test position of positional understanding

[yanquis1972]

why not just bxc6 after c6??

[Eelco de Groot]

then why did black win and where is whites error?

I think White just tried too much, but I can't really analyze it now because this position below is still running for testing purposes. You'd have to ask Lukas for a run with the Cluster if he can confirm that Rybka also does not see anything. In the position that I took as an example position, it already seems that both White and Black should go for a repetition of moves. Such a thing (prediction of move repetition by a program) can be inaccurate if the evaluation just does not see any advantage for Black, then it will just keep shuffling the Bishop from d7 to e8, but I do have some positional eval of the little mobility of the White Rook and Bishop. It only gives a very slight push I think to try to free them or give some indication of an advantage, but a repetition of moves is the only result. In the game however that did not happen, I suspect White just tried too hard to win when that seems a bit foolish. I have no idea how strong the players are?

[D]1r2brk1/2q1b1pp/p2p1p2/n3p3/P1pPPP2/1pP1B2Q/1P1N2PP/RB3R1K b - -

[yanquis1972]

some monte carlo analysis of the game position after black's f6 yielded results of about 45% (200 games at depth 11 & 2000 at depth 9), so i think you are correct. it would seem correct evaluation would be about what top engines have it at, which would be = with a very slight advantage to black (maybe the advantage could be a little bigger based on MC but i'm not clear on how to interpret results; at any rate draw is easily the most likely result, with black having an easier time winning)

[michiguel]

why not just bxc6 after c6??

White bishop become better, d5 will be a good post for both B and N, and the b and a pawns will roll forward. White will dominate the center completely.

Miguel

[Eelco de Groot]

some monte carlo analysis of the game position after black's f6 yielded results of about 45% (200 games at depth 11 & 2000 at depth 9), so i think you are correct. it would seem correct evaluation would be about what top engines have it at, which would be = with a very slight advantage to black (maybe the advantage could be a little bigger based on MC but i'm not clear on how to interpret results; at any rate draw is easily the most likely result, with black having an easier time winning)

Thanks for trying with Monte Carlo John! I just stopped my analysis, because I got no further iterations after 2914 min 28.718 sec. After entering 21... Bd7 in this side variation I only get apparent move repetitions.

[Event "European TC VII, Final"]
[Site "ICCF"]
[Date "2008.12.31"]
[Round "-"]
[White "Schneider, Wolfgang"]
[Black "Eiben, Mgr. Pavel"]
[WhiteElo "2532"]
[BlackElo "2505"]
[Result "0-1"]

1. e4 e5 2. Nf3 Nc6 3. Bb5 a6 4. Ba4 Nf6 5. O-O Be7 6. Nc3
b5 7. Bb3 d6 8. a4 b4 9. Nd5 Na5 10. Ba2 Nxd5 11. Bxd5 c6
12. Ba2 c5 13. c3 Rb8 14. d4 b3 15. Bb1 Qc7 16. Be3 c4
17. Nd2 O-O 18. f4 f6 (18... Bd7 19. Kh1 f6 20. Qh5 Be8
21. Qh3 {= according to Rainbow Serpent,
http://rybkaforum.net/cgi-bin/rybkaforu ... ?tid=13584}
21... Bd7 * {Some analysis of this postion below})
19. d5 Bd7 20. f5 Be8 21. Qg4 Bd8 22. Rf3 Qf7
23. Rh3 Bb6 24. Nf1 Qc7 25. Qh4 h6 26. Kh1 Bxe3 27. Nxe3
Kf7 28. Rg3 Rg8 29. Qg4 Ke7 30. Rh3 h5 31. Qe2 Rh8 32. Rh4
Bf7 33. Nf1 Rbg8 34. Ng3 Rh6 35. Qf2 Be8 36. Qe3 Kd8
37. Nf1 Kc8 38. Qf2 Kb7 39. Ne3 Qc5 40. Qe1 Kb6 41. Qe2 g5
42. fxg6 Rhxg6 43. Qd2 Rg5 44. Qe1 Bd7 45. g3 Nb7 46. Qf2
Ka7 47. Kg2 Kb8 48. Qe2 Bg4 49. Qe1 a5 50. Qf2 Nd8 51. Nf5
Qxf2+ 52. Kxf2 Nb7 53. Ra3 Ka7 54. Ne3 Rc8 55. Nf1 Kb6
56. Nd2 Nd8 57. h3 Bd7 58. g4 Be8 59. Nf3 Rg7 60. Nh2 hxg4
61. Rxg4 Rcc7 62. Rxg7 Rxg7 63. Nf3 Nf7 64. h4 Rg4 65. Nd2
Kc5 0-1


[D]1r3rk1/2qbb1pp/p2p1p2/n3p3/P1pPPP2/1pP1B2Q/1P1N2PP/RB3R1K w - -

Engine: Rainbow Serpent 1.5 Build 031 (Athlon 2009 MHz, 256 MB)
by Romstad, Costalba, Kiiski, De Groot

1.00 0:00 0.00 22.Qh5 (535)

2.00 0:00 0.00 22.Qh5 (536)

3.00 0:00 0.00 22.Qh5 (748)

4.00 0:00 0.00 22.Qh5 (797)

5.00 0:00 0.00 22.Qh5 (3.004)

6.00 0:00 0.00 22.Qh5 (3.081)

7.00 0:00 0.00 22.Qh5 (3.282)

8.00 0:00 0.00 22.Qh5 (142.992) 435

9.00 0:00 0.00 22.Qh5 (143.832) 438

10.01 0:06 0.00 22.Qh5 (2.866.399) 456

11.01 0:06 0.00 22.Qh5 (2.867.630) 455

12.01 0:06 0.00 22.Qh5 (2.872.021) 455

13.01 0:06 0.00 22.Qh5 (2.882.938) 454

14.01 0:06 0.00 22.Qh5 (2.902.583) 454

15.01 0:06 0.00 22.Qh5 (2.966.514) 453

16.01 0:06 0.00 22.Qh5 (2.966.553) 453

17.01 0:07 0.00 22.Qh5 (3.280.196) 455

18.01 0:08 0.00 22.Qh5 (3.679.132) 455

19.01 0:09 0.00 22.Qh5 (4.432.298) 459

20.01 0:11 0.00 22.Qh5 (5.469.889) 462

21.01 0:20 0.00 22.Qh5 (9.420.065) 470

22.01 1:06 0.00 22.Qh5 (31.238.385) 468

best move: Qh3-h5 time: 3:41.485 min n/s: 478.817 nodes: 106.020.233


Maybe we are just not seeing or the computer is not seeing the way for Black to activate its pieces, I don't know. My goal was just to test what the computer sees here, maybe later I can take another look. Build 31 is apart from this position untested, so this is not really a good reference, apart from your Monte Carlo testing It still would be interesting to see what the Cluster says about it.

Va Banque Regards,
Eelco

[playjunior]

The position seems incredible. We need a GM analysis.

[Howard E]

Thanks for posting this position. Black has so many options. Most every piece exchange favours black. Also, after black castles king side he can if feasible march the king back to the queenside. Then place the rooks on the kingside. Virtually two pieces up.

[Kirk]

how much time does your favourite program need to see that white is losing or even only to see only a significant advantage for black

[D]1rb1k2r/2q1bppp/p2p4/n3p3/P1pPP3/1pP1BN2/1P3PPP/RB1Q1RK1 w k - 0 17

This position if from the following game

http://www.iccf-webchess.com/MakeAMove.aspx?id=204281

White rook at a1 cannot move because the bishop at b1 cannot move so I believe that white is losing.

Uri

Only engine I tried so far that shows a small advantage for black has been Smarthink


Analysis by SmarThink 1.00:

1. =/+ (-0.57): 1...0-0 2.Qd2 Be6 3.Rd1 Rbd8 4.d5 Bd7 5.Qe2
2. =/+ (-0.51): 1...Bf6 2.Qd2 Be6 3.Rd1 Rd8 4.d5 Bd7 5.Qe2
3. =/+ (-0.40): 1...Bd7 2.Qd2
4. =/+ (-0.30): 1...h6 2.Qd2 Be6 3.Rd1 Rd8 4.d5 Bd7 5.Qe2
5. = (-0.25): 1...Be6 2.Qd2

Depth 14/51