Test position, Queen sac, that appears to be winning!

[PauloSoare]

Eelco, the problem with this would be the depth to find a high score?
I've been analyzing the position and I think that after 1. axb6 1 ... c3 can
hinder the search, I am not sure.
The position is so beautiful that I would dare say that further analysis of
Martin Thorens with Houdini, white can win, just do not ask me how =)

[Eelco de Groot]

Eelco, the problem with this would be the depth to find a high score?
I've been analyzing the position and I think that after 1. axb6 1 ... c3 can
hinder the search, I am not sure.
The position is so beautiful that I would dare say that further analysis of
Martin Thorens with Houdini, white can win, just do not ask me how =)

Hi Paulo!

Yes, cool queensac and a good testposition too, for the computers I mean! I think it is mostly a question of depth, because in the line that Rainbow Serpent saw, White gets an endgame where Black still has a passed pawn but it can not ever promote, bishop is guarding the promotion square and Black does not have pieces anymore, other than the King And on the other side of the board, White has no passed pawn but two connected pawns which are much stronger than the lone black pawn on the h-file. But there is no clear evaluation rule for that pair of pawns, it depends on how much support/interference there is from other pieces too, so it has to be searched and that would require just some extra depth.

I have seen that move ...c3 in deeper lines too, further along the PV I mean, but there it dropped in score after a while. But if you see a good line please post! It does not have to be a Houdini line, because I think the endgame should not be too difficult that you need Houdini's extra endgame knowledge for it. Just its extra depth

I will try to make a new verson of Rainbow Serpent that might be able to look just a little bit deeper, then I will post again, it should be possible to get an evaluation a little bit better than +0.16 I think. Watch out Houdini

Tommy Cooper engine in the making regards,
Eelco

[Robert Flesher]

Eelco, the problem with this would be the depth to find a high score?
I've been analyzing the position and I think that after 1. axb6 1 ... c3 can
hinder the search, I am not sure.
The position is so beautiful that I would dare say that further analysis of
Martin Thorens with Houdini, white can win, just do not ask me how =)

I am glad you like it, this is why I posted it. People who actually play chess, like us, will like this!

[zullil]

Does your favourite engine see axb6!!

[d]r1r3k1/p1Q4p/1n2qpp1/P2b4/2p4P/5PB1/3N2PK/R2R4 w - - 0 28

Code: Select all

Stockfish-2.0.1 (8 threads):

info depth 32 seldepth 43 multipv 1 score cp 169 time 975578 nodes 8448482672 nps 8659976 pv a5b6 c8c7 b6c7 g8f7 d2e4 d5e4 f3e4 a8c8 d1d8 c4c3 a1d1 c3c2 d1d7 e6d7 d8d7 f7e8 d7d8 c8d8 c7d8q e8d8 g3f4 d8d7 h2g3 d7c6 g3f3 c6b5 f3e2 b5c4 e2d2 a7a6 d2c2 c4d4 c2d2 d4e4 f4c7 e4f5 c7d8 g6g5 h4h5 f5g4 d8f6 g4h5 f6c3 

(The search is still running.)

[tomgdrums]

Eelco, the problem with this would be the depth to find a high score?
I've been analyzing the position and I think that after 1. axb6 1 ... c3 can
hinder the search, I am not sure.
The position is so beautiful that I would dare say that further analysis of
Martin Thorens with Houdini, white can win, just do not ask me how =)

I am glad you like it, this is why I posted it. People who actually play chess, like us, will like this!

Oddly enough Naum 4.2 picked the move instantly but could not be convinced it was winning. I probably should have given it more time. I decided to play a game against Junior instead.

[Eelco de Groot]

Does your favourite engine see axb6!!

[d]r1r3k1/p1Q4p/1n2qpp1/P2b4/2p4P/5PB1/3N2PK/R2R4 w - - 0 28

Code: Select all

Stockfish-2.0.1 (8 threads):

info depth 32 seldepth 43 multipv 1 score cp 169 time 975578 nodes 8448482672 nps 8659976 pv a5b6 c8c7 b6c7 g8f7 d2e4 d5e4 f3e4 a8c8 d1d8 c4c3 a1d1 c3c2 d1d7 e6d7 d8d7 f7e8 d7d8 c8d8 c7d8q e8d8 g3f4 d8d7 h2g3 d7c6 g3f3 c6b5 f3e2 b5c4 e2d2 a7a6 d2c2 c4d4 c2d2 d4e4 f4c7 e4f5 c7d8 g6g5 h4h5 f5g4 d8f6 g4h5 f6c3 

(The search is still running.)

I can't really compete with that 8 threads i7 from Louis but I'm pleased to announce that here he is, back on stage, the fish with the fez; Tommy Cooper 0.5

Paulo, it seems you were right, look at the variations of the new engine, suddenly it plays your 2... c3 several times!


[D]r1r3k1/p1Q4p/1n2qpp1/P2b4/2p4P/5PB1/3N2PK/R2R4 w - -

Engine: Tommy Cooper 0.5 (Athlon 2009 MHz, 256 MB)
heavily based on Stockfish 2.0.1 and Rainbow Serpent
by Tord Romstad, Marco Costalba, Joona Kiiski

1/01 0:00 -1.17 1.Re1 (1.285) 4

2/06 0:00 -1.17 1.Re1 Rxc7 2.Rxe6 Bxe6 3.Bxc7 Nd5 (14.337) 48

3/06 0:00 -1.81 1.Re1 Rxc7 2.Rxe6 Bxe6 3.Bxc7 Nd5 (34.698) 105

3/07 0:00 -1.17 1.Qd6 Qxd6 2.Bxd6 Nd7 3.Nb1 Be6
4.Re1 (57.335) 152

4/06 0:00 -1.25 1.Qd6 Qxd6 2.Bxd6 Nd7 3.Rab1 Ne5 (86.099) 204

5/07 0:00 -1.42 1.Qd6 Qxd6 2.Bxd6 Nd7 3.Nb1 Be6
4.Re1 (91.054) 207

6/10 0:00 -1.17++ 1.Qd6 Qxd6 2.Bxd6 Nd7 3.Nb1 Be6
4.Nc3 Nc5 5.Rab1 Nb3 (107.359) 228

6/11 0:00 -0.91++ 1.Qd6 Qxd6 2.Bxd6 Nd7 3.Nb1 Be6
4.Nc3 Nc5 5.g4 Nb3 6.Ra2 (244.109) 339

6/14 0:00 -1.01 1.Qd6 Qxd6 2.Bxd6 Nd7 3.Nb1 Be6
4.Re1 Kf7 5.Nc3 Rc6 6.Rad1 Nc5 7.Ne4 Nd3 (302.099) 371

7/18 0:01 -1.12 1.Qd6 Qxd6 2.Bxd6 Nd7 3.Nb1 Be6
4.Nc3 Nc5 5.Bxc5 Rxc5 6.Ne4 Rc7
7.Nxf6+ Kg7 8.Nd5 Bxd5 9.Rxd5 c3 (424.491) 411

8/12 0:01 -1.13 1.Qd6 Qxd6 2.Bxd6 Nd7 3.Nb1 Be6
4.Nc3 Nc5 5.Ne4 Nxe4 6.fxe4 c3 (460.318) 420

9/17 0:01 -1.05++ 1.Qd6 Qxd6 2.Bxd6 Nd7 3.Nb1 Be6
4.Nc3 Nc5 5.Kg3 Nb3 6.Ra2 Kf7 7.a6 Rc6
8.Bf4 Re8 9.Ne4 (514.487) 433

9/17 0:01 -1.17-- 1.Qd6 Qxd6 2.Bxd6 Nd7 3.Nb1 Be6
4.Nc3 Nc5 5.Kg3 Nb3 6.Ra2 Kf7 7.a6 Rc6
8.Bf4 Re8 9.Ne4 (614.247) 457

9/20 0:01 -1.01++ 1.Qd6 Qxd6 2.Bxd6 Nd7 3.Nb1 Be6
4.Nc3 Nc5 5.Kg3 Nb3 6.Ra2 Kf7 7.a6 Rc6
8.Bf4 Nc5 9.Rb1 Rd8 10.Rb7+ Rd7 (658.179) 462

9/15 0:01 -0.96 1.Qd6 Qxd6 2.Bxd6 Nd7 3.Nb1 Be6
4.Nc3 Nc5 5.Kg3 Nb3 6.Ra2 Kf7 7.Nb5 Rc6
8.Bb4 (763.520) 474

10/12 0:01 -1.13 1.Qd6 Qxd6 2.Bxd6 Nd7 3.Nb1 Be6
4.Nc3 Nc5 5.Kg3 Kg7 6.Rd2 Bf7 (851.368) 486

11/15 0:02 -1.09 1.Qd6 Qxd6 2.Bxd6 Nd7 3.Nb1 Be6
4.Nc3 Nc5 5.g4 Nb3 6.Ra2 Kf7 7.Kg3 Rd8
8.Nb5 (1.432.247) 517

12/20 0:03 -1.18 1.Qd6 Qxd6 2.Bxd6 Nd7 3.Nb1 Be6
4.Nc3 Nc5 5.g4 Nb3 6.Ra2 Kf7 7.Kg3 Rd8
8.Nb5 a6 9.Nc7 Ra7 10.Re2 Rxc7 (1.606.440) 524

13/17 0:03 -1.17 1.Qd6 Qxd6 2.Bxd6 Nd7 3.Nb1 Be6
4.Nc3 Nc5 5.g4 Nb3 6.Ra4 Re8 7.Kg2 Rad8
8.Rb4 Nxa5 9.Ra4 (2.063.376) 536

14/20 0:04 -1.20 1.Qd6 Qxd6 2.Bxd6 Nd7 3.Nb1 Be6
4.Nc3 Nc5 5.Kg3 Kf7 6.Bxc5 Rxc5
7.Rab1 f5 8.Rb7+ Kf6 9.Kf2 Rxa5
10.Rxh7 Ra3 (2.497.282) 547

15/20 0:05 -1.24 1.Qd6 Qxd6 2.Bxd6 Nd7 3.Nb1 Be6
4.Nc3 Nc5 5.Kg3 Kf7 6.Ra2 Nd3 7.Re2 Rd8
8.Nb5 Rd7 9.Nc7 Rc8 10.Rxe6 Rcxc7 (3.005.668) 554

16/17 0:06 -1.17 1.Qd6 Qxd6 2.Bxd6 Nd7 3.Nb1 Be6
4.Nc3 Nc5 5.Kg3 Kf7 6.a6 Nb3 7.Ra2 Rc6
8.Bf4 Re8 9.Be3 (3.674.725) 562

17/19 0:08 -1.17 1.Qd6 Qxd6 2.Bxd6 Nd7 3.Nb1 Be6
4.Nc3 Nc5 5.Kg3 Kf7 6.a6 Nb3 7.Ra2 Rc6
8.Bf4 Nc5 9.Be3 Re8 10.Ra5 (4.643.957) 570

18/11 0:20 -1.09++ 1.Qd6 Qxd6 2.Bxd6 Nd7 3.Rac1 Kg7
4.Re1 c3 5.Re7+ Bf7 6.Ne4 (12.106.877) 587

18/23 0:25 -1.21-- 1.Qd6 Qxd6 2.Bxd6 Nd7 3.Nb1 Be6
4.Nc3 Nc5 5.Kg3 Kf7 6.a6 Nb3 7.Ra2 Rc6
8.Bf4 Nc5 9.Rb1 Rd8 10.Rb7+ Rd7
11.Rxd7+ Bxd7 12.Be3 (14.813.782) 591

18/07 0:35 -1.01++ 1.axb6 Rxc7 2.bxc7 f5 3.Nxc4 Bxc4
4.Rd8+ (20.652.536) 583

18/05 0:37 -0.84++ 1.axb6 Rxc7 2.bxc7 Qb6 3.Rac1 (21.907.171) 585

18/21 0:43 -0.52++ 1.axb6 Rxc7 2.bxc7 Kg7 3.Nxc4 Bxc4
4.Rd8 Rc8 5.Rad1 Ba6 6.R1d7+ Kh6
7.Bf4+ g5 8.hxg5+ Kg6 9.Rd6 Qe2
10.Rxf6+ Kg7 11.Rd7+ (25.903.891) 591

18/31 1:03 -0.24 1.axb6 Rxc7 2.bxc7 c3 3.Ne4 Bxe4
4.Rd8+ Kg7 5.fxe4 Rc8 6.Rc1 g5
7.hxg5 fxg5 8.Rxc3 Qh6+ 9.Kg1 Qb6+
10.Bf2 Qb7 11.Rxc8 Qxc8 12.Bxa7 Kf7
13.Kh2 Ke6 14.Kg3 (38.261.312) 600

19/31 1:12 -0.22 1.axb6 Rxc7 2.bxc7 c3 3.Ne4 Bxe4
4.Rd8+ Kg7 5.fxe4 Rc8 6.Rc1 g5
7.hxg5 fxg5 8.Rxc3 Qh6+ 9.Kg1 Qb6+
10.Bf2 Qb7 11.Rxc8 Qxc8 12.Bxa7 Kf7
13.Kh2 Ke6 14.Kg3 (43.444.337) 601

20/29 1:23 -0.20 1.axb6 Rxc7 2.bxc7 c3 3.Ne4 Bxe4
4.Rd8+ Kg7 5.fxe4 Rc8 6.Rc1 g5
7.hxg5 fxg5 8.Rxc3 Qh6+ 9.Kg1 Qb6+
10.Bf2 Qb7 11.Rxc8 Qxc8 12.Bxa7 Kf7
13.Kh2 Ke6 14.Kg3 (50.230.297) 604

21/33 1:37 -0.12++ 1.axb6 Rxc7 2.bxc7 c3 3.Ne4 Bxe4
4.Rd8+ Kg7 5.fxe4 Rc8 6.Rc1 g5
7.hxg5 fxg5 8.Rxc3 Qh6+ 9.Kg1 Qb6+
10.Bf2 Qb7 11.Rxc8 Qxc8 12.Bxa7 Kf7
13.Kf2 Ke6 14.Kg3 (58.862.936) 606

21/34 1:52 -0.11 1.axb6 Rxc7 2.bxc7 c3 3.Ne4 Bxe4
4.Rd8+ Kg7 5.fxe4 Rc8 6.Rc1 g5
7.hxg5 fxg5 8.Rxc3 Qh6+ 9.Kg1 Qb6+
10.Bf2 Qb7 11.Rxc8 Qxc8 12.Bxa7 Kf7
13.Kf2 Ke6 14.Bd4 (68.291.812) 604

22/30 2:19 -0.11 1.axb6 Rxc7 2.bxc7 c3 3.Ne4 Bxe4
4.Rd8+ Kg7 5.fxe4 Rc8 6.Rc1 g5
7.hxg5 fxg5 8.Rxc3 Qh6+ 9.Kg1 Qb6+
10.Bf2 Qb7 11.Rxc8 Qxc8 12.Bxa7 Kf7
13.Kf2 Ke6 14.Kg3 (84.556.414) 605

23/21 2:55 -0.03++ 1.axb6 Rxc7 2.bxc7 c3 3.Ne4 Bxe4
4.Rd8+ Kg7 5.fxe4 Rc8 6.Rc1 g5
7.hxg5 fxg5 8.Rxc8 Qxc8 9.Rxc3 Kf7
10.Rd3 Qe6 11.Rd8 (106.318.978) 605

23/13 3:27 0.00 1.axb6 Rxc7 2.bxc7 Kf7 3.Nxc4 Bxc4
4.Rd8 Rc8 5.Rad1 Ke7 6.Bd6+ Kf7
7.Bg3 (125.769.530) 606

24/13 4:39 0.00 1.axb6 Rxc7 2.bxc7 Kf7 3.Nxc4 Bxc4
4.Rd8 Rc8 5.Rad1 Ke7 6.Bd6+ Kf7
7.Bg3 (168.815.646) 604

25/13 6:37 0.00 1.axb6 Rxc7 2.bxc7 Kf7 3.Nxc4 Bxc4
4.Rd8 Rc8 5.Rad1 Ke7 6.Bd6+ Kf7
7.Bg3 (239.009.797) 601

26/13 9:38 0.00 1.axb6 Rxc7 2.bxc7 Kf7 3.Nxc4 Bxc4
4.Rd8 Rc8 5.Rad1 Ke7 6.Bd6+ Kf7
7.Bg3 (346.620.826) 598

27/13 15:35 0.00 1.axb6 Rxc7 2.bxc7 Kf7 3.Nxc4 Bxc4
4.Rd8 Rc8 5.Rad1 Ke7 6.Bd6+ Kf7
7.Bg3 (557.955.054) 596

28/13 37:55 0.00 1.axb6 Rxc7 2.bxc7 Kf7 3.Nxc4 Bxc4
4.Rd8 Rc8 5.Rad1 Ke7 6.Bd6+ Kf7
7.Bg3 (1.349.851.091) 593

29/04 77:43 +0.08++ 1.axb6 Rxc7 2.bxc7 Kf7 (2.782.777.194) 596

29/03 104:49 +0.16++ 1.axb6 Rxc7 2.bxc7 (3.767.958.243) 599

29/05 169:17 +0.32++ 1.axb6 Rxc7 2.bxc7 Kf7 3.Ne4 (5.993.017.936) 589

29/04 204:13 +0.64++ 1.axb6 c3 2.Ne4 Rxc7 (7.252.181.924) 591

29/06 216:29 +0.52-- 1.axb6 Rxc7 2.bxc7 Kf7 3.Ne4 Bxe4 (7.708.554.196) 593

29/01 270:07 +1.24++ 1.axb6 (9.651.441.679) 595

29/06 306:03 +1.08-- 1.axb6 Rxc7 2.bxc7 c3 3.Ne4 Bxe4 (11.009.212.441) 599

29/32 408:21 +1.66 1.axb6 Rxc7 2.bxc7 c3 3.Ne4 Bxe4
4.Rd8+ Kf7 5.fxe4 Rc8 6.Rad1 c2
7.R1d7+ Qxd7 8.Rxd7+ Ke8 9.Rd8+ Rxd8
10.cxd8Q+ Kxd8 11.Bf4 Kd7 12.Kg1 Kc6
13.Kf2 Kc5 14.Ke2 (14.721.314.566) 600

Overall Tommy Cooper seems a slight improvement in the amount of nodes needed to reach completion of an iteration but it is difficult to be sure, because the variations are also different and Tommy Cooper could use 100% of the processor, no parallel Stockfish were running so the times can't really be compared. I'm just looking at the total nodenumbers.

Eelco

[kgburcham]

[D] r1r3k1/p1Q4p/1n2qpp1/P2b4/2p4P/5PB1/3N2PK/R2R4 w - -

HIARCS 13.2 MP (2048 MB)
by Mark John Uniacke

21/49 1:31 +0.70 1.axb6 Rxc7 2.bxc7 Rc8 3.Ne4 Bxe4 4.Rd8+ Kf7 5.fxe4 c3 6.Rad1 Ke7 7.Rxc8 (567.453.964) 6231
22/49 2:26 +0.93 1.axb6 Rxc7 2.bxc7 g5 3.hxg5 Rc8 4.Ne4 Bxe4 5.fxe4 (960.969.121) 6554
23/58 4:21 +1.15 1.axb6 Rxc7 2.bxc7 Rc8 3.Ne4 Bxe4 4.Rd8+ Kf7 5.fxe4 c3 6.Rad1 Ke7 7.Rxc8 (1.789.845.747) 6835
24/58 7:30 +1.27 1.axb6 Rxc7 2.bxc7 Kf7 3.Ne4 Bxe4 4.fxe4 g5 5.hxg5 (3.142.504.384) 6980
25/58 12:20 +1.30 1.axb6 Rxc7 2.bxc7 Rc8 3.Ne4 Bxe4 4.Rd8+ (5.268.183.538) 7112
26/58 22:52 +1.51 1.axb6 Rxc7 2.bxc7 Kf7 3.Ne4 Bxe4 4.fxe4 Ke7 5.Rab1 Rc8 6.Rb8 c3 7.Rd8 Rxc7 8.Re8+ (9.881.229.998) 7200
27/62 39:37 +1.53 1.axb6 Rxc7 2.bxc7 Rc8 3.Ne4 Bxe4 4.Rd8+ (17.196.499.416) 7233


1. axb6 Rxc7 2. bxc7 Rc8 3. Ne4
[D] 2r3k1/p1P4p/4qpp1/3b4/2p1N2P/5PB1/6PK/R2R4 b - -

Houdini 1.5ab-16 x64 (4096 MB)
by Robert Houdart

26/67 1:10 +1.29-- 3...Bxe4 4.Rd8+ (2.300.536.133) 32695
26/71 2:10 +1.49-- 3...Bxe4 4.Rd8+ (4.431.015.264) 33910
26/71 3:28 +1.60 3...Bxe4 4.Rd8+ Kf7 5.fxe4 c3 6.Rc1 g5 7.hxg5 fxg5 8.Rxc3 Qh6+ 9.Kg1 Qb6+ 10.Bf2 Qb7 11.Rxc8 Qxc8 12.Bxa7 Ke7 13.Be3 Kf6 14.Bd4+ Ke7 15.Rh3 Kd6 16.Rxh7 Qb7 (7.222.053.116) 34572
27/76 5:11 +1.70-- 3...Bxe4 4.Rd8+ (10.957.981.378) 35182

1... Bxe4 2. Rd8+ Kf7 3. fxe4 c3
[D] 2rR4/p1P2k1p/4qpp1/8/4P2P/2p3B1/6PK/R7 w - -

Houdini 1.5ab-16 x64 (4096 MB)
by Robert Houdart

26/66 1:03 +2.87++ 4.Rc1 (2.186.099.731) 34264
26/72 1:47 +3.13 4.Rc1 a5 5.Rxc3 a4 6.Rcd3 Ke7 7.Rxc8 Qxc8 8.Rd8 Qxd8 9.cxd8N Kxd8 10.Bd6 Kd7 11.Ba3 Ke6 12.Kg3 h6 13.Kf3 f5 14.Kf4 fxe4 15.Kxe4 Kf6 16.Bb2+ Kf7 17.Kd5 (3.766.101.195) 35043
27/72 2:20 +3.13 4.Rc1 a5 5.Rxc3 a4 6.Rcd3 Ke7 7.Rxc8 Qxc8 8.Rd8 Qxd8 9.cxd8N Kxd8 10.Bd6 Kd7 11.Ba3 Ke6 12.Kg3 h6 13.Kf3 f5 14.Kf4 fxe4 15.Kxe4 Kf6 16.Bb2+ Ke6 17.g3 (4.956.528.914) 35271
28/72 2:44 +3.13 4.Rc1 a5 5.Rxc3 a4 6.Rcd3 Ke7 7.Rxc8 Qxc8 8.Rd8 Qxd8 9.cxd8N Kxd8 10.Bd6 Kd7 11.Ba3 Ke6 12.Kg3 h6 13.Kf3 f5 14.Kf4 fxe4 15.Kxe4 Kf6 16.Bb2+ Ke6 17.g3 (5.816.525.387) 35280
29/78 6:28 +3.28++ 4.Rc1 (14.397.769.262) 37106
29/78 7:54 +3.28 4.Rc1 a5 5.Rxc3 a4 6.Rcd3 Ke7 7.Rxc8 Qxc8 8.Rd8 Qxd8 9.cxd8N Kxd8 10.Bd6 Kd7 11.Ba3 h5 12.Kg3 g5 13.Kf3 Ke8 14.Ke3 Kf7 15.g3 Kg7 16.Bb2 Kg6 17.Kd4 (17.650.853.492) 37179

[zullil]

Stockfish-2.0.1 (8 threads):

Code: Select all

info depth 38 seldepth 44 multipv 1 score cp 315 time 16021512 nodes 147949944374 nps 9234455 pv a5b6 c4c3 d2e4 d5e4 d1d8 c8d8 a1a7 d8d7 c7d7 e6d7 a7d7 c3c2 d7c7 e4b7 c7c2 a8a6 g3f2 a6a8 c2c7 a8b8 f2g3 b7a6 c7c6 b8b7 c6f6 b7d7 h4h5 g6h5 f6f5 a6b7 f5a5 b7c6 a5h5 d7d5 h5h6 c6b7 g3f4 d5c5 f4e3 c5e5 e3d4 e5g5 h6h4 g5f5 

[PauloSoare]

Eelco, the problem with this would be the depth to find a high score?
I've been analyzing the position and I think that after 1. axb6 1 ... c3 can
hinder the search, I am not sure.
The position is so beautiful that I would dare say that further analysis of
Martin Thorens with Houdini, white can win, just do not ask me how =)

I am glad you like it, this is why I posted it. People who actually play chess, like us, will like this!

I do not play chess OTB long ago, Robert, maybe 6 years ago.
I'm currently playing in engine room, a lot of work to prepare
best openings. OTB I am not a good player.

[PauloSoare]

I told one .... c3, Eelco. I was wrong.