Bug in Top5 Engines?

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

Moderators: Harvey Williamson, bob, hgm

Forum rules
This textbox is used to restore diagrams posted with the [d] tag before the upgrade.
Post Reply
Spliffjiffer
Posts: 248
Joined: Thu Aug 02, 2012 5:48 pm
Location: Germany

Bug in Top5 Engines?

Post by Spliffjiffer » Wed Oct 24, 2012 5:11 pm

V.Anand vs M.Sadler:

1. e4 c5 2. Nf3 d6 3. d4 cxd4 4. Nxd4 Nf6 5. Nc3 a6 6. Be3 e5 7. Nb3 Be6 8.f3 Be7 9. Qd2 h5 10. O-O-O Nbd7 11. g3 Rc8 12. Nd5 Bxd5 13. exd5 Nb6 14.Bxb6 Qxb6 15. Bh3 Rc7 16. Kb1 h4 17. Qe1 Kf8 18. f4 exf4 19. gxf4 Rh5 20.Bg2 Rc4 21. Rf1 Rf5 22. Qxh4 Nh5 23. Qh3 g6 24. Rfe1 Bf6 25. Be4 Rxf4 26.Qd3 Rc7 27. Bf3 Rh4 28. Qd2 Kg7 29. c3 Be5 30. Rxe5 1/2:1/2

[D]8/1pr2pk1/pq1p2p1/3PR2n/7r/1NP2B2/PP1Q3P/1K1R4 b - - 0 30

Sadler threw his hands up in horror, he had thrown away his winning position

Instead of 29. ...Be5? black should have played 29. ...Rcc4 the players agreed in the analysis afterwards...(also 29. ...a5 looks very promising)

So far so good...but not for the top 5 engines...


Houdini 1.5, Critter 1.6a, Komodo 3, Stockfish 2.3.1 and Rybka 3 dynamic (well...if R3 dy. is not a top 5 engine you can probably use R4) have a common bug here:
After 29. ...Be5 they not only have big problems to find 30. Rxe5, they also have the same bug when you go backwards from the position which was reached after:

30.Qg5 Bf6, 31.Qd2 Rcc4, 32.Rf1...

[D]8/1p3pk1/pq1p1bp1/3P3n/2r4r/1NP2B2/PP1Q3P/1K1R1R2 b - - 6 32


Note: Use the "playing mode" and NOT the "analysis mode" if you use Houdini 1.5

Now...choose any timecontrol, let's say 40/20 or even more, and let the engines compute on the 32. move for white and wait until the engine makes its move.
You will see that the engine is quite unconfotable with the position.
After any move was played, go back to the 31. move for white and let the engine compute again...wait for the move...
Go back to the 30. move for white and let it compute again.
Now the engine should know that 30.Qg5 is very unpleasant because of the quite forced Bf6, Qd2 - maneuver and the consequences you were computing for in move 32
...but it don't...
It will think that the position is drawish after 30.Qg5 !!!

This works also in the "analysis mode" of Critter, Komodo,Stockfish and Rybka 3 dynamic !!!
Only houdini 1.5 is different in "analysis mode".

Maybe this is not a bug but anyway...the engines would probably loose the game after 30.Qg5??.

Do you have the same bug?

I used Windows XP prof. (32 bit),E6600 Core 2 Duo @ 2,4 GHz



Regards
Wahrheiten sind Illusionen von denen wir aber vergessen haben dass sie welche sind.

User avatar
Ajedrecista
Posts: 1484
Joined: Wed Jul 13, 2011 7:04 pm
Location: Madrid, Spain.
Contact:

Re: Bug in Top-5 engines?

Post by Ajedrecista » Wed Oct 24, 2012 6:08 pm

Hello Peer:
Spliffjiffer wrote:V.Anand vs M.Sadler:

1. e4 c5 2. Nf3 d6 3. d4 cxd4 4. Nxd4 Nf6 5. Nc3 a6 6. Be3 e5 7. Nb3 Be6 8.f3 Be7 9. Qd2 h5 10. O-O-O Nbd7 11. g3 Rc8 12. Nd5 Bxd5 13. exd5 Nb6 14.Bxb6 Qxb6 15. Bh3 Rc7 16. Kb1 h4 17. Qe1 Kf8 18. f4 exf4 19. gxf4 Rh5 20.Bg2 Rc4 21. Rf1 Rf5 22. Qxh4 Nh5 23. Qh3 g6 24. Rfe1 Bf6 25. Be4 Rxf4 26.Qd3 Rc7 27. Bf3 Rh4 28. Qd2 Kg7 29. c3 Be5 30. Rxe5 1/2:1/2

[D]8/1pr2pk1/pq1p2p1/3PR2n/7r/1NP2B2/PP1Q3P/1K1R4 b - - 0 30

Sadler threw his hands up in horror, he had thrown away his winning position

Instead of 29. ...Be5? black should have played 29. ...Rcc4 the players agreed in the analysis afterwards...(also 29. ...a5 looks very promising)

So far so good...but not for the top 5 engines...


Houdini 1.5, Critter 1.6a, Komodo 3, Stockfish 2.3.1 and Rybka 3 dynamic (well...if R3 dy. is not a top 5 engine you can probably use R4) have a common bug here:
After 29. ...Be5 they not only have big problems to find 30. Rxe5, they also have the same bug when you go backwards from the position which was reached after:

30.Qg5 Bf6, 31.Qd2 Rcc4, 32.Rf1...

[D]8/1p3pk1/pq1p1bp1/3P3n/2r4r/1NP2B2/PP1Q3P/1K1R1R2 b - - 6 32


Note: Use the "playing mode" and NOT the "analysis mode" if you use Houdini 1.5

Now...choose any timecontrol, let's say 40/20 or even more, and let the engines compute on the 32. move for white and wait until the engine makes its move.
You will see that the engine is quite unconfotable with the position.
After any move was played, go back to the 31. move for white and let the engine compute again...wait for the move...
Go back to the 30. move for white and let it compute again.
Now the engine should know that 30.Qg5 is very unpleasant because of the quite forced Bf6, Qd2 - maneuver and the consequences you were computing for in move 32
...but it don't...
It will think that the position is drawish after 30.Qg5 !!!

This works also in the "analysis mode" of Critter, Komodo,Stockfish and Rybka 3 dynamic !!!
Only houdini 1.5 is different in "analysis mode".

Maybe this is not a bug but anyway...the engines would probably loose the game after 30.Qg5??.

Do you have the same bug?

I used Windows XP prof. (32 bit),E6600 Core 2 Duo @ 2,4 GHz



Regards
Sorry for this long post. I tried this position:

[d]8/1pr2pk1/pq1p2p1/3Pb2n/7r/1NP2B2/PP1Q3P/1K1RR3 w - - 0 30

Which I think it is when 29.- ..., Be5 was played (please correct me if I am wrong). I only tried Hannibal, Quazar, Spark, Naum and Spike: all of them find RxB in few seconds (I cleared hash between analysis and analysis).

Analysis by Hannibal 1.2 w32 (single core) in an Intel Pentium D930 (3 GHz), with 512 MB of hash, running under Arena 2.0.1 GUI:
FEN: 8/1pr2pk1/pq1p2p1/3Pb2n/7r/1NP2B2/PP1Q3P/1K1RR3 w - - 0 30

Hannibal1.2x32:
8 00:00 11.437 368.935 +0,11 Re1xe5 d6xe5 d5d6 Rc7d7 Qd2e1 Qb6d8 Qe1xe5+ Qd8f6
9 00:00 16.140 520.645 +0,10 Re1xe5 d6xe5 d5d6 Rc7d7 Qd2e1 e5e4 Qe1xh4 e4xf3 Rd1f1 Qb6xd6 Rf1xf3
10- 00:00 28.400 364.102 -0,06 Re1xe5 d6xe5 d5d6 Rc7d7 Qd2e1 e5e4 Qe1xh4 e4xf3 Qh4h3 Rd7xd6 Rd1d3
10 00:00 115.427 321.523 -0,14 Qd2g5 Be5f6 Qg5g2 Nh5f4 Qg2g3 Rc7c8 Bf3g4 Rc8c4 Re1e8 Rc4a4
11 00:00 210.422 328.784 -0,01 Re1xe5 d6xe5 d5d6 Rc7d7 Qd2e1 e5e4 Qe1xh4 e4xf3 Nb3d4 Rd7xd6 Rd1d3 f3f2 Qh4xf2 Kg7g8 Rd3d1 Nh5f6 Qf2f4
12 00:00 240.563 335.045 -0,01 Re1xe5 d6xe5 d5d6 Rc7d7 Qd2e1 e5e4 Qe1xh4 e4xf3 Nb3d4 Rd7xd6 Rd1d3 f3f2 Qh4xf2 Kg7g8 Rd3d1 Nh5f6 Qf2f4
13 00:01 425.031 362.654 -0,16 Re1xe5 d6xe5 d5d6 Rc7d7 Qd2e1 e5e4 Qe1xh4 e4xf3 Nb3d4 Rd7xd6 Rd1d3 f3f2 Qh4xf2 Rd6f6 Rd3f3 Qb6d6 Qf2d2 Nh5f4 Nd4f5+ g6xf5 Qd2xf4 Qd6xf4 Rf3xf4
14 00:02 775.567 384.896 -0,21 Re1xe5 d6xe5 d5d6 Rc7d7 Qd2e1 e5e4 Qe1xh4 e4xf3 Nb3d4 Rd7xd6 Rd1d3 f3f2 Rd3f3 Rd6f6 Rf3xf2 Rf6xf2 Qh4xf2 Qb6c7 Qf2h4 Qc7e5
15 00:03 1.250.638 400.204 -0,21 Re1xe5 d6xe5 d5d6 Rc7d7 Qd2e1 e5e4 Qe1xh4 e4xf3 Nb3d4 Rd7xd6 Rd1d3 f3f2 Rd3f3 Rd6f6 Rf3xf2 Rf6xf2 Qh4xf2 Qb6c7 Qf2g2 Nh5f6 Nd4f5+ Kg7f8
16- 00:05 2.348.748 403.010 -0,37 Re1xe5 d6xe5 d5d6 Rc7d7 Qd2e1 e5e4 Qe1xh4 e4xf3 Nb3d4 Rd7xd6 Rd1d3 Rd6f6 Nd4xf3 Qb6b5 Qh4d4 Nh5f4 Rd3d1 Qb5e2 Qd4e5 Qe2xe5 Nf3xe5 g6g5 Kb1c2 Rf6h6 Rd1e1
16 00:10 4.074.904 401.231 -0,40 Re1xe5 d6xe5 d5d6 Rc7d7 Qd2e1 e5e4 Qe1xh4 e4xf3 Nb3d4 Rd7xd6 Rd1f1 Rd6f6 Rf1f2 Qb6d6 Qh4e4 b7b6 Qe4b7 b6b5 a2a3 Rf6f4 Rf2xf3 Rf4xf3 Nd4xf3 Nh5f4
17 00:22 8.620.294 387.167 -0,34 Re1xe5 d6xe5 d5d6 Rc7d7 Qd2e1 e5e4 Qe1xh4 e4xf3 Qh4d4+ Qb6xd4 Nb3xd4 f3f2 Rd1f1 Rd7xd6 Rf1xf2 Nh5f6 Kb1c2 Rd6d8 Rf2e2 Nf6d5 Nd4f3 Kg7f6 Re2e4 Kf6f5 Re4e5+ Kf5g4
18 00:33 13.114.130 390.918 -0,40 Re1xe5 d6xe5 d5d6 Rc7d7 Qd2e1 e5e4 Qe1xh4 e4xf3 Qh4d4+ Qb6xd4 Nb3xd4 f3f2 Rd1f1 Rd7xd6 Rf1xf2 Nh5f6 Kb1c2 Rd6d8 Rf2e2 Nf6d5 Nd4f3 Kg7f6 c3c4 Nd5f4 Re2d2 Rd8xd2+ Kc2xd2
19 00:41 16.538.105 399.114 -0,40 Re1xe5 d6xe5 d5d6 Rc7d7 Qd2e1 e5e4 Qe1xh4 e4xf3 Qh4d4+ Qb6xd4 Nb3xd4 f3f2 Rd1f1 Rd7xd6 Rf1xf2 Nh5f6 Kb1c2 Rd6d8 Rf2e2 Nf6d5 Kc2b3 g6g5 Re2d2 Kg7f6 Nd4c2 a6a5 c3c4 a5a4+ Kb3a3
20 00:56 23.336.845 413.385 -0,35 Re1xe5 d6xe5 d5d6 Rc7d7 Qd2e1 e5e4 Qe1xh4 e4xf3 Qh4d4+ Qb6xd4 Nb3xd4 f3f2 Rd1f1 Rd7xd6 Rf1xf2 Nh5f6 Kb1c2 Nf6e4 Rf2g2 Rd6d8 Rg2g1 Rd8h8 h2h4 Ne4c5 Nd4f3 Rh8e8 Rg1e1 Re8xe1 Nf3xe1 Kg7f6 Ne1d3 Nc5xd3 Kc2xd3
21 01:33 40.133.772 427.810 -0,40 Re1xe5 d6xe5 d5d6 Rc7d7 Qd2e1 e5e4 Qe1xh4 e4xf3 Qh4d4+ Qb6xd4 Nb3xd4 f3f2 Rd1f1 Nh5f6 Rf1xf2 Nf6e4 Rf2g2 Ne4xd6 Kb1c2 Rd7e7 Rg2e2 Re7xe2+ Nd4xe2 Kg7f6 Kc2d3 Kf6e5 Ne2d4 f7f5 Nd4f3+ Ke5f4 Nf3d4 b7b6 Nd4e6+ Kf4e5
22 02:29 63.198.550 423.886 -0,40 Re1xe5 d6xe5 d5d6 Rc7d7 Qd2e1 e5e4 Qe1xh4 e4xf3 Qh4d4+ Qb6xd4 Nb3xd4 f3f2 Rd1f1 Nh5f6 Rf1xf2 Nf6e4 Rf2g2 Ne4xd6 Kb1c2 Rd7e7 Kc2d3 f7f5 c3c4 Kg7f6 c4c5 Nd6e4 Kd3c4 g6g5 b2b4 Re7d7 a2a3 Rd7d8 Rg2e2 g5g4
23 03:23 85.716.184 421.371 -0,40 Re1xe5 d6xe5 d5d6 Rc7d7 Qd2e1 e5e4 Qe1xh4 e4xf3 Qh4d4+ Qb6xd4 Nb3xd4 f3f2 Rd1f1 Nh5f6 Rf1xf2 Nf6e4 Rf2g2 Ne4xd6 Kb1c2 Rd7e7 Kc2d3 f7f5 c3c4 Kg7f6 c4c5 Nd6e4 Kd3c4 g6g5 b2b4 Re7d7 a2a3 Rd7d8 Rg2e2 g5g4
24 04:45 122.964.071 430.132 -0,37 Re1xe5 d6xe5 d5d6 Rc7d7 Qd2e1 e5e4 Qe1xh4 e4xf3 Qh4d4+ Qb6xd4 Nb3xd4 f3f2 Rd1f1 Nh5f6 Rf1xf2 Nf6e4 Rf2g2 Kg7h6 Kb1c2 f7f5 Rg2g1 Rd7xd6 Kc2b3 b7b6 Kb3c4 Rd6d7 a2a4 f5f4 Rg1d1 Ne4f6 Rd1f1 g6g5 h2h3 Rd7c7+ Kc4d3
I stopped the analysis after five minutes.

------------------------

Analysis by Quazar 0.4 w32 (single core) in an Intel Pentium D930 (3 GHz), with 512 MB of hash, running under Arena 2.0.1 GUI:
FEN: 8/1pr2pk1/pq1p2p1/3Pb2n/7r/1NP2B2/PP1Q3P/1K1RR3 w - - 0 30

Quazar_0.4_w32:
1 00:01 6 4 +0,08 Re1e2
2 00:01 44 35 -0,05 Re1e2 f7f5 Re2g2
3 00:01 184 148 -0,05 Re1e2 f7f5 Nb3d4 Be5xd4 c3xd4
4 00:01 398 318 -0,05 Re1e2 f7f5 Nb3d4 Be5xd4 c3xd4
5 00:01 1.116 892 -0,39 Re1e2 f7f5 Nb3d4 Be5xd4 c3xd4 Nh5f4
6 00:01 2.907 2.296 -0,17 Re1e2 f7f5 Qd2e1 Rh4c4 Re2xe5 d6xe5 d5d6 Rc7d7 Qe1xe5+ Nh5f6
7 00:01 6.012 4.578 -0,33 Re1xe5 d6xe5 d5d6 Rc7d7 Qd2e1 e5e4 Qe1xh4 e4xf3 Qh4g4 Rd7xd6 Qg4xf3 Rd6xd1+ Qf3xd1
8 00:01 14.679 10.207 -0,11 Re1xe5 d6xe5 d5d6 Rc7d7 Qd2e1 e5e4 Qe1xh4 e4xf3 Qh4g4 Rd7xd6 Nb3d4 Rd6xd4 Qg4xd4+ Qb6xd4 c3xd4 Nh5f6
9 00:01 18.726 12.747 -0,20 Re1xe5 d6xe5 d5d6 Rc7d7 Qd2e1 e5e4 Qe1xh4 e4xf3 Qh4g4 Rd7xd6 Rd1xd6 Qb6xd6 Qg4xf3 b7b5
10 00:01 30.292 18.641 -0,20 Re1xe5 d6xe5 d5d6 Rc7d7 Qd2e1 e5e4 Qe1xh4 e4xf3 Nb3d4 Rd7xd6 Rd1d3 Qb6d8 Qh4xd8 Rd6xd8 Rd3xf3 b7b5 b2b4 Rd8d5
11 00:01 47.347 26.115 -0,39 Re1xe5 d6xe5 d5d6 Rc7d7 Qd2e1 e5e4 Qe1xh4 e4xf3 Nb3d4 Rd7xd6 Rd1d3 Qb6d8 Qh4e1 f7f6 Qe1d2 Qd8d7 Nd4e6+ Qd7xe6 Rd3xd6 Qe6e4+ Kb1c1
12 00:02 67.517 32.983 -0,30 Re1xe5 d6xe5 d5d6 Rc7d7 Qd2e1 e5e4 Qe1xh4 e4xf3 Nb3d4 Rd7xd6 Rd1d3 Rd6f6 Qh4f2 Qb6a5 Rd3xf3 Qa5e5 Rf3xf6 Qe5e4+ Kb1c1 Nh5xf6
13- 00:02 107.242 40.377 -0,38 Re1xe5 d6xe5 d5d6 Rc7d7 Qd2e1 e5e4 Qe1xh4 e4xf3 Nb3d4 Rd7xd6 Rd1d3 Rd6f6 Nd4xf3 Qb6e6
13 00:02 112.495 41.618 -0,30 Re1xe5 d6xe5 d5d6 Rc7d7 Qd2e1 e5e4 Qe1xh4 e4xf3 Nb3d4 Rd7xd6 Rd1d3 Rd6f6 Qh4f2 Qb6a5 Rd3xf3 Qa5e5 Rf3xf6 Qe5e4+ Kb1c1 Nh5xf6
14- 00:03 140.627 46.152 -0,38 Re1xe5 d6xe5 d5d6 Rc7d7 Qd2e1 e5e4 Qe1xh4 e4xf3 Nb3d4 Rd7xd6 Rd1d3 Rd6f6 Qh4f2 Qb6a5 Rd3xf3 Qa5e5 Rf3xf6 Qe5e4+ Kb1c1 Nh5xf6
14 00:03 149.857 47.709 -0,25 Re1xe5 d6xe5 d5d6 Rc7d7 Qd2e1 e5e4 Qe1xh4 e4xf3 Nb3d4 Rd7xd6 Rd1d3 Rd6f6 Qh4f2 Qb6a5 Rd3xf3 Qa5e5 Rf3xf6 Nh5xf6 Kb1c2
15- 00:03 189.819 52.363 -0,33 Re1xe5 d6xe5 d5d6 Rc7d7 Qd2e1 e5e4 Qe1xh4 e4xf3 Nb3d4 Rd7xd6 Rd1d3 Rd6f6 Qh4f2 Qb6a5 Rd3xf3 Qa5e5 Rf3xf6 Nh5xf6 Kb1c2 Qe5e4+ Kc2b3 Nf6d5
15- 00:04 221.407 54.708 -0,41 Re1xe5 d6xe5 d5d6 Rc7d7 Qd2e1 e5e4 Qe1xh4 e4xf3 Nb3d4 Rd7xd6 Rd1d3 Rd6f6 Qh4f2 Qb6a5 Rd3xf3 Qa5e5 Rf3xf6 Nh5xf6 Kb1c2 Nf6e4
15 00:04 263.035 58.857 -0,42 Re1xe5 d6xe5 d5d6 Rc7d7 Qd2e1 e5e4 Qe1xh4 e4xf3 Qh4d4+ Qb6xd4 Nb3xd4 Rd7xd6 Nd4f5+ g6xf5 Rd1xd6 f3f2 Rd6d1 Nh5f6 Rd1f1 Nf6e4 Kb1c2 Kg7g6 Kc2d3 Kg6g5 a2a3 b7b5 Kd3e3 f7f6 Rf1xf2 Ne4xf2 Ke3xf2
16- 00:07 477.332 66.554 -0,51 Re1xe5 d6xe5 d5d6 Rc7d7 Qd2e1 e5e4 Qe1xh4 e4xf3 Qh4d4+ Qb6xd4 Nb3xd4 Rd7xd6 Nd4f5+ g6xf5 Rd1xd6 f3f2 Rd6d1 Nh5f6 Rd1f1 Nf6e4 Kb1c2 b7b5 Kc2d3 Kg7f6 a2a3 Kf6g5 Kd3e3 f7f6 Rf1xf2 Ne4xf2 Ke3xf2 Kg5h4
16 00:07 501.415 67.841 -0,45 Re1xe5 d6xe5 d5d6 Rc7d7 Qd2e1 e5e4 Qe1xh4 e4xf3 Qh4d4+ Qb6xd4 Nb3xd4 Rd7xd6 Nd4f5+ g6xf5 Rd1xd6 f3f2 Rd6d1 Nh5f6 Rd1f1 Nf6e4 Kb1c2 b7b5 Kc2d3 Kg7g6 b2b3 Kg6g5 c3c4 Kg5f4 a2a4 b5xa4 b3xa4 Ne4c5+ Kd3e2
17 00:10 777.635 72.976 -0,48 Re1xe5 d6xe5 d5d6 Rc7d7 Qd2e1 e5e4 Qe1xh4 e4xf3 Nb3d4 Rd7xd6 Qh4g5 Rd6f6 Qg5e5 Nh5f4 Rd1d2 f3f2 Rd2xf2 Nf4d3 Qe5xf6+ Qb6xf6 Rf2xf6 Kg7xf6 Kb1c2 Nd3f4 a2a3 Kf6e5 h2h4 Nf4g2
18+ 00:17 1.255.135 73.762 -0,40 Qd2g5 Qb6f2 Re1f1 Qf2xh2 Rf1h1 Be5f6 Rh1xh2 Bf6xg5 Rh2xh4 Bg5xh4 Rd1h1 Rc7c4 Nb3a5 Rc4f4 Na5xb7 Rf4xf3 Rh1xh4
18 00:18 1.377.171 73.942 -0,41 Qd2g5 Qb6f2 Re1f1 Qf2xh2 Rf1h1 Be5f6 Rh1xh2 Bf6xg5 Rh2xh4 Bg5xh4 Rd1h1 Bh4f2 Bf3xh5 g6xh5 Rh1h2 Bf2e3 Rh2xh5 Rc7e7 Kb1c2 Kg7g6 Rh5h3 f7f5 Nb3d4 f5f4 Nd4e6 Kg6f5 Kc2d3
19- 00:22 1.635.985 73.889 -0,50 Qd2g5 Qb6f2 Re1f1 Qf2xh2 Rf1h1 Be5f6 Rh1xh2 Bf6xg5 Rh2xh4 Bg5xh4 Rd1h1 Bh4g5 Bf3xh5 g6xh5 Rh1xh5 Kg7g6 Rh5h8 Rc7e7 Nb3d4 Re7e5 Kb1c2 Re5xd5 Kc2d3 Rd5e5
19 00:22 1.706.428 74.545 -0,46 Qd2g5 Qb6f2 Re1f1 Qf2xh2 Rf1h1 Be5f6 Rh1xh2 Bf6xg5 Rh2xh4 Bg5xh4 Rd1h1 Bh4g5 Bf3xh5 g6xh5 Rh1xh5 Kg7g6 Rh5h8 Rc7e7 Rh8b8 Kg6f5 Kb1c2 Kf5f4 a2a3 Re7e3 Nb3d4
20- 00:29 2.151.520 73.280 -0,55 Qd2g5 Qb6f2 Re1f1 Qf2xh2 Rf1h1 Be5f6 Rh1xh2 Bf6xg5 Rh2xh4 Bg5xh4 Rd1h1 Bh4g5 Bf3xh5 g6xh5 Rh1xh5 Kg7g6 Rh5h8 Rc7e7 Rh8b8 Kg6f5 Kb1c2 Bg5f4 Nb3d4+ Kf5e4 Nd4e2 Bf4e5
20 00:31 2.331.609 73.761 -0,47 Qd2g5 Qb6f2 Re1f1 Qf2xh2 Rf1h1 Be5f6 Rh1xh2 Bf6xg5 Rh2xh4 Bg5xh4 Rd1h1 Bh4g5 Bf3xh5 g6xh5 Rh1xh5 Kg7g6 Rh5h8 Rc7e7 Rh8b8 Kg6f5 Kb1c2 Bg5f4 Nb3d4+ Kf5e5 Nd4e2 f7f5 Ne2xf4 Ke5xf4
21- 00:38 2.883.141 74.372 -0,55 Qd2g5 Qb6f2 Re1f1 Qf2xh2 Rf1h1 Be5f6 Rh1xh2 Bf6xg5 Rh2xh4 Bg5xh4 Rd1h1 Bh4g5 Bf3xh5 g6xh5 Rh1xh5 Kg7g6 Rh5h8 Rc7e7 Rh8b8 Kg6f5 Kb1c2 Bg5f4 Nb3d4+ Kf5e5 Nd4e2 Bf4e3
21 00:40 3.040.688 75.253 -0,54 Qd2g5 Qb6f2 Re1f1 Qf2xh2 Rf1h1 Be5f6 Rh1xh2 Bf6xg5 Rh2xh4 Bg5xh4 Rd1h1 Bh4g5 Bf3xh5 g6xh5 Rh1xh5 Kg7g6 Rh5h8 Rc7e7 Rh8b8 Kg6f5 Kb1c2 Bg5f4 a2a4 Kf5e4 c3c4 f7f6 Rb8h8
22+ 00:49 3.793.943 76.985 -0,45 Re1xe5 d6xe5 d5d6 Rc7c4 d6d7
22+ 00:53 4.213.924 78.420 -0,37 Re1xe5 d6xe5 d5d6 Rc7c4 d6d7
22 01:05 5.206.558 79.375 -0,55 Re1xe5 d6xe5 d5d6 Rc7d7 Qd2e1 e5e4 Qe1xh4 e4xf3 Nb3d4 Rd7xd6 Qh4g5 Rd6f6 Qg5e5 Nh5f4 Nd4xf3 Qb6e6 Rd1d4 Qe6f5+ Qe5xf5 Rf6xf5 Rd4d2 Nf4d5 Nf3d4 Rf5h5 a2a3 Kg7f6 Kb1c2 Rh5e5 Kc2d3
23+ 01:29 7.212.470 80.347 -0,46 Re1xe5 d6xe5 d5d6 Rc7d7 Qd2e1 e5e4 Qe1xh4 e4xf3 Nb3d4 Rd7xd6 Qh4f2 Rd6f6 Rd1d3 Qb6a5 Rd3xf3 Qa5e5 a2a4 Rf6xf3 Qf2xf3 Qe5xh2 Qf3xb7 Nh5f4 Qb7e4
23 02:04 9.737.037 78.071 -0,53 Re1xe5 d6xe5 d5d6 Rc7d7 Qd2e1 e5e4 Qe1xh4 e4xf3 Nb3d4 Rd7xd6 Qh4f2 Rd6f6 Rd1d3 Qb6a5 Rd3xf3 Qa5e5 a2a4 Qe5e4+ Kb1c1 Rf6xf3 Qf2xf3 Nh5f6 h2h4 Qe4xf3 Nd4xf3 Nf6e4 Kc1d1 f7f5 Kd1e2 Kg7f6 Ke2e3 b7b5
24+ 02:43 13.326.110 81.653 -0,45 Re1xe5 d6xe5 d5d6 Rc7d7 Qd2e1 e5e4 Qe1xh4 e4xf3 Qh4d4+ Qb6xd4 Nb3xd4 f3f2 Rd1f1 f7f5 Nd4e6+ Kg7f7 Ne6c5 Rd7xd6 Rf1xf2 b7b5 Kb1c2 f5f4 Rf2d2 Rd6xd2+ Kc2xd2
24+ 02:46 13.761.574 82.528 -0,37 Re1xe5 d6xe5 d5d6 Rc7d7 Qd2e1 e5e4 Qe1xh4 e4xf3 Qh4d4+ Qb6xd4 Nb3xd4 f3f2 Rd1f1 f7f5 Nd4e6+ Kg7f7 Ne6c5 Rd7xd6 Rf1xf2 b7b5 Kb1c2 f5f4 Rf2d2 Rd6xd2+ Kc2xd2
24- 04:16 20.371.246 79.294 -0,61 Re1xe5 d6xe5 d5d6 Rc7d7 Qd2e1 e5e4 Qe1xh4 e4xf3 Qh4d4+ Qb6xd4 Nb3xd4 f3f2 Rd1f1 f7f5 Nd4e6+ Kg7f7 Ne6c5 Rd7xd6 Rf1xf2 b7b5 Kb1c2 Nh5f6 Rf2d2 Rd6c6 b2b4 g6g5 Rd2d4 Kf7g6 h2h3 f5f4 Kc2d3 Kg6f5 Kd3e2 a6a5
24 04:25 21.017.863 79.093 -0,55 Re1xe5 d6xe5 d5d6 Rc7d7 Qd2e1 e5e4 Qe1xh4 e4xf3 Qh4d4+ Qb6xd4 Nb3xd4 f3f2 Rd1f1 f7f5 Nd4e6+ Kg7f7 Ne6c5 Rd7xd6 Rf1xf2 b7b5 Kb1c2 Nh5f6 Rf2d2 Rd6c6 b2b4 g6g5 Rd2d4 Nf6g4 h2h3 Ng4e5 Rd4d5 Kf7f6 Rd5d8 Ne5c4 Rd8d7 Kf6e5 Rd7e7+ Ke5d5 Kc2d3 f5f4
I stopped the analysis after five minutes.

------------------------

Analysis by Spark 1.0 w32 (using two cores) in an Intel Pentium D930 (3 GHz), with 512 MB of hash, running under Arena 2.0.1 GUI:
FEN: 8/1pr2pk1/pq1p2p1/3Pb2n/7r/1NP2B2/PP1Q3P/1K1RR3 w - - 0 30

Spark-1.0-win32-mp:
1 00:00 7 0 -3,53 a2a4 Qb6xb3
1 00:00 15 0 -3,33 a2a3 Qb6xb3
1 00:00 20 0 +0,18 h2h3 Rh4xh3
1 00:00 64 0 +0,24 Qd2g5 Rh4xh2
1 00:00 96 0 +0,44 Bf3xh5 Rh4xh5
1 00:00 111 0 +0,70 Re1h1
1 00:00 113 0 +0,98 Re1e2
2/1 00:00 191 0 +0,44 Re1e2 Nh5f4
3/2 00:00 467 0 +0,69 Re1e2 Nh5f4 Re2e4
3/3 00:00 865 0 +0,85 Qd2g5 Qb6f2 Re1xe5 Qf2xf3
4/3- 00:00 2.416 0 +0,53 Qd2g5 Qb6f2 Re1xe5 Qf2xf3
4/4 00:00 4.374 0 +0,36 Qd2g5 Be5f6 Qg5g2 Nh5f4
4/4 00:00 5.089 0 +0,59 Re1e2 Nh5f4 Re2e4 Kg7g8
5/4 00:00 6.441 0 +0,52 Re1e2 Nh5f4 Re2e4 f7f5 Re4b4
6/8 00:00 15.090 0 +0,54 Re1e2 Nh5f4 Re2e4 f7f5 Re4b4 Qb6a7
7/10 00:00 31.732 0 +0,51 Re1e2 Nh5f6 Rd1e1 Rh4h3 Bf3h1 Kg7g8 Re1f1
8/11 00:00 64.274 589.669 +0,43 Re1e2 Nh5f4 Re2e4 f7f5 Re4b4 Qb6a7 Nb3d4 Be5xd4 Rb4xd4
8/12+ 00:00 84.660 677.280 +0,76 Re1xe5 d6xe5
8/12 00:00 91.786 655.614 +0,48 Re1xe5 d6xe5 d5d6 Rc7d7 Qd2e1 e5e4 Qe1xh4 e4xf3 Rd1f1 Qb6xd6 Rf1xf3
9/12- 00:00 137.728 736.513 +0,23 Re1xe5 d6xe5 d5d6 Rc7d7 Qd2e1 e5e4 Qe1xh4 e4xf3 Rd1f1 Qb6xd6 Rf1xf3
9/13 00:00 185.624 793.264 +0,30 Re1e2 Nh5f6 Rd1e1 Be5f4 Qd2c2 Rh4h3 Bf3h1 Rh3e3 Re2xe3 Bf4xe3
10/17 00:00 343.531 846.135 +0,30 Re1e2 Nh5f6 Rd1e1 Be5f4 Qd2c2 Rh4h3 Bf3h1 Rh3e3 Re2xe3 Bf4xe3
11/17 00:00 540.795 962.268 +0,39 Re1e2 Nh5f6 Rd1e1 Be5f4 Qd2c2 Rh4h3 Re1f1 Rh3h4 Bf3h1 a6a5 Re2f2
12/20- 00:01 1.042.345 1.011.003 +0,14 Re1e2 Nh5f6 Rd1e1 Be5f4 Qd2c2 Rh4h3 Re1f1 Rh3h4 Bf3h1 a6a5 Re2f2
12/20 00:01 2.117.069 1.067.071 -0,06 Re1e2 Nh5f6 Qd2g5 Rh4h3 Bf3g4 Rh3xh2 Re2xh2 Be5xh2 Rd1h1 Qb6f2 Nb3d4 Nf6xg4 Qg5xg4
13/27 00:04 4.751.223 1.101.860 -0,17 Re1e2 Nh5f6 Kb1a1 a6a5 Qd2e1 Rh4xh2 Re2xh2 Be5xh2 Qe1d2 Bh2g1 Nb3xa5 Bg1e3 Qd2e2 Rc7e7
13/28 00:05 6.021.332 1.094.787 -0,06 Re1xe5 d6xe5 d5d6 Rc7d7 Qd2e1 e5e4 Qe1xh4 e4xf3 Qh4h3 Rd7xd6 Rd1xd6 Qb6xd6 Qh3xf3 Qd6xh2 Qf3xb7
14/28+ 00:06 7.307.822 1.121.691 +0,19 Re1xe5 d6xe5
14/28+ 00:06 7.537.051 1.132.369 +0,44 Re1xe5 d6xe5
14/28 00:06 7.952.177 1.138.627 +0,61 Re1xe5 d6xe5 d5d6 Rc7d7 Qd2e1 e5e4 Qe1xh4 e4xf3 Qh4g4 Rd7xd6 Nb3d4 Nh5f6 Qg4xf3 Nf6d5 Nd4b3
15/28- 00:08 9.668.683 1.169.834 +0,36 Re1xe5 d6xe5 d5d6 Rc7d7 Qd2e1 e5e4 Qe1xh4 e4xf3 Qh4g4 Rd7xd6 Nb3d4 Nh5f6 Qg4xf3 Nf6d5 Nd4b3
15/30- 00:11 13.543.899 1.200.593 +0,11 Re1xe5 d6xe5 d5d6 Rc7d7 Qd2e1 e5e4 Qe1xh4 e4xf3 Qh4g4 Rd7xd6 Nb3d4 Nh5f6 Qg4xf3 Nf6d5 Nd4b3
15/34 00:14 17.744.547 1.242.528 +0,05 Re1xe5 d6xe5 d5d6 Rc7d7 Qd2e1 e5e4 Qe1xh4 e4xf3 Qh4g4 Rd7xd6 Nb3d4 Rd6f6 Nd4xf3 Qb6f2 Nf3e5 Qf2xh2
16/34 00:17 22.387.631 1.294.307 +0,11 Re1xe5 d6xe5 d5d6 Rc7d7 Qd2e1 e5e4 Qe1xh4 e4xf3 Qh4g4 Rd7xd6 Nb3d4 Rd6f6 Nd4xf3 Qb6e6 Qg4xe6 Rf6xe6 Rd1d7
17/36 00:30 37.955.604 1.249.567 +0,11 Re1xe5 d6xe5 d5d6 Rc7d7 Qd2e1 e5e4 Qe1xh4 e4xf3 Nb3d4 Rd7xd6 Rd1d3 Qb6c7 Rd3xf3 Rd6f6 Nd4f5+ Rf6xf5 Rf3xf5 g6xf5 Qh4g5+ Kg7f8 Qg5xh5
18/40 00:43 54.597.117 1.243.500 +0,24 Re1xe5 d6xe5 d5d6 Rc7d7 Qd2e1 e5e4 Qe1xh4 e4xf3 Nb3d4 Rd7xd6 Rd1d3 Qb6c7 a2a3 b7b5 Qh4f2 Rd6f6 Nd4xf3 Nh5f4 Rd3d4
19/40- 01:09 84.048.572 1.210.969 -0,01 Re1xe5 d6xe5 d5d6 Rc7d7 Qd2e1 e5e4 Qe1xh4 e4xf3 Nb3d4 Rd7xd6 Rd1d3 Qb6c7 a2a3 b7b5 Qh4f2 Rd6f6 Nd4xf3 Nh5f4 Rd3d4
19/46 02:36 181.480.204 1.158.344 -0,09 Re1xe5 d6xe5 d5d6 Rc7d7 Qd2e1 e5e4 Qe1xh4 e4xf3 Nb3d4 Rd7xd6 Rd1d3 Rd6f6 Nd4xf3 Nh5f4 Rd3d2 Nf4d5 Rd2d3 Nd5e3 Qh4e4 Rf6e6
20/46 03:37 260.014.844 1.195.212 0,00 Re1xe5 d6xe5 d5d6 Rc7d7 Qd2e1 e5e4 Qe1xh4 e4xf3 Nb3d4 Rd7xd6 Rd1d3 Rd6f6 Nd4xf3 Nh5f4 Rd3d2 Nf4d5 Rd2d3 Nd5f4
20/46 04:59 355.919.582 1.186.647 0,00 Re1xe5 d6xe5 d5d6 Rc7d7 Qd2e1 e5e4 Qe1xh4 e4xf3 Nb3d4 Rd7xd6 Rd1d3 Rd6f6 Nd4xf3 Nh5f4 Rd3d2 Nf4d5 Rd2d3 Nd5f4
I stopped the analysis after five minutes more less.

------------------------

Analysis by Naum 4.2 w32 (using two cores) in an Intel Pentium D930 (3 GHz), with 512 MB of hash, running under Arena 2.0.1 GUI:
FEN: 8/1pr2pk1/pq1p2p1/3Pb2n/7r/1NP2B2/PP1Q3P/1K1RR3 w - - 0 30

Naum 4.2 x32:
1/9 00:01 236 236.000 +0,07 Re1e2
2/7 00:01 391 24.437 -0,09 Re1e2 Nh5f4
3/12 00:01 777 48.562 -0,12 Re1e2 Nh5f4 Re2e4
4/10 00:01 1.641 102.562 -0,16 Re1e2 Nh5f4 Re2e4 Rc7e7
5/11 00:01 2.103 131.437 -0,16 Re1e2 Nh5f4 Re2e4 f7f5 Re4e3
5/14 00:01 5.494 343.375 -0,02 Qd2g5 Be5f6 Qg5d2 Nh5f4 Re1e8
6/17 00:01 18.419 391.893 -0,29 Qd2g5 Qb6f2 Re1f1 Qf2xh2 Rf1h1 Be5f6
6/17 00:01 19.490 414.680 -0,11 Re1e2 Nh5f4 Re2e4 f7f5 Re4b4 Qb6a7
7/14 00:01 24.154 383.396 -0,02 Re1e2 Nh5f4 Re2e4 Rc7c8 Rd1e1 Rc8h8 Re4b4
8/16 00:01 40.620 520.769 -0,06 Re1e2 Rc7c4 Rd1e1 Nh5f4 Re2e3 Rh4h8 Bf3e4 Rh8e8
9/17 00:01 66.477 604.336 -0,07 Re1e2 Rc7c4 Rd1e1 Nh5f4 Re2e3 Rh4h8 Bf3e4 f7f5 Be4f3
10/20 00:01 102.041 649.942 -0,13 Re1e2 Rc7c4 Rd1e1 Nh5f4 Re2e3 Rh4h8 Bf3e4 f7f5 Be4f3 Rh8e8
11/22 00:01 229.891 774.043 -0,15 Re1e2 Rc7c4 Rd1e1 Nh5f4 Re2e3 Rh4h8 Bf3e4 f7f5 Be4f3 a6a5 Re1g1
12/34 00:02 933.464 947.679 -0,10 Re1xe5 d6xe5 d5d6 Rc7d7 Qd2e1 e5e4 Qe1xh4 e4xf3 Qh4d4+ Qb6xd4 Nb3xd4 f3f2
13/39 00:02 1.509.826 995.927 -0,17 Re1xe5 d6xe5 d5d6 Rc7d7 Qd2e1 e5e4 Qe1xh4 e4xf3 Nb3d4 Rd7xd6 Rd1d3 Rd6f6 Qh4f2
14/35 00:03 2.758.641 1.050.910 -0,19 Re1xe5 d6xe5 d5d6 Rc7d7 Qd2e1 e5e4 Qe1xh4 e4xf3 Qh4d4+ Qb6xd4 Nb3xd4 f3f2 Rd1f1 Nh5f6
15/41 00:06 6.029.990 1.128.366 -0,20 Re1xe5 d6xe5 d5d6 Rc7d7 Qd2e1 e5e4 Qe1xh4 e4xf3 Qh4d4+ Qb6xd4 Nb3xd4 f3f2 Rd1f1 Nh5f6 Rf1xf2
16/43 00:12 13.033.290 1.169.849 -0,18 Re1xe5 d6xe5 d5d6 Rc7d7 Qd2e1 e5e4 Qe1xh4 e4xf3 Qh4d4+ Qb6xd4 Nb3xd4 f3f2 Rd1f1 Nh5f6 Rf1xf2 Nf6e4
17/45 00:20 22.643.875 1.194.738 -0,16 Re1xe5 d6xe5 d5d6 Rc7d7 Qd2e1 e5e4 Qe1xh4 e4xf3 Qh4d4+ Qb6xd4 Nb3xd4 f3f2 Rd1f1 Nh5f6 Rf1xf2 Nf6e4 Rf2g2
18/45 00:34 41.033.429 1.218.042 -0,18 Re1xe5 d6xe5 d5d6 Rc7d7 Qd2e1 e5e4 Qe1xh4 e4xf3 Qh4d4+ Qb6xd4 Nb3xd4 f3f2 Rd1f1 Nh5f6 Rf1xf2 Nf6e4 Rf2g2 Ne4xd6
19/50 01:06 81.263.813 1.245.422 -0,15 Re1xe5 d6xe5 d5d6 Rc7d7 Qd2e1 e5e4 Qe1xh4 e4xf3 Qh4d4+ Qb6xd4 Nb3xd4 f3f2 Rd1f1 Nh5f6 Rf1xf2 Nf6e4 Rf2g2 Ne4xd6 h2h4
20/54 02:30 186.176.107 1.247.018 -0,19 Re1xe5 d6xe5 d5d6 Rc7d7 Qd2e1 e5e4 Qe1xh4 e4xf3 Qh4d4+ Qb6xd4 Nb3xd4 f3f2 Rd1f1 f7f5 Rf1xf2 Kg7f6 Kb1c2 Rd7xd6 Rf2e2 Nh5f4
21/53 04:25 329.743.523 1.246.667 -0,16 Re1xe5 d6xe5 d5d6 Rc7d7 Qd2e1 e5e4 Qe1xh4 e4xf3 Qh4d4+ Qb6xd4 Nb3xd4 f3f2 Rd1f1 f7f5 Rf1xf2 Kg7f6 Kb1c2 Rd7xd6 Rf2e2 Nh5f4 Re2e8
I stopped the analysis after five minutes.

------------------------

Analysis by Spike 1.2 Turin w32 (single core) in an Intel Pentium D930 (3 GHz), with 512 MB of hash, running under Arena 2.0.1 GUI:
FEN: 8/1pr2pk1/pq1p2p1/3Pb2n/7r/1NP2B2/PP1Q3P/1K1RR3 w - - 0 30

Spike 1.2:
0 00:00 0 0 +1,10
1 00:00 550 34.375 +0,99 Re1e2
1 00:00 628 39.250 +0,99 Re1e2
2 00:00 734 45.875 +0,72 Re1e2 Kg7g8
2 00:00 995 62.187 +0,72 Re1e2 Kg7g8
3 00:00 1.610 100.625 +0,83 Re1e2 Nh5f4 Qd2e1
3 00:00 3.167 197.937 +0,83 Re1e2 Nh5f4 Qd2e1
4 00:00 4.887 305.437 +0,54 Re1e2 Be5f4 Qd2c2 Nh5f6
4 00:00 8.568 267.750 +0,71 Qd2g5 Rh4h3 Bf3e2 Rh3xh2
4 00:00 10.510 328.437 +0,71 Qd2g5 Rh4h3 Bf3e2 Rh3xh2
5 00:00 19.126 303.587 +0,58 Qd2g5 Rh4h3 Re1xe5 d6xe5 d5d6 Rh3xf3 Qg5xe5+ f7f6 d6xc7 f6xe5 c7c8Q
5 00:00 23.328 370.285 +0,66 Re1e2 Nh5f4 Re2e4 f7f5 Re4b4
5 00:00 28.620 362.278 +0,66 Re1e2 Nh5f4 Re2e4 f7f5 Re4b4
6 00:00 36.253 385.670 +0,58 Re1e2 Nh5f4 Re2e4 f7f5 Re4b4 Qb6a7
6 00:00 52.881 375.042 +0,58 Re1e2 Nh5f4 Re2e4 f7f5 Re4b4 Qb6a7
7 00:00 76.520 407.021 +0,75 Re1e2 Nh5f4 Re2e4 f7f5 Re4b4 Qb6a7 Rd1e1
7 00:00 93.770 399.021 +0,75 Re1e2 Nh5f4 Re2e4 f7f5 Re4b4 Qb6a7 Rd1e1
8 00:00 128.217 409.638 +0,54 Re1e2 Nh5f4 Qd2e1 Rh4h3 Re2xe5 Rh3xf3 Re5e8 Nf4d3
8 00:01 314.509 437.425 +0,54 Re1e2 Nh5f4 Qd2e1 Rh4h3 Re2xe5 Rh3xf3 Re5e8 Nf4d3
9 00:01 513.287 432.059 +0,32 Re1e2 Nh5f6 Rd1e1 Rh4h3 Re2f2 Be5f4 Qd2c2 Bf4e3 Rf2f1
9 00:02 773.626 438.066 +0,39 Qd2g5 Be5f6 Qg5g2 Bf6e5 Re1h1 Rc7c4 Rd1e1 Kg7g8 h2h3
9 00:02 936.290 443.739 +0,39 Qd2g5 Be5f6 Qg5g2 Bf6e5 Re1h1 Rc7c4 Rd1e1 Kg7g8 h2h3
10 00:03 1.143.401 448.920 +0,44 Qd2g5 Be5f6 Qg5g2 Bf6e5 Re1xe5 d6xe5 d5d6 Rc7d7 Qg2g5 Qb6f2 Rd1d2 Qf2f1+ Nb3c1
10 00:03 1.342.668 449.805 +0,44 Qd2g5 Be5f6 Qg5g2 Bf6e5 Re1xe5 d6xe5 d5d6 Rc7d7 Qg2g5 Qb6f2 Rd1d2 Qf2f1+ Nb3c1
11- 00:04 1.893.073 445.428 +0,14 Qd2g5 Be5f6 Qg5g2 Bf6e5 Re1xe5 d6xe5 d5d6 Rc7d7 Qg2g5 Qb6f2 Rd1d2 Qf2f1+ Nb3c1
11 00:06 2.964.752 459.366 +0,25 Re1xe5 d6xe5 d5d6 Rc7d7 Qd2e1 e5e4 Qe1xh4 e4xf3 Qh4g4 Rd7xd6 Nb3d4 Rd6d5 Qg4xf3 Nh5f6
11 00:08 3.847.240 456.808 +0,25 Re1xe5 d6xe5 d5d6 Rc7d7 Qd2e1 e5e4 Qe1xh4 e4xf3 Qh4g4 Rd7xd6 Nb3d4 Rd6d5 Qg4xf3 Nh5f6
12 00:11 4.987.457 473.552 +0,35 Re1xe5 d6xe5 d5d6 Rc7d7 Qd2e1 e5e4 Qe1xh4 e4xf3 Qh4g4 Rd7xd6 Nb3d4 Rd6f6 Nd4xf3 Rf6f4 Qg4g2
12 00:12 5.831.646 476.052 +0,35 Re1xe5 d6xe5 d5d6 Rc7d7 Qd2e1 e5e4 Qe1xh4 e4xf3 Qh4g4 Rd7xd6 Nb3d4 Rd6f6 Nd4xf3 Rf6f4 Qg4g2
13 00:16 7.743.411 484.902 +0,24 Re1xe5 d6xe5 d5d6 Rc7d7 Qd2e1 e5e4 Qe1xh4 e4xf3 Nb3d4 Rd7xd6 Rd1d3 Qb6c5 Rd3xf3 Nh5f6 Qh4f4 Kg7g8
13 00:20 9.508.781 481.823 +0,24 Re1xe5 d6xe5 d5d6 Rc7d7 Qd2e1 e5e4 Qe1xh4 e4xf3 Nb3d4 Rd7xd6 Rd1d3 Qb6c5 Rd3xf3 Nh5f6 Qh4f4 Kg7g8
14 00:27 13.330.135 497.448 +0,37 Re1xe5 d6xe5 d5d6 Rc7d7 Qd2e1 e5e4 Qe1xh4 e4xf3 Nb3d4 Rd7xd6 Rd1d3 Rd6d5 Rd3xf3 Qb6d6 Rf3e3 Rd5e5 Re3e4
14 00:33 16.209.536 494.238 +0,37 Re1xe5 d6xe5 d5d6 Rc7d7 Qd2e1 e5e4 Qe1xh4 e4xf3 Nb3d4 Rd7xd6 Rd1d3 Rd6d5 Rd3xf3 Qb6d6 Rf3e3 Rd5e5 Re3e4
15 00:48 24.594.502 508.077 +0,10 Re1xe5 d6xe5 d5d6 Rc7d7 Qd2e1 e5e4 Qe1xh4 e4xf3 Nb3d4 Rd7xd6 Rd1d3 Rd6d5 Rd3xf3 Qb6d6 Rf3e3 Rd5e5 Re3xe5 Qd6xe5
15 01:18 39.413.531 503.481 +0,10 Re1xe5 d6xe5 d5d6 Rc7d7 Qd2e1 e5e4 Qe1xh4 e4xf3 Nb3d4 Rd7xd6 Rd1d3 Rd6d5 Rd3xf3 Qb6d6 Rf3e3 Rd5e5 Re3xe5 Qd6xe5
16 01:41 51.699.227 512.507 +0,24 Re1xe5 d6xe5 d5d6 Rc7d7 Qd2e1 e5e4 Qe1xh4 e4xf3 Nb3d4 Rd7xd6 Rd1d3 Rd6d5 Rd3xf3 Qb6d8 Qh4xd8 Rd5xd8 Rf3e3 Nh5f6 Re3e7
16 01:58 60.088.824 509.901 +0,24 Re1xe5 d6xe5 d5d6 Rc7d7 Qd2e1 e5e4 Qe1xh4 e4xf3 Nb3d4 Rd7xd6 Rd1d3 Rd6d5 Rd3xf3 Qb6d8 Qh4xd8 Rd5xd8 Rf3e3 Nh5f6 Re3e7
17 03:02 94.194.198 518.393 +0,08 Re1xe5 d6xe5 d5d6 Rc7d7 Qd2e1 e5e4 Qe1xh4 e4xf3 Nb3d4 Rd7xd6 Rd1d3 Rd6d5 Rd3xf3 Rd5e5 Rf3f1 Nh5f6 Qh4h3 Qb6d8 Qh3g3 Qd8d6 Nd4f5+
17 04:27 137.442.636 515.519 +0,08 Re1xe5 d6xe5 d5d6 Rc7d7 Qd2e1 e5e4 Qe1xh4 e4xf3 Nb3d4 Rd7xd6 Rd1d3 Rd6d5 Rd3xf3 Rd5e5 Rf3f1 Nh5f6 Qh4h3 Qb6d8 Qh3g3 Qd8d6 Nd4f5+
18 05:00 155.898.484 519.769 +0,08 Re1xe5 d6xe5 d5d6 Rc7d7 Qd2e1 e5e4 Qe1xh4 e4xf3 Nb3d4 Rd7xd6 Rd1d3 Rd6d5 Rd3xf3 Rd5e5 Rf3f1 Nh5f6 Qh4h3 Qb6d8 Qh3g3 Qd8d6 Nd4f5+
I stopped the analysis after five minutes.

------------------------

I know that Hannibal, Quazar, Spark, Naum and Spike are not in the Top 5.

Regards from Spain.

Ajedrecista.

Spliffjiffer
Posts: 248
Joined: Thu Aug 02, 2012 5:48 pm
Location: Germany

Re: Bug in Top-5 engines?

Post by Spliffjiffer » Wed Oct 24, 2012 6:40 pm

Hello Jesus :-)

Your diagram is correct (position before 30.Rxe5)

Thanks for the very intersting analysis snapshots of these not top 5 engines!

Its very interesting to see these engines solve this little puzzle (for us humans) that fast.
Especially whenHoudini&Co have much much more problems with it....must be something code-related i think.
Wahrheiten sind Illusionen von denen wir aber vergessen haben dass sie welche sind.

Dan Astrachan
Posts: 134
Joined: Fri May 28, 2010 9:07 am

Re: Bug in Top-5 engines?

Post by Dan Astrachan » Thu Oct 25, 2012 11:53 am

[D]8/1pr2pk1/pq1p2p1/3Pb2n/7r/1NP2B2/PP1Q3P/1K1RR3 w - - 0 29


This is the position after 29. ...Be5

All the engines I've tried (Ivanhoe 9.46h, Critter 1.6a, Stockfish 2.3.1, Strelka 5.5, Houdini 1.5a, Robbolito 0.10 and Fire 2.2) find Re5 very quickly.
Computer used: Intel Core i7 875K @ 2.93GH, 4 cores, 12 Gb of RAM

ernst
Posts: 336
Joined: Thu Mar 09, 2006 5:00 pm

Re: Bug in Top-5 engines?

Post by ernst » Thu Oct 25, 2012 12:59 pm

Fwiw.
  • 8192 MB Large Page Hash
    Nalimov 5 men EGTB available - 32 MB cache
    Engine: Houdini 3 Pro x64 (8192MB)
    by Robert Houdart

    19/46 0:01 -0.60 29...a5 30.Bxh5 Rxh5 31.Re4 Qb5 32.Qf2 Rxd5 33.Nd4 Bxd4 34.Rexd4 Qc5 35.Qf4 Rxd4 36.Qxd4+ Qxd4 37.Rxd4 Rc6 38.Kc2 f5 39.Rd5 b6 40.Kd3 Kf6 41.h4 Ke6 42.Rb5 d5 (18.214.170) 12518
    20/49 0:02 -0.63 29...a5 30.Bxh5 Rxh5 31.Re4 Qb5 32.Qf2 Rxd5 33.Nd4 Bxd4 34.Rexd4 Qc5 35.Qf4 Rxd4 36.Qxd4+ Qxd4 37.Rxd4 Rc6 38.Kc2 f5 39.Rd5 b6 40.h4 Kf6 41.Rd1 a4 42.Kd3 Rc7 (26.696.596) 12878
    21/51 0:02 -0.63 29...a5 30.Bxh5 Rxh5 31.Re4 Qb5 32.Qf2 Rxd5 33.Nd4 Bxd4 34.Rexd4 Qc5 35.Qf4 Rxd4 36.Qxd4+ Qxd4 37.Rxd4 Rc6 38.Kc2 f5 39.Rd5 b6 40.h4 Kf6 41.Rd1 a4 42.Kd3 Rc7 (36.320.455) 13140
    22/54 0:06 -0.69 29...a5 30.Bxh5 Rxh5 31.Re4 Qb5 32.Qf2 Rxd5 33.Nd4 Bxd4 34.Rexd4 Qc5 35.Qf4 Rxd4 36.Qxd4+ Qxd4 37.Rxd4 Rd7 38.Kc2 f5 39.Kd3 d5 40.h4 b6 41.Ke3 Kf6 42.Kf3 Ke5 (91.429.499) 13355
    23/54 0:18 -0.60-- 29...a5 30.Re4 (200.353.450) 10691
    23/59 0:57 -0.80++ 29...Rcc4 (718.425.814) 12455
    23/62 1:27 -0.71 29...Rcc4 30.Re8 a5 31.Qe3 Qxe3 32.Rxe3 a4 33.Na5 Rc7 34.Bxh5 Rxh5 35.Re4 b5 36.Nc6 Rxh2 37.a3 Rh4 38.Rxh4 Bxh4 39.Rd4 g5 40.Re4 Bf2 41.Kc2 Bc5 42.Kd3 Kf6 (1.109.004.370) 12697
    24/62 1:31 -0.71 29...Rcc4 30.Re8 a5 31.Qe3 Qxe3 32.Rxe3 a4 33.Na5 Rc7 34.Bxh5 Rxh5 35.Re4 b5 36.Nc6 Rxh2 37.a3 Rh4 38.Rxh4 Bxh4 39.Rd4 g5 40.Re4 Bf2 41.Kc2 Bc5 42.Kd3 Kf6 (1.159.281.296) 12729
    25/62 1:41 -0.71 29...Rcc4 30.Re8 a5 31.Qe3 Qxe3 32.Rxe3 a4 33.Na5 Rc7 34.Bxh5 Rxh5 35.Re4 b5 36.Nc6 Rxh2 37.a3 Rh4 38.Rxh4 Bxh4 39.Rd4 g5 40.Re4 Bf2 41.Kc2 Bc5 42.Kd3 Kf6 (1.299.272.969) 12845
    26/62 2:03 -0.71 29...Rcc4 30.Re8 a5 31.Qe3 Qxe3 32.Rxe3 a4 33.Na5 Rc7 34.Bxh5 Rxh5 35.Re4 b5 36.Nc6 Rxh2 37.a3 Rh4 38.Rxh4 Bxh4 39.Rd4 g5 40.Re4 Bf2 41.Kc2 Bc5 42.Kd3 Kf6 (1.603.754.032) 12981
    27/62 2:40 -0.73 29...Rcc4 30.Re8 a5 31.Qe3 Qxe3 32.Rxe3 a4 33.Na5 Rc7 34.Bxh5 Rxh5 35.Re4 b5 36.Nc6 Rxh2 37.a3 Rb7 38.Rb4 Rh4 39.Rg1 Rxb4 40.axb4 Rc7 41.Kc2 Rc8 42.Rg2 Ra8 (2.129.474.705) 13270
    27/62 2:54 -0.73 29...Rcc4 30.Re8 a5 31.Qe3 Qxe3 32.Rxe3 a4 33.Na5 Rc7 34.Bxh5 Rxh5 35.Re4 b5 36.Nc6 Rxh2 37.a3 Rb7 38.Rb4 Rh4 39.Rg1 Rxb4 40.axb4 Rc7 41.Kc2 Rc8 42.Rg2 Ra8 (2.333.143.380) 13344
    best move: Rc7-c4 time: 2:55.281 min n/s: 13.344.000 CPU 99.1% n/s(1CPU): 13.465.186 nodes: 2.333.143.380

    Hash cleared

    19/47 0:01 -0.23 30.Rxe5 dxe5 31.d6 Rd7 32.Qe1 e4 33.Qxh4 exf3 34.Qd4+ Qxd4 35.Nxd4 f2 36.Rf1 Rxd6 37.Rxf2 Rf6 38.Rd2 Rf1+ 39.Kc2 Nf4 40.Nb3 Ne6 41.Na5 Nc5 42.Re2 Rf3 43.Nb3 (22.276.101) 12671
    20/49 0:02 -0.24 30.Rxe5 dxe5 31.d6 Rd7 32.Qe1 e4 33.Qxh4 exf3 34.Qd4+ Qxd4 35.Nxd4 f2 36.Rf1 Rxd6 37.Rxf2 Rf6 38.Rd2 Rf1+ 39.Kc2 Nf4 40.Nb3 Ne6 41.Na5 Nc5 42.Re2 Rf3 43.Nb3 (36.650.690) 13136
    21/55 0:06 -0.26 30.Rxe5 dxe5 31.d6 Rd7 32.Qe1 e4 33.Qxh4 exf3 34.Qd4+ Qxd4 35.Nxd4 f2 36.Rf1 Nf6 37.Rxf2 Ne4 38.Rg2 Kh6 39.Kc2 f5 40.Kd1 Kh5 41.Ke2 Nxd6 42.Kf3 Re7 43.Re2 (84.696.976) 13564
    22/58 0:09 -0.26 30.Rxe5 dxe5 31.d6 Rd7 32.Qe1 e4 33.Qxh4 exf3 34.Qd4+ Qxd4 35.Nxd4 f2 36.Rf1 Nf6 37.Rxf2 Ne4 38.Rg2 Kh6 39.Kc2 f5 40.Kd1 Kh5 41.Ke2 Nxd6 42.Kf3 Re7 43.Re2 (123.784.618) 13682
    23/59 0:16 -0.21 30.Rxe5 dxe5 31.d6 Rd7 32.Qe1 e4 33.Qxh4 exf3 34.Qd4+ Qxd4 35.Nxd4 f2 36.Rf1 Rxd6 37.Rxf2 Rf6 38.Rd2 Rf4 39.Kc2 Nf6 40.Re2 Nd5 41.Kb3 Rf1 42.Kc2 Rf6 43.Rd2 (224.421.178) 13711
    24/59 0:19 -0.21 30.Rxe5 dxe5 31.d6 Rd7 32.Qe1 e4 33.Qxh4 exf3 34.Qd4+ Qxd4 35.Nxd4 f2 36.Rf1 Rxd6 37.Rxf2 Rf6 38.Rd2 Rf4 39.Kc2 Nf6 40.Re2 Nd5 41.Kb3 Rf1 42.Kc2 Rf6 43.Rd2 (264.663.014) 13768
    25/63 0:36 -0.29-- 30.Rxe5 dxe5 (504.379.112) 13932
    25/63 0:49 -0.28 30.Rxe5 dxe5 31.d6 Rd7 32.Qe1 e4 33.Qxh4 exf3 34.Qd4+ Qxd4 35.Nxd4 f2 36.Rf1 Nf6 37.Rxf2 Ne4 38.Rg2 Nxd6 39.Re2 f5 40.h4 Kh6 41.Kc2 Ne4 42.Ne6 Rd6 43.Nf4 (689.541.486) 13981
    26/63 1:04 -0.28 30.Rxe5 dxe5 31.d6 Rd7 32.Qe1 e4 33.Qxh4 exf3 34.Qg4 Rxd6 35.Qxf3 Rxd1+ 36.Qxd1 Nf6 37.Nd4 Qc7 38.Qf3 b5 39.Qc6 Qxc6 40.Nxc6 g5 41.Ne5 Ne4 42.Kc2 f5 43.Kd3 (900.718.570) 14009
    27/67 2:38 -0.30 30.Rxe5 dxe5 31.d6 Rd7 32.Qe1 e4 33.Qxh4 exf3 34.Qd4+ Qxd4 35.Nxd4 f2 36.Rf1 Nf6 37.Rxf2 Ne4 38.Rg2 Nxd6 39.Re2 f5 40.Ne6+ Kf6 41.Nc5 Re7 42.Rxe7 Kxe7 43.Kc2 (2.203.698.476) 13915
    27/67 3:12 -0.30 30.Rxe5 dxe5 31.d6 Rd7 32.Qe1 e4 33.Qxh4 exf3 34.Qd4+ Qxd4 35.Nxd4 f2 36.Rf1 Nf6 37.Rxf2 Ne4 38.Rg2 Nxd6 39.Re2 f5 40.Ne6+ Kf6 41.Nc5 Re7 42.Rxe7 Kxe7 43.Kc2 (2.683.545.526) 13919
    best move: Re1xe5 time: 3:12.812 min n/s: 13.919.000 CPU 99.7% n/s(1CPU): 13.960.882 nodes: 2.683.545.526

Uri Blass
Posts: 8796
Joined: Wed Mar 08, 2006 11:37 pm
Location: Tel-Aviv Israel

Re: Bug in Top5 Engines?

Post by Uri Blass » Thu Oct 25, 2012 2:03 pm

Spliffjiffer wrote:V.Anand vs M.Sadler:

1. e4 c5 2. Nf3 d6 3. d4 cxd4 4. Nxd4 Nf6 5. Nc3 a6 6. Be3 e5 7. Nb3 Be6 8.f3 Be7 9. Qd2 h5 10. O-O-O Nbd7 11. g3 Rc8 12. Nd5 Bxd5 13. exd5 Nb6 14.Bxb6 Qxb6 15. Bh3 Rc7 16. Kb1 h4 17. Qe1 Kf8 18. f4 exf4 19. gxf4 Rh5 20.Bg2 Rc4 21. Rf1 Rf5 22. Qxh4 Nh5 23. Qh3 g6 24. Rfe1 Bf6 25. Be4 Rxf4 26.Qd3 Rc7 27. Bf3 Rh4 28. Qd2 Kg7 29. c3 Be5 30. Rxe5 1/2:1/2

[D]8/1pr2pk1/pq1p2p1/3PR2n/7r/1NP2B2/PP1Q3P/1K1R4 b - - 0 30

Sadler threw his hands up in horror, he had thrown away his winning position

Instead of 29. ...Be5? black should have played 29. ...Rcc4 the players agreed in the analysis afterwards...(also 29. ...a5 looks very promising)

So far so good...but not for the top 5 engines...


Houdini 1.5, Critter 1.6a, Komodo 3, Stockfish 2.3.1 and Rybka 3 dynamic (well...if R3 dy. is not a top 5 engine you can probably use R4) have a common bug here:
After 29. ...Be5 they not only have big problems to find 30. Rxe5, they also have the same bug when you go backwards from the position which was reached after:

30.Qg5 Bf6, 31.Qd2 Rcc4, 32.Rf1...

[D]8/1p3pk1/pq1p1bp1/3P3n/2r4r/1NP2B2/PP1Q3P/1K1R1R2 b - - 6 32


Note: Use the "playing mode" and NOT the "analysis mode" if you use Houdini 1.5

Now...choose any timecontrol, let's say 40/20 or even more, and let the engines compute on the 32. move for white and wait until the engine makes its move.
You will see that the engine is quite unconfotable with the position.
After any move was played, go back to the 31. move for white and let the engine compute again...wait for the move...
Go back to the 30. move for white and let it compute again.
Now the engine should know that 30.Qg5 is very unpleasant because of the quite forced Bf6, Qd2 - maneuver and the consequences you were computing for in move 32
...but it don't...
It will think that the position is drawish after 30.Qg5 !!!

This works also in the "analysis mode" of Critter, Komodo,Stockfish and Rybka 3 dynamic !!!
Only houdini 1.5 is different in "analysis mode".

Maybe this is not a bug but anyway...the engines would probably loose the game after 30.Qg5??.

Do you have the same bug?

I used Windows XP prof. (32 bit),E6600 Core 2 Duo @ 2,4 GHz



Regards
No bug in houdini and also no bug in other programs.
The problem is simply because you give the program all the game and not only the position and you give them a game that they cannot play
so it is not a situation from a real game that they played.


Note that playing mode is not supposed to be analysis mode and analysis mode in many programs is identical to playing mode except maybe contempt=0(the last thing is not a bug because usually programmars care about playing and not about analyzing and it is not that they invested time in improving their analysis mode).

Playing mode is supposed to be relevant only for playing conditions.
In playing a game houdini is not going to get the same game in the first place.

Note that the line 30.Qg5 Bf6 31.Qd2 leads to repetition of the position
after 29.c3 and usually programs evaluate it as a draw and it is no bug.

If houdini plays with white
There are mainly only 2 options:
1)It plays against a weaker opponent so it will never get the position after 29.c3 with white in the first place because this position is better for black and the opponent is not smart enough to get a better position.

2)It plays against a strong opponent so there is no way that the opponent is going to miss a better move than 29...Be5

In both cases houdini is not going to get the situation when it does the mistake Qg5.

The only reason that houdini and other top programs gets the situation when it does the mistake Qg5 is that you put it in illogical conditions(meaning condition that cannot happen in their game and ask them to play like they are in a game when the game has mistakes by both sides).

If you give every top program the position without the game then it is going to find easily the move Rxe5.

Uri Blass
Posts: 8796
Joined: Wed Mar 08, 2006 11:37 pm
Location: Tel-Aviv Israel

Re: Bug in Top5 Engines?

Post by Uri Blass » Thu Oct 25, 2012 2:26 pm

Spliffjiffer wrote:V.Anand vs M.Sadler:

Maybe this is not a bug but anyway...the engines would probably loose the game after 30.Qg5??.

Regards
I answered but here is a shorter answer.

The engines are not going to play 30.Qg5 in a real game.
The reason that they play 30.Qg5 in playing mode when you give them the game is simply that you give them a game that is not their game and they evaluate a repetition of previous game position as a draw.

If it is their game and their opponent is a computer who got a better position then the opponent usually is not going to allow them to repeat a previous game position(meaning the opponent is not going to play 29...Be5 but 29...Rcc4 and there is no reason that a computer prefers the line 29....Be5 30.Qg5 Bf6 31.Qd2 Rcc4).

Spliffjiffer
Posts: 248
Joined: Thu Aug 02, 2012 5:48 pm
Location: Germany

Re:Bug in Top5 Engines?

Post by Spliffjiffer » Thu Oct 25, 2012 3:20 pm

Thank you Uri for the good explanation...i understand the problem now :-)

If i only copy and paste the fen-string of the position after 29...Be5 and let the program think, it will find the 30.Rxe5 move.

But if i copy and paste the "whole game" until 31.Rf1 (in my example), where a 1-fold-repetition was already made (30.Qg5 Bf6, 31.Qd2), the program "thinks" it should repeat the moves because its position is much worse...and so...why not to go for a 3-fold-repetition.

But for analysis this behavior is still of corse "terrible"...i mean...critter,komodo,stockfish and rybka are quite pointless in analysing different moves at around move 30 when you go back and forward through the position.




Regards from germany
Wahrheiten sind Illusionen von denen wir aber vergessen haben dass sie welche sind.

Post Reply