Test position

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

Moderator: Ras

Ron Murawski
Posts: 397
Joined: Sun Oct 29, 2006 4:38 am
Location: Schenectady, NY

Test position

Post by Ron Murawski »

[d]r1b1r1k1/2p1q1bp/1p1ppnp1/p1nP1pB1/2P1P3/2N2PPP/PPQNB1K1/3R1R2 b - - 0 18

18...Nxd5!

Perhaps this is easier for humans than engines? But I'm sure *some* engine(s) will solve it quickly! :)

[Event "Edited game"]
[Site "EINSTEIN"]
[Date "2007.07.16"]
[Round "-"]
[White "-"]
[Black "-"]
[Result "*"]

1. c4 Nf6 2. d4 g6 3. Nc3 Bg7 4. Nf3 O-O 5. e4 d6 6. Be2 Nc6 7. d5 Nb8 8.
h3 Na6 9. Be3 Nc5 10. Qc2 a5 11. O-O Nfd7 12. Nd2 e6 13. Rad1 Re8 14. f3 b6
15. g3 f5 16. Kh2 Nf6 17. Kg2 Qe7 18. Bg5 Nxd5
*
Stan Arts

Re: Test position

Post by Stan Arts »

My engine seems to like it from depth 8 on. But not quickly because this position seems to give it a bit of a tree explosion going from depth 7 to 8. (and before that some nervous gitters about the knightfork on c7) Urgh.

Code: Select all

 Legal moves: 39   Static score: -9
Ply Score  Time     Nodes      Best move and expected line
------------------------------------------------------------
2   1      0        587        Bb7 Kh2 exd5
2   39     0        828        Nxd5 Nxd5 Qxg5
3   39     1        2731       Nxd5 Nxd5 Qxg5 Nxc7
4   39     3        6608       Nxd5 Nxd5 Qxg5 Nxc7
5   15     11       25780      Nxd5 < - > Nxd5 Qxg5 Nxc7 Kf7
5   -60    11       26365      Nxd5 < - > Nxd5 Bc3
5   -130   15       34288      Nxd5 Nxd5 Qxg5 Nxc7 Ba6 Nxa8
5   -129   17       36303      Bb7 < + > Kh2 Re8-d8
5   -54    18       38594      Bb7 < + > Kh2 Re8-d8
5   14     31       69810      Bb7 Kh2 exd5 cxd5 fxe4 Bxf6
6   5      108      265718     Bb7 Kh2 c6 Nb3 exd5 Nxc5 bxc5
6   6      114      279665     Nxd5 < + > Nxd5 Qxg5 Kh1
6   28     118      289355     Nxd5 Nxd5 Qxg5 Nxc7 f4 Kh1
7   4      162      384258     Nxd5 < - > Nxd5 Qxg5 Nxc7 f4 Kh1 Kf7
7   -71    164      387719     Nxd5 < - > Nxd5 exd5 Bxe7 Rxe7
7   -138   179      419084     Nxd5 Nxd5 Qxg5 Nxc7 Ba6 f4 Qe7 Nxa8
7   -137   181      424332     Bb7 < + > Kh2 c6 Nb3 exd5 Nxc5 bxc5
7   -62    186      431189     Bb7 < + > Kh2 c6 Nb3 exd5 Nxc5 bxc5
7   14     415      990806     Bb7 Kg1 exd5 cxd5 Qe5 Kh2 fxe4 Bxf6
8   13     1489     3604489    Bb7 Rf1-e1 c6 Kh2 cxd5 cxd5 h6 Bxf6 Qxf6
8   14     1515     3664382    Nxd5 < + > Nxd5 Qxg5 Nxc7 f4 a3 Bxb2 Nxe8 Qxg3+
8   48     1543     3736583    Nxd5 Nxd5 Qxg5 Nxc7 f4 g4 Qh4 Nxe8 Qg3+ Kh1 Qxh3+ Kg1 Bd4+ Rf2 Qg3+ Kh1
9   26     1690     4041198    Nxd5 Nxd5 Qxg5 Nxc7 f4 Rh1 Qxg3+ Kf1 Bb7 Nxa8 Rxa8
10  48     3347     7919444    Nxd5 Nxd5 Qxg5 Nxc7 f4 g4 Qh4 Nxe8 Qg3+ Kh1 Qxh3+ Kg1 Bd4+ Rf2 Qg3+ Kh1
11  48     4483     10440919   Nxd5 Nxd5 Qxg5 Nxc7 f4 g4 Qh4 Nxe8 Qg3+ Kh1 Qxh3+ Kg1 Bd4+ Rf2 Qg3+ Kh1
12  72     7203     17499819   Nxd5 < + > Nxd5 Qxg5 Nxc7 f4 g4 Qh4 Nxe6 Bxe6 Rf1-e1
12  121    24709    58476037   Nxd5 cxd5 Qxg5 f4 Qe7 Bb5 Rf8 Bc6 Ba6 Bxa8 Bd3 Qc1 Bxf1+ Rxf1
13  124    52242    120476290  Nxd5 cxd5 Qxg5 Bb5 Rf8 f4 Qe7 Bc6 Rb8 Kh2 Ba6 Rf1-e1 fxe4 Nd2xe4
14  116    90681    219557274  Nxd5 cxd5 Qxg5 Bb5 Rf8 exf5 Qxf5 Qxf5 Rxf5 Nd2-e4 exd5 Nxd5 Rf7 Nxc5 bxc5
Total nodes:291449911 n/sec:243700  (q-nodes:39% max depth:34)
See you at the wcrcc tournament !
Stan


Ron Murawski wrote:[d]r1b1r1k1/2p1q1bp/1p1ppnp1/p1nP1pB1/2P1P3/2N2PPP/PPQNB1K1/3R1R2 b - - 0 18

18...Nxd5!

Perhaps this is easier for humans than engines? But I'm sure *some* engine(s) will solve it quickly! :)
bob
Posts: 20943
Joined: Mon Feb 27, 2006 7:30 pm
Location: Birmingham, AL

Re: Test position

Post by bob »

what is the punch line after Nxd5 Nxd5???
JVMerlino
Posts: 1404
Joined: Wed Mar 08, 2006 10:15 pm
Location: San Francisco, California

Re: Test position

Post by JVMerlino »

CM9_R1 needs just over two minutes to come back to Nxd5 on a P4-2.4:

Code: Select all

Time	Depth	Score	Positions	Moves
0:00	1/4	-0.74	10683		18...Nxd5 19.Nxd5 Qxg5 20.Nxc7
0:00	1/5	-0.58	22361		18...Nxd5 19.Nxd5 Qxg5 20.Nxc7
					f4 21.Nxe8 Qxg3+ 22.Kh1 Qxh3+ 23.Kg1
					Bd4+ 24.Rf2 Bxf2+ 25.Kxf2
0:00	1/6	-0.58	38574		18...Nxd5 19.Nxd5 Qxg5 20.Nxc7
					f4 21.Nxe8 Qxg3+ 22.Kh1 Qxh3+ 23.Kg1
					Bd4+ 24.Rf2 Bxf2+ 25.Kxf2
0:01	1/7	0.00	147038		18...Nxd5 19.Nxd5 Qxg5 20.Nxc7
					f4 21.g4 Qh4 22.Nxe8 Qg3+ 23.Kh1
					Qxh3+ 24.Kg1 Qg3+ 25.Kh1 Qh3+
0:05	1/7	-0.22	778285		18...a4 19.h4 Bb7 20.b4 axb3ep
					21.axb3 Qf7 22.Rfe1 fxe4 23.fxe4
					exd5 24.cxd5
0:07	1/7	-0.36	1025120		18...Ba6 19.f4 fxe4 20.Ndxe4 Ncxe4
					21.Nxe4 exd5
0:12	1/8	-0.49	1774594		18...Ba6 19.Nb3 Nxd5 20.cxd5 Qxg5
					21.f4 Qe7 22.Nxc5 Bxe2 23.Qxe2 bxc5
0:14	1/8	-0.54	1963331		18...Ng4 19.Bf4 e5 20.fxg4 exf4
					21.Rxf4 Bh6 22.gxf5 Bxf4 23.gxf4
					gxf5
0:21	1/8	-0.58	2920370		18...Bb7 19.Bh4 h6 20.f4 exd5 21.Nxd5
					Qf7 22.Nxf6+ Bxf6 23.Bxf6 Qxf6
0:33	1/9	-0.40	4747528		18...Bb7 19.Rfe1 Ng4 20.fxg4 Qxg5
					21.Nf3 Qh6 22.exf5 exf5 23.gxf5
1:12	1/10	-0.39	10188711	18...Bb7 19.Rfe1 h6 20.Be3 fxe4
					21.Bxc5 dxc5 22.Ndxe4 Nxe4 23.fxe4
					exd5 24.cxd5 Bxc3 25.Qxc3 Qxe4+
					26.Bf3
1:44	1/10	-0.51	14881577	18...Ng4 19.hxg4 Qxg5 20.f4 Qf6
					21.dxe6 Bxe6 22.gxf5 gxf5 23.Bh5
2:04	1/10	-0.92	18012257	18...Nxd5 19.cxd5 Qxg5 20.Bb5 Ba6
					21.h4 Qe7 22.Bxa6 Rxa6 23.Rc1 Qd7
2:20	1/11	-0.92	20597200	18...Nxd5 19.cxd5 Qxg5 20.Bb5 Ba6
					21.h4 Qe7 22.Bxa6 Rxa6 23.Rc1 Raa8
3:32	1/12	-0.67	32140420	18...Nxd5 19.Nxd5 Qxg5 20.Nxc7
					f4 21.g4 Qh4 22.Nb3 Qg3+ 23.Kh1
					Qxh3+ 24.Kg1 Nxb3 25.axb3 Qg3+
					26.Kh1 Bb7 27.Nxa8 Rxa8 28.Rxd6
jm
User avatar
Mike S.
Posts: 1480
Joined: Thu Mar 09, 2006 5:33 am

Re: Test position

Post by Mike S. »

bob wrote:what is the punch line after Nxd5 Nxd5???
It looks like 18...Nxd5 is good for a draw only, and probably a silent move like Bb7 may be better. But it is somewhat tricky, because after 18... Nxd5 19.Nxd5 Qxg5 20.Nxc7 f4, and again after 21.g4 Qh4, White does not have time to capture one of the rooks because of mate. But if he plays 22.Nb3 which defends against Bg7-d4, I think Black's best choice is to go for the perpetual with Qg3+ etc. Although I am not 100% sure if deeper analysis couldn't find an improvement for Black.

[Event "?""]
[Site "?"]
[Date "2007.07.17"]
[Round "?"]
[White "?"]
[Black "?"]
[Result "*"]
[SetUp "1"]
[FEN "r1b1r1k1/2p1q1bp/1p1ppnp1/p1nP1pB1/2P1P3/2N2PPP/PPQNB1K1/3R1R2 b - - 0 18"]
[PlyCount "14"]

18... Nxd5 19. Nxd5 Qxg5 20. Nxc7 f4 21. g4 (21. Nxa8 Qxg3+ 22. Kh1 Bd4 {
(no defense against Qxh3#)}) (21. Nxe8 Qxg3+ 22. Kh1 Qxh3+ {is a long mate.})
21... Qh4 22. Nb3 (22. Rh1 Qg3+ 23. Kf1 Bd4 {followed by Qf2#}) (22. Nxa8 Qg3+
23. Kh1 Bd4 {(again no defense against Qxh3#)}) (22. Nxe8 Qg3+ 23. Kh1 Bd4 24.
Nf6+ Kh8 {(same)}) 22... Qg3+ 23. Kh1 Qxh3+ 24. Kg1 Qg3+ 25. Kh1 *
Regards, Mike
Ron Murawski
Posts: 397
Joined: Sun Oct 29, 2006 4:38 am
Location: Schenectady, NY

Re: Test position

Post by Ron Murawski »

Mike S. wrote:
bob wrote:what is the punch line after Nxd5 Nxd5???
It looks like 18...Nxd5 is good for a draw only, and probably a silent move like Bb7 may be better. But it is somewhat tricky, because after 18... Nxd5 19.Nxd5 Qxg5 20.Nxc7 f4, and again after 21.g4 Qh4, White does not have time to capture one of the rooks because of mate. But if he plays 22.Nb3 which defends against Bg7-d4, I think Black's best choice is to go for the perpetual with Qg3+ etc. Although I am not 100% sure if deeper analysis couldn't find an improvement for Black.

[Event "?""]
[Site "?"]
[Date "2007.07.17"]
[Round "?"]
[White "?"]
[Black "?"]
[Result "*"]
[SetUp "1"]
[FEN "r1b1r1k1/2p1q1bp/1p1ppnp1/p1nP1pB1/2P1P3/2N2PPP/PPQNB1K1/3R1R2 b - - 0 18"]
[PlyCount "14"]

18... Nxd5 19. Nxd5 Qxg5 20. Nxc7 f4 21. g4 (21. Nxa8 Qxg3+ 22. Kh1 Bd4 {
(no defense against Qxh3#)}) (21. Nxe8 Qxg3+ 22. Kh1 Qxh3+ {is a long mate.})
21... Qh4 22. Nb3 (22. Rh1 Qg3+ 23. Kf1 Bd4 {followed by Qf2#}) (22. Nxa8 Qg3+
23. Kh1 Bd4 {(again no defense against Qxh3#)}) (22. Nxe8 Qg3+ 23. Kh1 Bd4 24.
Nf6+ Kh8 {(same)}) 22... Qg3+ 23. Kh1 Qxh3+ 24. Kg1 Qg3+ 25. Kh1 *
Mike's analysis looks flawless. I missed the white knight fork at c7; I was too enamoured of the black threat of simultaneously forking king, queen, and 2 rooks. I really should run these positions overnight before posting! The position is busted! :(

But it's quite possible that Nxd5 is the best move...

Ron
bob
Posts: 20943
Joined: Mon Feb 27, 2006 7:30 pm
Location: Birmingham, AL

Re: Test position

Post by bob »

Mike S. wrote:
bob wrote:what is the punch line after Nxd5 Nxd5???
It looks like 18...Nxd5 is good for a draw only, and probably a silent move like Bb7 may be better. But it is somewhat tricky, because after 18... Nxd5 19.Nxd5 Qxg5 20.Nxc7 f4, and again after 21.g4 Qh4, White does not have time to capture one of the rooks because of mate. But if he plays 22.Nb3 which defends against Bg7-d4, I think Black's best choice is to go for the perpetual with Qg3+ etc. Although I am not 100% sure if deeper analysis couldn't find an improvement for Black.

[Event "?""]
[Site "?"]
[Date "2007.07.17"]
[Round "?"]
[White "?"]
[Black "?"]
[Result "*"]
[SetUp "1"]
[FEN "r1b1r1k1/2p1q1bp/1p1ppnp1/p1nP1pB1/2P1P3/2N2PPP/PPQNB1K1/3R1R2 b - - 0 18"]
[PlyCount "14"]

18... Nxd5 19. Nxd5 Qxg5 20. Nxc7 f4 21. g4 (21. Nxa8 Qxg3+ 22. Kh1 Bd4 {
(no defense against Qxh3#)}) (21. Nxe8 Qxg3+ 22. Kh1 Qxh3+ {is a long mate.})
21... Qh4 22. Nb3 (22. Rh1 Qg3+ 23. Kf1 Bd4 {followed by Qf2#}) (22. Nxa8 Qg3+
23. Kh1 Bd4 {(again no defense against Qxh3#)}) (22. Nxe8 Qg3+ 23. Kh1 Bd4 24.
Nf6+ Kh8 {(same)}) 22... Qg3+ 23. Kh1 Qxh3+ 24. Kg1 Qg3+ 25. Kh1 *
that was my point. I thought this was a move to win the game by black, but it is a drawing move with that white response. Hence my question...
Dirt
Posts: 2851
Joined: Wed Mar 08, 2006 10:01 pm
Location: Irvine, CA, USA

Re: Test position

Post by Dirt »

Mike S. wrote:It looks like 18...Nxd5 is good for a draw only, and probably a silent move like Bb7 may be better. But it is somewhat tricky, because after 18... Nxd5 19.Nxd5 Qxg5 20.Nxc7 f4, and again after 21.g4 Qh4, White does not have time to capture one of the rooks because of mate. But if he plays 22.Nb3 which defends against Bg7-d4, I think Black's best choice is to go for the perpetual with Qg3+ etc. Although I am not 100% sure if deeper analysis couldn't find an improvement for Black.

[Event "?""]
[Site "?"]
[Date "2007.07.17"]
[Round "?"]
[White "?"]
[Black "?"]
[Result "*"]
[SetUp "1"]
[FEN "r1b1r1k1/2p1q1bp/1p1ppnp1/p1nP1pB1/2P1P3/2N2PPP/PPQNB1K1/3R1R2 b - - 0 18"]
[PlyCount "14"]

18... Nxd5 19. Nxd5 Qxg5 20. Nxc7 f4 21. g4 (21. Nxa8 Qxg3+ 22. Kh1 Bd4 {
(no defense against Qxh3#)}) (21. Nxe8 Qxg3+ 22. Kh1 Qxh3+ {is a long mate.})
21... Qh4 22. Nb3 (22. Rh1 Qg3+ 23. Kf1 Bd4 {followed by Qf2#}) (22. Nxa8 Qg3+
23. Kh1 Bd4 {(again no defense against Qxh3#)}) (22. Nxe8 Qg3+ 23. Kh1 Bd4 24.
Nf6+ Kh8 {(same)}) 22... Qg3+ 23. Kh1 Qxh3+ 24. Kg1 Qg3+ 25. Kh1 *
After following your main line to 23. Kh1, Spike 1.2 thinks black can do better than forcing a draw and wants 23.. Bb7. I don't see any really convincing continuation there, but I'm not sure there isn't one.
bob
Posts: 20943
Joined: Mon Feb 27, 2006 7:30 pm
Location: Birmingham, AL

Re: Test position

Post by bob »

Dirt wrote:
Mike S. wrote:It looks like 18...Nxd5 is good for a draw only, and probably a silent move like Bb7 may be better. But it is somewhat tricky, because after 18... Nxd5 19.Nxd5 Qxg5 20.Nxc7 f4, and again after 21.g4 Qh4, White does not have time to capture one of the rooks because of mate. But if he plays 22.Nb3 which defends against Bg7-d4, I think Black's best choice is to go for the perpetual with Qg3+ etc. Although I am not 100% sure if deeper analysis couldn't find an improvement for Black.

[Event "?""]
[Site "?"]
[Date "2007.07.17"]
[Round "?"]
[White "?"]
[Black "?"]
[Result "*"]
[SetUp "1"]
[FEN "r1b1r1k1/2p1q1bp/1p1ppnp1/p1nP1pB1/2P1P3/2N2PPP/PPQNB1K1/3R1R2 b - - 0 18"]
[PlyCount "14"]

18... Nxd5 19. Nxd5 Qxg5 20. Nxc7 f4 21. g4 (21. Nxa8 Qxg3+ 22. Kh1 Bd4 {
(no defense against Qxh3#)}) (21. Nxe8 Qxg3+ 22. Kh1 Qxh3+ {is a long mate.})
21... Qh4 22. Nb3 (22. Rh1 Qg3+ 23. Kf1 Bd4 {followed by Qf2#}) (22. Nxa8 Qg3+
23. Kh1 Bd4 {(again no defense against Qxh3#)}) (22. Nxe8 Qg3+ 23. Kh1 Bd4 24.
Nf6+ Kh8 {(same)}) 22... Qg3+ 23. Kh1 Qxh3+ 24. Kg1 Qg3+ 25. Kh1 *
After following your main line to 23. Kh1, Spike 1.2 thinks black can do better than forcing a draw and wants 23.. Bb7. I don't see any really convincing continuation there, but I'm not sure there isn't one.
looks to me like you end up with the same draw there as well... If Bb7 then white takes on a8 and now black has to find a way to recover the material or else take the draw and run.
Dirt
Posts: 2851
Joined: Wed Mar 08, 2006 10:01 pm
Location: Irvine, CA, USA

Re: Test position

Post by Dirt »

bob wrote:looks to me like you end up with the same draw there as well... If Bb7 then white takes on a8 and now black has to find a way to recover the material or else take the draw and run.
Just taking Spike's preferred move after a few seconds leads to this:
[Event "Edited game"]
[Site "GREGSIMP"]
[Date "2007.07.17"]
[Round "-"]
[White "-"]
[Black "-"]
[Result "*"]
[FEN "r1b1r1k1/2p1q1bp/1p1ppnp1/p1nP1pB1/2P1P3/2N2PPP/PPQNB1K1/3R1R2 b - - 0 1"]
1... Nxd5 2. Nxd5 Qxg5 3. Nxc7 f4 4. g4 Qh4 5. Nb3 Qg3+ 6. Kh1 Bb7 7. Nxa8
Qxh3+ 8. Kg1 Qg3+ 9. Kh1 Bxa8 10. Rd2 Qh4+ 11. Kg1 h5 12. Bd1 hxg4 13. Rh2
Qg5 14. Nxc5 g3 15. Rd2 Qxc5+ 16. Kg2 Bc6
*
[d]4r1k1/6b1/1pbpp1p1/p1q5/2P1Pp2/5Pp1/PPQR2K1/3B1R2 w - - 0 17

Black has two pawns and the two bishops for the exchange, and Spike is quite happy. There are many variations possible, and this is unlikely to be best play. Even so, I'm not sure there's any real advantage here.