interesting playchess game (Naum 4.1-Rybka 3)

[jdart]

[Event "?"]
[Site "?"]
[Date "2010.02.28"]
[Round "?"]
[White "Future Breeze, Naum 4.1"]
[Black "Ultrapower, Rybka 3"]
[Result "1-0"]
[ECO "E01"]
[WhiteElo "2622"]
[BlackElo "2638"]
[PlyCount "41"]
[EventDate "2010.03.01"]
[TimeControl "7200"]

1. d4 {0} d5 {0.00/0 0} 2. Nf3 {0} Nf6 {0.00/0 0} 3. c4 {0} dxc4 {0.00/0 0} 4.
g3 {0} e6 {0.00/0 0} 5. Bg2 {0} c6 {0.00/0 0} 6. Nc3 {0} Nbd7 {0.00/0 0} 7. O-O
{0} b5 {0.00/0 0} 8. e4 {0} Bb7 {0.00/0 0} 9. e5 {0} Nd5 {0.00/0 0} 10. Ng5 {0}
Be7 {0.00/0 0} 11. Qh5 {0} g6 {0.00/0 0} 12. Qh6 {0} Bf8 {0.00/0 0} 13. Qh3 {0}
Nxc3 {0.00/0 0} 14. bxc3 {0} Be7 {0.00/0 0} 15. Ne4 {0} O-O {0.00/0 0} 16. Qh6
{208} Rb8 {0.29/19 460} 17. Re1 {0 (Ng5)} Re8 {0.34/19 93} 18. Bg5 {230} Bf8 {
0.36/21 0} 19. Qh4 {1691 (Bxd8)} Qc7 {0.21/17 101} 20. g4 {0 (a4)} c5 {2.27/14
270} 21. Re3 {0 Ultrapower,Rybka 3 resigns (Lag: Av=0.18s, max=0.4s)} 1-0

This was long time control (120 minutes/game).

19. Qh4 was a very interesting move. Naum played this after a long think, but the PGN doesn't show Naum's score.

After 19. .. Qh4 we have this position:

[D] 1r1qrbk1/pb1n1p1p/2p1p1p1/1p2P1B1/2pPN2Q/2P3P1/P4PBP/R3R1K1 b - - 0 19

and here I think 19. .. Be7 may save Black - White is better but I don't see a win (if there is one, then 19. Qh4 was truly brilliant).

This is a good test position (play Be7, avoid Qc7, Qb6, Qa5). 20. g4! (clearly winning, and itself a nice test) is not too hard to find, but backing up a ply makes it much harder for engines to see.

[jdart]

Since .. Be7 allows a repetition, I think Rybka 3 is misbehaving on this position in ChessBase because of the repetition scoring bug it has (score shown as 0 when repetition is possible, even when there is a win).

It is not actually so hard, although Arasan loses it then finds it again:

Arasan 12.2 Copyright 1994-2010 by Jon Dart
test be7.epd 1200 -v
Warning: could not open book file
found 5-man tablebases in directory c:\chess\tb
bm Be7
0 0.01 Qc7 +0.12 16072
0 0.01 Be7 +0.20 16092
1 0.01 Be7 -0.08 16138
1 0.01 Qc7 -0.04 16139
1 0.01 Qa5 +0.08 16171
2 0.01 Qa5 +0.08 16173
3 0.01 Qa5 +0.08 16208
4 0.03 Qa5 +0.12 18818
5 0.04 Qa5 +0.28 38598
6 0.09 Qa5 +0.20 80032
7 0.14 Qa5 +0.40 155841
8 0.25 Qa5 +0.40 381482
9 0.45 Qa5 +0.41 915470
10 0.82 Qa5 +0.45 1705331
11 1.51 Qa5 +0.28 3463900
12 2.72 Qa5 +0.18 6560795
13 5.59 Qa5 +0.18 13222218
14 13.82 Qa5 +0.20 34553583
15 21.65 Qa5 -0.16 55614944
15 29.32 Qb6 +0.00 76095368
16 64.79 Qc7 +0.00 163955349
17 84.87 Qc7 -0.48 215592994
17 131.18 Be7 -0.44 319499187
18 140.82 Be7 -0.60 347153687
18 165.40 Qc7 -0.48 417776463
19 182.84 Qc7 -0.97 466062665
19 252.09 Be7 -0.51 664514780
20 288.87 Be7 -0.57 770894955
21 450.89 Be7 -0.48 1244047989
22 735.42 Be7 -0.59 2067714390
2.85M nodes/second.
0 tablebase probes, 0 tablebase hits
6362 splits, average thread usage=4.00
result: Be7 score: -0.59 ++ solved in 252.09 sec. (664.51M nodes)

Be7 a4 c5 axb5 Bxe4

solution times:
0 1 2 3 4 5 6 7 8 9

0 | 252.09

correct : 1/1
nodes to solution : 664.51M
depth to solution : 18.00
time to solution : 252.09 sec.
test complete

[mhalstern]

I'm bored and have spent most of the day looking at this. Stockfish likes Be7 and Zappa weirdly Likes Re7, both engines showing Black as worse. Look at this line that I think is best:

After Be7

20 Nxf6+ Nxf6 21 exf6 Bf8 22 g4!

It looks like a forced loss for Black here. Please correct me if I am wrong. If this is correct, then it is hard to see exactly where Black went wrong in the opening. Instinctively, I don't like 13 .. Nxc3, as I think the centralized knight is needed for defense.

[jdart]

I think you are right about that.

Engines just go bonkers on this game though.

I looked at:
20. .. Be7 21. Nf6+ Nxf6 22. exf6 Bf8 23. g4! h6

[D] 1r1qrbk1/pb3p2/2p1pPpp/1p4B1/2pP2PQ/2P5/P4PBP/R3R1K1 w - - 0 23

and now several engines I tried want to play Bf4 for a long time - when the obvious Bxh6 is winning.

[Eelco de Groot]

Please correct me if I am wrong.

I could not find anything with Rainbow Serpent so I think there is a bug, especially if Stockfish can find Be7. I did check with free Rybka 2.3.2a ripped from the Peshka application and it, she, agrees with you, although it took quite a while to show a convincing score, almost 1600 minutes but for Rybka this is indicating a lost position:


1r1qrbk1/pb1n1p1p/2p1p1p1/1p2P1B1/2pPN2Q/2P3P1/P4PBP/R3R1K1 b - -

Engine: Rybka 2.3.2a PV-tips32-bit (256 MB)
by Vasik Rajlich

5.00 0:00 +0.07 19...Qa5 20.Bf3 b4 21.Rec1 Qa3
22.Bf6 c5 23.Ng5 Nxf6 24.Bxb7 (6.344) 58

6.00 0:00 +0.12 19...Qa5 20.Bf6 b4 21.Rec1 bxc3
22.Rxc3 c5 23.Rac1 Bd5 24.Ng5 Nxf6
25.exf6 (16.998) 61

7.00 0:00 +0.11 19...Qa5 20.Re2 b4 21.cxb4 Qxb4
22.a3 Qb5 23.Nd6 Bxd6 24.exd6 c3
25.Qg4 c5 (53.323) 61

8.00 0:01 +0.13 19...Qa5 20.Re2 b4 21.cxb4 Qxb4
22.a3 Qb5 23.Bf6 h5 24.a4 Qb3 25.Qf4 (73.465) 63

9.00 0:03 +0.04 19...Qa5 20.Red1 b4 21.cxb4 Qxb4
22.Rdc1 Ba6 23.a3 Qb2 24.Nd6 Bxd6
25.exd6 Rb6 26.Bd8 Rb8 (218.368) 60

10.00 0:05 +0.06 19...Qa5 20.Red1 Qa3 21.Bc1 Qa4
22.Bg5 b4 23.cxb4 Bxb4 24.a3 Bf8
25.Rdb1 c5 26.Nd6 Bxd6 27.Bxb7 (314.172) 62

11.00 0:06 +0.03 19...Qa5 20.Red1 Qa3 21.Bc1 Qa4
22.Be3 b4 23.cxb4 Be7 24.Ng5 Bxg5
25.Bxg5 Qxb4 26.Rdb1 Qa3 27.Be4 (406.069) 63

12.01 0:10 +0.05 19...Qa5 20.Red1 Qa3 21.Bc1 Qa4
22.Bf4 b4 23.cxb4 Bxb4 24.Nf6+ Nxf6
25.exf6 Rbd8 26.Rab1 c3 27.Qh6 Bf8 (676.421) 63

13.01 0:17 0.00 19...Qa5 20.Red1 Qa3 21.Bc1 Qa4
22.Bh6 Be7 23.Bg5 Bf8 24.a3 Qc2
25.Rdc1 Qe2 26.a4 b4 27.Bf1 (1.069.840) 63

14.01 0:36 0.00 19...Qa5 20.Red1 Qa3 21.Bc1 Qa5
22.Bg5 Qa3 23.Bc1 Qa5 24.Bg5 Qa3
25.Bc1 Qa5 26.Bg5 Qa3 27.Bc1 Qa5
28.Bg5 Qa3 29.Bc1 Qa5 30.Bg5 Qa3
31.Bc1 Qa5 32.Bg5 Qa3 (2.330.126) 64

15.01 3:25 -1.84 19...Qa5 20.g4 Ba3 21.Re3 f6 22.Rh3 Re7
23.Bxf6 Rbe8 24.Bxe7 Rxe7 25.Nf6+ Nxf6
26.exf6 Rd7 27.Re1 Bf8 28.Rxe6 Qxa2
29.Re8 Qa1+ 30.Bf1 c5 31.dxc5 h6 (10.881.818) 54

15.02 7:26 -0.49 19...Be7 20.a4 a6 21.Qf4 Bxg5
22.Nxg5 Qe7 23.axb5 cxb5 24.Bxb7 Rxb7
25.Rxa6 b4 26.cxb4 h6 27.Ne4 Qxb4
28.Rea1 g5 29.Qf3 Qb3 30.Qh5 (27.978.861) 64

15.18 11:29 -0.14 19...Qc7 20.a4 a5 21.Qf4 b4 22.Re3 Rec8
23.Rf3 Nb6 24.Nf6+ Kh8 25.cxb4 axb4
26.a5 Nd5 27.Nxd5 cxd5 28.Qxf7 Qxf7
29.Rxf7 (43.297.752) 64

16.01 11:50 -0.18 19...Qc7 20.a4 a5 21.Qf4 b4 22.Re3 Rec8
23.Rf3 f5 24.exf6 Qxf4 25.Bxf4 c5
26.Re1 cxd4 27.Bxb8 Rxb8 28.cxd4 c3
29.Rf4 (44.686.683) 64

17.01 12:41 -0.17 19...Qc7 20.a4 a5 21.Qf4 b4 22.Re3 Rec8
23.cxb4 axb4 24.a5 b3 25.Nd6 f6
26.Bh3 f5 27.Nxc4 Ba6 (47.942.081) 64

18.01 27:47 -0.23 19...Qc7 20.Nf6+ Nxf6 21.exf6 Qa5
22.Re3 c5 23.g4 Bxg2 24.Kxg2 Qd8
25.Rh3 h6 26.Bxh6 Bxh6 27.Qxh6 Qxf6
28.dxc5 Qg7 29.Qf4 e5 30.Qe4 Qf6
31.c6 (104.502.436) 64

19.01 42:35 -0.42 19...Qc7 20.Nf6+ Nxf6 21.exf6 c5
22.Bf4 Qa5 23.Bxb8 Bxg2 24.Bc7 Qxc7
25.Kxg2 cxd4 26.Qxd4 Rd8 27.Qe5 Qc6+
28.Qe4 Qxe4+ 29.Rxe4 Ba3 30.Re2 a6
31.Rae1 Rd3 (155.982.466) 62

20.01 70:07 -1.85 19...Qc7 20.g4 c5 21.Re3 Bxe4 22.Rh3 h6
23.Bxe4 cxd4 24.cxd4 Nc5 25.Bc2 Red8
26.dxc5 Qxe5 27.Bf6 Qe2 28.Bxd8 Qxc2
29.Be7 (239.714.071) 58

20.02 95:38 -0.49 19...Be7 20.a4 a6 21.Qf4 Bxg5
22.Nxg5 Qe7 23.axb5 cxb5 24.Bxb7 Rxb7
25.Rxa6 Nb6 26.Ne4 Nd5 27.Qh6 Qf8
28.Qxf8+ Rxf8 29.Nd6 Rbb8 30.Rc1 Kg7 (347.733.733) 62

21.01 119:48 -0.46 19...Be7 20.a4 a6 21.Qf4 Bxg5
22.Nxg5 Qe7 23.axb5 cxb5 24.Bxb7 Rxb7
25.Rxa6 Nb6 26.Ne4 Nd5 27.Qh6 Qf8
28.Qxf8+ Rxf8 29.Nd6 Rbb8 30.Rc1 Kg7 (440.872.815) 62

22.01 147:11 -0.51 19...Be7 20.a4 a6 21.Qf4 Bxg5
22.Nxg5 Qe7 23.axb5 cxb5 24.Bxb7 Rxb7
25.Rxa6 Nb6 26.Ne4 Nd5 27.Qh6 Qf8
28.Qxf8+ Rxf8 29.Nd6 Rd7 30.Rc1 Rb8
31.Ra5 (547.894.087) 63

23.01 228:09 -0.47 19...Be7 20.a4 a6 21.Qf4 Rf8 22.Bxe7 Qxe7
23.Qh6 f5 24.Nd6 Nb6 25.axb5 axb5
26.h4 Na4 27.Rec1 Ba8 28.Nxc4 Nxc3
29.Rxc3 bxc4 30.Rxc4 Rb6 31.Rac1 (878.575.431) 65

24.01 424:13 -0.74 19...Be7 20.Nf6+ Nxf6 21.exf6 Bf8
22.g4 h6 23.Bf4 b4 24.cxb4 Qxd4
25.Be5 Qd2 26.Red1 Qxb4 27.Bxb8 Rxb8
28.Rab1 Qa3 29.Qg3 Qxg3 30.hxg3 Bc5
31.Rdc1 Bb6 32.Rxc4 (1.639.324.329) 65

25.01 1558:13 -2.20 19...Be7 20.Nf6+ Bxf6 21.exf6 e5
22.Qh6 Nxf6 23.Qh4 exd4 24.Bxf6 Qd6
25.Qxd4 Qxd4 26.Bxd4 a5 27.Rxe8+ Rxe8
28.Bb6 Ra8 29.Rd1 b4 30.cxb4 axb4
31.Rd7 c3 32.Rxb7 c2 (5.544.632.665) 60

best move: Bf8-e7 time: 1625:58.734 min n/s: 60.138 nodes: 5.729.508.205