Not logged inRybka Chess Community Forum
Up Topic The Rybka Lounge / Computer Chess / Is my game proof of human limitations of foresight?
- - By rocket (****) Date 2019-08-15 21:36
I cannot envision how g4 wins, instead of going for the repetition.

I can evaluate based on my experience and positional understanding that it looks completely winning, but I could not for the life of me see a concrete sequence of moves. All engines can.

[Event "120'/40"]
[Site "?"]
[Date "????.??.??"]
[Round "?"]
[White "me"]
[Black "Deep Fritz 10.1 2CPU"]
[Result "1/2-1/2"]
[PlyCount "74"]
[TimeControl "40/7200:0/0:0/0"]

304M♗, Fritz11.ctg, DES♔TO♙-♘DU6FUC 1. e4 e5 2. Nf3 Nc6 3. Bb5 Nf6 4. d3 Bc5 5. O-O d6 170 6. Nbd2 55 O-O 173 7. Re1 34 Ng4 282 8. Re2 20 f5 176 9. h3 294 Nxf2 376 10. Rxf2 53 Bxf2+ 140 11. Kxf2 5 fxe4 333 12. dxe4 94 Be6 170 13. Kg1 194 Qf6 182 14. c3 133 a6 0 15. Bc4 75 Nd8 120 16. Qe2 22 Kh8 220 17. Bxe6 25 Nxe6 54 18. Nf1 50 Qg6 184 19. Nh4 75 Qf7 133 20. Be3 60 Nf4 145 21. Bxf4 46 Qxf4 69 22. Nf3 23 Rad8 268 23. Rd1 6 h6 456 24. Qe3 73 Qf7 194 25. b3 22 Qg6 439 26. Kh2 80 Qe6 262 27. Ng3 118 b6 309 28. Qd3 216 a5 168 29. c4 84 Ra8 90 30. a4 1268 Rae8 250 31. Nf5 40 Qf6 259 32. Qd2 28 Re6 341 33. Qe3 47 Qg6 140 34. N3h4 368 Qh7 177 35. Nf3 90 Qg6 207 36. N3h4 7 Qh7 134 37. Nf3 6 Qg6 248 1/2-1/2
Parent - - By Labyrinth (*****) Date 2019-08-15 23:08
Well it's not the only win, g4 doesn't even feature in the engine line a move later:

34. N3h4 Qh7 35. Kg1 Rff6 36. Qg3 Re8 37. Qg4 g6 38. Ne3 Rg8 39. Nd5 Rff8 40. Rd3 Qg7 41. Rg3 Kh7 42. Nf3 +1.70

In this case the knight settled on d5, rook on g3, queen on g4 and the other knight on f3. This is clearly an improvement to the original position after 34..Qh7. There isn't really a need for a concrete line here, just steady improvements.

5r1k/2p3pq/1p1pr2p/p3pN2/P1P1P3/1P2QN1P/6PK/3R4 b - - 0 35


5rr1/2p3qk/1p1p2pp/p2Np3/P1P1P1Q1/1P3NRP/6P1/6K1 b - - 0 42
Parent - - By rocket (****) Date 2019-08-17 17:07
Your final position is indeed winning?
Parent - By Labyrinth (*****) Date 2019-08-17 22:58
I think so. Here is a sample engine game:

5rr1/2p3qk/1p1p2pp/p2Np3/P1P1P1Q1/1P3NRP/6P1/6K1 b - - 0 42


42...c6 43.Nxb6 Rf4 44.Qe6 Rf6 45.Qd7 Qxd7 46.Nxd7 Rf7 47.Nb6 c5 48.Nh4 Re7 49.Rd3 Re6 50.Nd5 Kg7 51.Nc3 Rc8 52.Nb5 Rd8 53.Nf3 Re7 54.h4 Kf6 55.h5 gxh5 56.Rd5 Re6 57.Nh4 Kg5 58.Nf5 h4 59.Nbxd6 Rb8 60.Nb5 Kf4 61.Nxh4 Kxe4 62.Rxc5 Rg8 63.Nf3 Kf4 64.Nc3 Ke3 65.Rd5 h5 66.Nd1+ Kf4 67.Nf2 Re7 68.Nd3+ Ke3 69.Ndxe5 Reg7 70.Rd2 h4 71.Kh2 Rb7 72.Rd3+ Kf2 73.Nxh4 Re8 74.Rf3+ Ke2 75.Nc6 Rh7 76.Nd4+ Kd2 77.Rh3 Rg7 78.c5 Rg4 79.Ndf5 Kc1 80.c6 Kb2 81.Nf3 Reg8 82.Ne7 Rxg2+ 83.Kh1 Kxb3 84.c7 Kxa4 85.c8Q Rxc8 86.Nxc8 Rg8 87.Ne7 Re8 88.Nc6 Kb5 89.Nfd4+ Kb6 90.Rb3+ Kc7 91.Nxa5 White wins per tablebase (Black loses in at most 23 moves with perfect play from both sides)

Anyways, only white can be better, and white can continue to press with little risk, so should do so if game time allows.
Up Topic The Rybka Lounge / Computer Chess / Is my game proof of human limitations of foresight?

Powered by mwForum 2.27.4 © 1999-2012 Markus Wichitill