Gingko - Hiarcs, WCCC 2015

[zullil]

Stockfish (even newer than latest developmental version) (1 thread, 1 GB hash):

Is this version somewhere available?

Wait a day or two, or compile in the change shown here:
https://github.com/official-stockfish/S ... h/pull/373

[S.Taylor]

In the first round of the World Computer Chess Championship Gingko found an interesting exchange sac to build a fortress.
[d]
42.Rxe4! does the job. After over 40 moves later Hiarcs agreed to a draw.

It's very interesting to see that all engines act very differently with this position. There are a couple of engines, which see the sac instantly:
- BlackMamba 2.0
- Fire 3.0 (not Fire 4.0)
- Firenzina 2.3.2
- Firenzina 2.4.1
- Maverick 1.0
- PanChess 00.611
- Rodent 1.7

Some engines don't find it immediately, but within 10 seconds:

Hardly strong enough to challenge an average schoolchild today!

[Jesse Gersenson]

I showed this position to a few GM's and it took them a few seconds to find Rxe4 draw. One of them laughed and said Gingko got lucky, finding the move but not realizing it was a fortress.

[Jesse Gersenson]

I'm curious to see Komodo 9.1's result on 1 core with
2 GB Hash. Could you please test it?

P9600 @ 2.66GHz

23+ 00:53 44,491k 827k -3.98 a4-a5
23- 00:54 45,249k 827k -4.05 a4-a5 Qd6-c7
23- 01:02 51,320k 826k -4.15 a4-a5 Qd6-c7
23+ 01:07 55,662k 826k -4.06 a4-a5
23 01:08 56,325k 827k -3.96 a4-a5 Qd6-c7 Re1-c1 Qc7-h7 Bg5-h4 Qh7-h6 Rc1-d1 Be4-f3 Bc4-d5+ Bf3xd5 Rd1xd5 Qh6-c1+ Kg1-g2 Qc1-c3 Rd5-d7+ Kb7-c6 Rd7-a7 Qc3xa5 Bh4-f6 Qa5-b6 Ra7-e7 Qb6xb3 Re7xe5 Qb3-f3+ Kg2-g1 c5-c4 Re5-e6+ Kc6-b5 Re6-e3 Qf3-d1+ Kg1-g2 Qd1-d5+ Kg2-g1 Kb5-b4 Bf6-c3+ Kb4-b3 Kg1-h2 Kb3-c2
24- 01:09 57,715k 826k -4.03 a4-a5 Qd6-c7
24- 01:12 60,050k 827k -4.10 a4-a5 Qd6-c7
24- 01:19 65,446k 828k -4.20 a4-a5 Qd6-c7
24+ 01:22 68,333k 830k -4.03 Re1xe4
24+ 01:22 68,786k 831k -3.66 Re1xe4
24 01:24 70,078k 831k -3.70 Re1xe4 f5xe4 Bg5-e3 Kb7-b6 Kg1-g2 a6-a5 Kg2-g1 Qd6-d1+ Kg1-g2 Kb6-c6 Kg2-h2 Kc6-d6 Kh2-g2 Qd1-e1 Kg2-h2 Qe1-a1 Bc4-e2 Qa1-c3 Be2-c4 Qc3-e1 Kh2-g2 Qe1-d1 Kg2-h2 Kd6-c6 Bc4-b5+ Kc6-b6 Bb5-c4 Qd1-f3 Kh2-g1 Kb6-c6 Kg1-h2 Qf3-f5
25+ 01:24 70,252k 832k -3.63 Re1xe4
25- 01:24 70,598k 832k -3.70 Re1xe4 f5xe4
25+ 01:26 72,003k 832k -3.63 Re1xe4
25+ 01:26 72,129k 832k -3.48 Re1xe4
25- 01:26 72,219k 832k -3.59 Re1xe4 f5xe4
25 01:27 72,721k 833k -3.66 Re1xe4 f5xe4 Bg5-e3 a6-a5 Kg1-g2 Kb7-b6 Kg2-g1 Qd6-d1+ Kg1-g2 Kb6-c6 Bc4-b5+ Kc6-d5 Bb5-c4+ Kd5-d6 Kg2-h2 Qd1-f3 Bc4-b5 Qf3-f5 Kh2-g2 Qf5-h7 Kg2-g1 Qh7-g6 Bb5-c4 Kd6-c6 Kg1-f1 Qg6-h7 Kf1-g1 Qh7-f5 Kg1-h2 Qf5-f8
26 01:27 73,020k 833k -3.66 Re1xe4 f5xe4 Bg5-e3 a6-a5 Kg1-g2 Kb7-b6 Kg2-g1 Qd6-d1+ Kg1-g2 Qd1-c2 Kg2-g1 Qc2-b1+ Kg1-h2 Qb1-e1 Bc4-b5 Qe1-d1 Bb5-c4 Kb6-c6 Bc4-b5+ Kc6-d5 Bb5-c4+ Kd5-d6 Be3-g5 Qd1-e1 Kh2-g2 Kd6-c6 Bg5-e3 Qe1-d1 Bc4-b5+ Kc6-d5 Bb5-c4+ Kd5-d6 Kg2-h2 Qd1-f3 Bc4-b5
27 01:31 76,228k 833k -3.66 Re1xe4 f5xe4 Bg5-e3 Kb7-c6 Kg1-g2 a6-a5 Kg2-g1 Qd6-d1+ Kg1-h2 Kc6-d6 Kh2-g2 Qd1-e1 Bc4-b5 Qe1-b1 Bb5-c4 Kd6-c6 Kg2-h2 Kc6-b6 Kh2-g2 Qb1-d1 Kg2-h2 Qd1-f3 Kh2-g1 Kb6-c6 Bc4-b5+ Kc6-d6 Kg1-h2 Qf3-d1 Bb5-c4 Qd1-e1 Kh2-g2 Qe1-d1
28- 01:35 79,530k 833k -3.73 Re1xe4 f5xe4
28+ 01:41 84,363k 834k -3.66 Re1xe4
28+ 01:44 87,414k 838k -3.56 Re1xe4
28- 01:44 87,682k 838k -3.64 Re1xe4 f5xe4
28 01:45 88,713k 839k -3.61 Re1xe4 f5xe4 Bg5-e3 Kb7-c6 Kg1-g2 a6-a5 Bc4-b5+ Kc6-b6 Bb5-e2 Qd6-d7 Be2-c4 Qd7-d1 Kg2-h2 Qd1-e1 Kh2-g2 Kb6-c6 Bc4-b5+ Kc6-d6 Bb5-c4 Qe1-d1 Kg2-h2 Kd6-c6 Bc4-b5+ Kc6-d5 Bb5-c4+ Kd5-d6 Be3-g5 Kd6-c7 Bg5-e3 Kc7-b6 Be3-g5 Qd1-f3 Bg5-e3 Kb6-c6
29 01:46 89,817k 840k -3.61 Re1xe4 f5xe4 Bg5-e3 Kb7-c6 Kg1-g2 a6-a5 Bc4-b5+ Kc6-b6 Bb5-e2 Qd6-d7 Be2-c4 Qd7-d1 Kg2-h2 Qd1-e1 Kh2-g2 Kb6-c6 Bc4-b5+ Kc6-d6 Bb5-c4 Qe1-d1 Kg2-h2 Qd1-f3 Kh2-g1 Kd6-c6 Kg1-h2 Qf3-d1 Kh2-g2 Qd1-b1 Kg2-h2 Qb1-e1 Bc4-b5+ Kc6-d6 Bb5-c4 Qe1-d1
30- 01:57 98,446k 839k -3.68 Re1xe4 f5xe4

[reflectionofpower]

Zappa Mexico II x64 -- Dissident Aggressor settings finds it instantly. running at 4 cores.

[Harvey Williamson]

The Baron found the win in this position. When Hiarcs played e5 a move most other engines play including komodo it is a draw. ...Bd5 I think wins

[Eelco de Groot]

That seems a very good find from The Baron!

I infer this was the position in the game?
[d]8/1k6/p2qp3/2p2pB1/P1B1b1p1/1P4P1/5P2/4R1K1 b - -

then after not playing e5 but Bd5:

[d]8/1k6/p2qp3/2pb1pB1/P1B3p1/1P4P1/5P2/4R1K1 w - -

Engine: Sf20150607_015 MOD MP (512 MB)
by Tord Romstad, Marco Costalba and Joona

38 120:47 -4.60 2.Bf4 Qd7 3.Rd1 Qc6 4.Rc1 Qb6 5.Bd3 Kc6
6.Be3 Qxb3 7.Bxa6 Qxa4 8.Bd3 c4
9.Be2 Qb4 10.Kf1 Kd6 11.Kg1 c3
12.Bd3 Be4 13.Bxe4 fxe4 14.Rd1+ Kc6
15.Rd4 (27.649.299.119) 3815

38 120:47 -5.15 2.a5 Be4 3.Kf1 Bf3 4.Rxe6 Qd1+ 5.Re1 Bg2+
6.Kxg2 Qxe1 7.Bd8 f4 8.gxf4 Qe4+
9.Kf1 Qxf4 10.Ke2 Qe5+ 11.Kf1 Qd4
12.Bg5 Qd1+ 13.Kg2 Qf3+ 14.Kg1 Kc6
15.Be3 (27.649.299.119) 3815

38 120:47 -5.15 2.Bf1 Qb6 3.Re3 Bxb3 4.a5 Qb4 5.Bg2+ Kc8
6.Bf6 Bd5 7.Bc3 Qb1+ 8.Bf1 Kd7 9.Re1 Qc2
10.Re3 Qc1 11.Be1 c4 12.Bc3 Kc6
13.Bb4 Kb5 14.Be1 Qa1 15.Bc3 (27.649.299.119) 3815

best move: Bg5-f4 time: 142:33.113 min n/s: 3.815.270 nodes: 32.698.656.300

[zullil]

In the first round of the World Computer Chess Championship Gingko found an interesting exchange sac to build a fortress.
[d]
42.Rxe4! does the job. After over 40 moves later Hiarcs agreed to a draw.

It's very interesting to see that all engines act very differently with this position. There are a couple of engines, which see the sac instantly:
- BlackMamba 2.0
- Fire 3.0 (not Fire 4.0)
- Firenzina 2.3.2
- Firenzina 2.4.1
- Maverick 1.0
- PanChess 00.611
- Rodent 1.7

Some engines don't find it immediately, but within 10 seconds:
- Matacz 1.4
- Stockfish 250714MZ++
- Stockfish 010614 Human (own setting - the default version doesn't find this in 10 seconds)

Stockfish 130615 takes 41 seconds, Komodo 6, Gull 3 and Houdini 3 don't find it in a minute.

What says your engine of choice?

A review of the game shows that White had the same sac available several moves earlier but didn't play it. More proof that neither engine had a clue about White constructing a fortress.

Here is the position after Black's 38th move:
[d]2k5/8/p2qp3/2p2p2/P1B1b1p1/1P4P1/5P1K/2B1R3 w - - 0 39

Here's the full game:
[pgn]
[Event "WCCC 2015"]
[Site "Leiden, The Netherlands"]
[Date "2015.06.29"]
[Round "1.2"]
[White "Ginkgo"]
[Black "Hiarcs"]
[Result "1/2-1/2"]
[ECO "D43"]
[LiveChessVersion "1.4.6"]

1. d4 Nf6 2. c4 c6 3. Nc3 d5 4. Nf3 e6 5. Bg5 h6 6. Bh4 dxc4 7. e4 g5 8. Bg3 b5 9. Be2 Bb7 10. O-O Nbd7 11. Ne5 Bg7 12. Nxd7 Nxd7 13. Bd6 a6 14. a4 b4 15. Bxb4 Qb6 16. Ba3 Qxd4 17. Qc2 c5 18. Rad1 Qe5 19. Bxc4 Qc7 20. b3 Be5 21. h3 Rg8 22. Ne2 h5 23. Qd3 g4 24. h4 Qc6 25. Ng3 O-O-O 26. Qc2 f5 27. Bc1 Rdf8 28. Nxh5 Rh8 29. Nf4 Rxh4 30. Ng6 Bh2+ 31. Kh1 Rh7 32. Nxf8 Nxf8 33. Rfe1 Qe8 34. Qd3 Nd7 35. Qxd7+ Rxd7 36. Rxd7 Qxd7 37. Kxh2 Qd6+ 38. g3 Bxe4 39. Bg5 Kb7 40. Kg1 e5 41. Rxe4 fxe4 42. Kh1 Kb6 43. Kh2 Ka5 44. Be3 Kb4 45. Kg2 Qb6 46. Bd2+ Ka3 47. Be3 Kb2 48. Bd5 Qc7 49. Bc4 Kc2 50. Kf1 Kc3 51. Kg2 Qb6 52. Bd5 Kb4 53. Bc1 Qb8 54. Bd2+ Ka3 55. Be3 Qd6 56. Bg8 Qf8 57. Be6 Qe7 58. Bg8 Qf8 59. Be6 Qd6 60. Bg8 Qb6 61. Bc4 Qc6 62. Bc1+ Ka2 63. Be3 Qd6 64. Bg5 Qb6 65. Be3 Qc6 66. Bf7 Qb7 67. Bg8 Qa8 68. Bc4 Qc8 69. Bf7 Qc7 70. Bc4 Qb7 71. Bg8 Qc8 72. Bf7 Qf8 73. Be6 Qd6 74. Bc4 Kb2 75. Kf1 Qe7 76. Kg1 Qf8 77. Kg2 Qa8 78. Kg1 Qa7 79. Kg2 Kc2 80. Bd5 Qe7 81. Bg8 a5 82. Bc4 Kc3 83. Bd5 1/2-1/2[/pgn]

[zullil]

The Baron found the win in this position. When Hiarcs played e5 a move most other engines play including komodo it is a draw. ...Bd5 I think wins

Gee, Stockfish chooses Bd5 almost instantly. Are you sure Komodo does not?

Code: Select all

Stockfish 290615 64 POPCNT by Tord Romstad, Marco Costalba and Joona Kiiski
position fen 8/1k6/p2qp3/2p2pB1/P1B1b1p1/1P4P1/5P2/4R1K1 b - - 3 40
d

 +---+---+---+---+---+---+---+---+
 |   |   |   |   |   |   |   |   |
 +---+---+---+---+---+---+---+---+
 |   | k |   |   |   |   |   |   |
 +---+---+---+---+---+---+---+---+
 | p |   |   | q | p |   |   |   |
 +---+---+---+---+---+---+---+---+
 |   |   | p |   |   | p | B |   |
 +---+---+---+---+---+---+---+---+
 | P |   | B |   | b |   | p |   |
 +---+---+---+---+---+---+---+---+
 |   | P |   |   |   |   | P |   |
 +---+---+---+---+---+---+---+---+
 |   |   |   |   |   | P |   |   |
 +---+---+---+---+---+---+---+---+
 |   |   |   |   | R |   | K |   |
 +---+---+---+---+---+---+---+---+

Fen: 8/1k6/p2qp3/2p2pB1/P1B1b1p1/1P4P1/5P2/4R1K1 b - - 3 40
Key: D74B512474A1793B
Checkers: 
go infinite
info depth 1 seldepth 1 multipv 1 score cp 187 nodes 63 nps 2333 tbhits 0 time 27 pv e6e5
info depth 2 seldepth 2 multipv 1 score cp 302 nodes 141 nps 3439 tbhits 0 time 41 pv e6e5 e1e4 f5e4
info depth 3 seldepth 4 multipv 1 score cp 228 nodes 355 nps 4034 tbhits 0 time 88 pv e4d5 g1h2 d5c4
info depth 4 seldepth 5 multipv 1 score cp 244 nodes 510 nps 4396 tbhits 0 time 116 pv e4d5 a4a5 e6e5 g1h2
info depth 5 seldepth 6 multipv 1 score cp 227 nodes 1895 nps 7260 tbhits 0 time 261 pv d6d7 g5f6 e4d5 a4a5 d5c4
info depth 6 seldepth 6 multipv 1 score cp 227 nodes 2223 nps 7613 tbhits 0 time 292 pv d6d7 g5f6 e4d5 a4a5 d5c4 b3c4
info depth 7 seldepth 8 multipv 1 score cp 250 nodes 5427 nps 7992 tbhits 0 time 679 pv d6d7 g5f6 d7d2 e1e4 f5e4 c4e6 d2d1 g1h2
info depth 8 seldepth 10 multipv 1 score cp 220 nodes 11376 nps 10370 hashfull 0 tbhits 0 time 1097 pv d6d7 g5e3 d7h7 g1f1 e4d5 c4d5 e6d5 e3c5 h7h1 f1e2
info depth 9 seldepth 12 multipv 1 score cp 248 nodes 35729 nps 22014 hashfull 0 tbhits 0 time 1623 pv e4d5 g5f4 d6b6 f4d2 d5c4 b3c4 b6c6 e1b1 b7c7 d2f4 c7d7
info depth 10 seldepth 16 multipv 1 score cp 253 nodes 54860 nps 31766 hashfull 0 tbhits 0 time 1727 pv e4d5 g5f4 d6d7 a4a5 b7c6 c4a6 d7h7 g1f1 h7h1 f1e2 h1e4 e2f1 e4g2 f1e2 d5b3

[Eelco de Groot]

May be a question of some luck. Not with Komodo, but in the first run I see nothing but e5, and even after switching to three moves Multi PV it does not see it. The next run I tried, now with three moves Bd5 is quickly on top, depth 26 after 28 seconds. But there is no 40...e5 anymore in the first three moves.