analysing 10. Rh3 !?...will report later...is it possible to give a line after 10...Ng4 for u, thx in advance :-)
Wahrheiten sind Illusionen von denen wir aber vergessen haben dass sie welche sind.

Spliffjiffer
Posts: 149
Joined: Thu Aug 02, 2012 5:48 pm
Location: Germany

Re: No chance

Post by Spliffjiffer » Thu Sep 06, 2018 6:25 pm

McCain showed me the way :-)
i can refute now the last chance for black to hold the draw after:
[D]1r3rk1/5p2/1qb1pPpp/3pP1N1/p1pP4/PpP4R/1Pn1B2P/3RB2K w - - 0 9
9. Rxh6 Ne3 (this was the last move to hope for a draw) 10. Rh3! Nf5 (10. .. Nxd1
11. Bxd1 Rfd8 12. Bd2 Rd7 13. Kg2 Qc7 14. Kf2 +-) 11. Bf2+-
(for example: 11...Rfc8 12. Rf1 Qd8 13.Bg4 Nd6 14. Be3 Ne8 15. Rff3 Rc7 16. Rh4 Nxf6
17. exf6 Ra7 18. Nxe6 Qxf6 19. Rxf6 Re8 20. Bf4 Bd7 21. Be5 Rxe6 22. Bxe6 Bxe6
23. Rxe6 fxe6 24. Rh8+ Kf7 25. Rh7+ Ke8 26. Rxa7 Kd8 27. Kg2 Kc8 28. Kg3 Kd8 29. Kg4 Kc8
30. Kg5 Kd8 31. Kf6 Kc8 32. Kxe6 g5 33. Bc7 g4 34. Bg3 Kd8 35. Ra8# )

according to my analysis there is no possibility left for black to hold the draw after 8...h6...if this can be confirmed then congratulation to Lyudmil...he found probably another beautyful queensac that is absolutely sound !!?

only one thing left that is confusing me: how on earth white wins after 9...Ne3 10.Nf3 (George says its winning in all lines)...10...Ng4 !?
thx for the position + lines Lyudmil...i had big fun with it :-)
Wahrheiten sind Illusionen von denen wir aber vergessen haben dass sie welche sind.

Spliffjiffer
Posts: 149
Joined: Thu Aug 02, 2012 5:48 pm
Location: Germany

Re: No chance

Post by Spliffjiffer » Thu Sep 06, 2018 11:08 pm

little update:
after some further insight it seems to me that 10.Rh3! is not the only winning move after 9...Ne3 !?
10.Rh7 for example looks winning as well...eg coming back with the rook just later...what a misery...i just was not tenacious enough, forgive me friends :-) but we have clearence now at least ;-)

but guys, tell me the truth about this 9...Ne3 10.Nf3 Ng4 line...i simply dont get it !!
Wahrheiten sind Illusionen von denen wir aber vergessen haben dass sie welche sind.

Spliffjiffer
Posts: 149
Joined: Thu Aug 02, 2012 5:48 pm
Location: Germany

Re: No chance

Post by Spliffjiffer » Mon Sep 17, 2018 6:14 pm

im still trying to make LT's 8.Ng5 work (i think i could show that 8.Bd2!? is winning instead) and im not successful so far :-(...after 8.Ng5 Nxa3 !? we have this:
[D]1r3rk1/5p1p/1qb1pPp1/3pP1N1/p1pP4/npP4R/1P2B2P/3RB2K w - - 0 9
can somebody help me here please, i cannot make white win at this point
Wahrheiten sind Illusionen von denen wir aber vergessen haben dass sie welche sind.

User avatar
MikeB
Posts: 2683
Joined: Thu Mar 09, 2006 5:34 am
Location: Pen Argyl, Pennsylvania

Re: No chance

Post by MikeB » Tue Sep 18, 2018 3:50 am

Spliffjiffer wrote:
Mon Sep 17, 2018 6:14 pm
im still trying to make LT's 8.Ng5 work (i think i could show that 8.Bd2!? is winning instead) and im not successful so far :-(...after 8.Ng5 Nxa3 !? we have this:
[D]1r3rk1/5p1p/1qb1pPp1/3pP1N1/p1pP4/npP4R/1P2B2P/3RB2K w - - 0 9
can somebody help me here please, i cannot make white win at this point
I believe the only way that White might win is to play Bf2 :

and if Black plays h6, that is a sure win for white:

Code: Select all

dep	score	nodes	time	(not shown:  tbhits	knps	seldep)
 31	  #14 	27.9M  	0:00.76	Rg1 Qa5 Rgg3 Qxc3 Rxc3 Nb5 Rcg3 Nxd4 Bxd4 c3 Rxh6 c2 Be3 d4+ Kg1 Bf3 Nxf3 c1=Q+ Bxc1 Rbc8 Rgh3 Rxc1+ Bf1 Rxf1+ Kxf1 a3 Rh8# 
The only moves that could possibly not lose for black are Nb1, Nb5 and h5.

So we'll do a "deep think" and see what happens by tomorrow night, by then we would have 1.8 trillion moves searched, hopefully , it doesn't crash.
multipv is set to 3

Code: Select all

dep	score	nodes	time	(not shown:  tbhits	knps	seldep)
 34	  0.00 	6.27G	4:09.70	Nb1 Rxh7 Nd2 Rg7+ Kh8 Rh7+ 
 34	  0.00 	6.22G	4:07.82	Nb5 Rxh7 Nxc3 Rg7+ Kh8 Rh7+ 
 34	  0.00 	6.18G	4:06.21	h5 Bxh5 Qc7 Rg1 Nb5 Bd1 Nd6 exd6 Qc8 Rh7 e5 Bg4 Bd7 Rg7+ Kh8 Rh7+ 
 34	+2.92 	6.57G	4:21.93	Qb4 cxb4 c3 Rxc3 Bb5 bxa3 Bxe2 Rd2 Bc4 Rh3 h5 Rc3 Rb7 Rb2 Rc8 Be3 Kf8 Kg1 Rd7 Kf2 Ke8 Ke1 Kd8 Kd2 Rc6 Kc1 Kc7 Rg2 Kb7 Kb2 Kb8 Rd2 Rdc7 Rd1 Kb7 Rdc1 Ka6 Rg1 Kb5 Rb1 
 33	  0.00 	5.22G	3:27.32	Nb1 Rxh7 Nd2 Rg7+ Kh8 Rh7+ 
 33	  0.00 	5.22G	3:27.31	Nb5 Rxh7 Nxc3 Rg7+ Kh8 Rh7+ 
 33	  0.00 	5.16G	3:24.89	h5 Bxh5 Qc7 Rg1 Nb5 Bd1 Nd6 exd6 Qc8 Rh7 e5 Bg4 Bd7 Rg7+ Kh8 Rh7+ 
 33	+2.53 	5.41G	3:35.14	Qa5 Rg1 Qxc3 Rxc3 h6 Nxe6 Nb5 Rcg3 a3 bxa3 Bd7 Nxf8 Kxf8 Rb1 Nxa3 Rb2 Bf5 Be1 Nb5 Bc3 Ke8 Kg2 Kd7 Re3 Ke6 Bd1 Na3 Kf2 h5 h4 Bc2 Be2 Bf5 Rg3 Nb5 Bd1 Rc8 Bxb3 cxb3 Rxb3 
 33	+2.88 	5.44G	3:36.48	Qb4 cxb4 c3 Rxc3 Bb5 bxa3 Bxe2 Rd2 Bc4 Rh3 h5 Rc3 Rb7 Rb2 Rc8 Be3 Kf8 Kg1 Rd7 Kf2 Ke8 Ke1 Kd8 Kd2 Rc6 Kc1 Kc7 Rg2 Ba6 Rc5 Rxc5+ bxc5 Kc6 Kb2 Rc7 Kc3 Bf1 Rg1 Bb5 Rd1 Bc4 Kb2 Bb5 
 33	+12.80 	5.89G	3:54.76	Qc7 Rxh7 Nc2 Rg7+ Kh8 Rg1 Qxe5 dxe5 d4+ Bf3 Bxf3+ Nxf3 Rg8 Rxf7 Rgf8 Ra7 Ra8 Bxd4 Nxd4 Rxa8 Rxa8 Nxd4 a3 bxa3 Rb8 Nxe6 Kg8 Ng5 b2 Rb1 Rb5 e6 Rxg5 Rxb2 Rf5 Rb8+ Kh7 e7 Re5 f7 Rxe7 
 33	+13.05 	6.15G	4:05.00	Nc2 Rxh7 Qc7 Rg7+ Kh8 Rg1 Qxe5 dxe5 d4+ Bf3 Bxf3+ Nxf3 Rg8 Rxf7 Rgf8 Ra7 Ra8 Bxd4 Rxa7 Bxa7 a3 Bc5 Ra8 bxa3 Rc8 Bd6 Ne3 Ng5 Nf5 f7 b2 f8=Q+ Rxf8 Bxf8 Kg8 Bc5 Nh6 Nxe6 Ng4 Bd4 Nh6 Rb1 Nf5 Rxb2 Kf7 Ng5+ Ke7 Bc5+ Kd7 Rb6 
 32	  0.00 	3.05G	1:59.96	Nb1 Rxh7 Nd2 Rg7+ Kh8 Rh7+ 
 32	  0.00 	3.04G	1:59.43	Nb5 Rxh7 Nxc3 Rg7+ Kh8 Rh7+ 
 32	  0.00 	3.04G	1:59.26	h5 Bxh5 Qc7 Rg1 Nb5 Bd1 Nd6 exd6 Qc8 Rh7 e5 Bg4 Bd7 Rg7+ Kh8 Rh7+ 
 32	+2.56 	3.10G	2:01.96	Qa5 Rg1 Qxc3 Rxc3 h6 Nxe6 Nb5 Rcg3 a3 Nxf8 Kxf8 bxa3 Bd7 Rb1 Nxa3 Rb2 Bf5 Be1 Nb5 Bc3 Ke8 Kg2 Kd7 Re3 Ke6 Bd1 Na3 Kf2 Ra8 Rg3 Ra7 Ke1 Nb5 Bxb3 cxb3 
 32	+2.84 	3.28G	2:09.57	Qb4 cxb4 c3 Rxc3 Bb5 bxa3 Bxe2 Rd2 Bc4 Rh3 h5 Rc3 Rb7 Rb2 Rc8 Be3 Kf8 Kg1 Rd7 Kf2 Ke8 Ke1 Kd8 Kd2 Rc6 Kc1 Rdc7 Rb1 Kc8 Kb2 Kb7 Rg1 Kb8 Rgc1 
 32	+12.02 	4.26G	2:48.11	Nc2 Rxh7 Qc7 Rg7+ Kh8 Rg1 Qxe5 dxe5 d4+ Bf3 Bxf3+ Nxf3 Rg8 Rxf7 Rgf8 Ra7 Ra8 Bxd4 Rxa7 Bxa7 a3 Bc5 Rc8 Bxa3 Nxa3 Nd4 Rb8 Nxe6 Nb5 f7 Kh7 Rg4 g5 Rxg5 Na7 Rg7+ Kh8 Rg6 Kh7 f8=Q Rxf8 Nxf8+ Kh8 Rg1 

Post Reply