Man vs machine - how to beat chess engine

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

Moderator: Ras

overlord
Posts: 198
Joined: Sun Jun 03, 2012 6:46 pm
Location: Trinec, Czech Republic

Re: Success against chess engine was never so easy :)

Post by overlord »

Compared with other top engines, Stockfish is very sensitive to blunders (especially at lower depths). For example, in free style games it is valuable to use Stockfish just a standing engine (for deep depths) not for analysis (as an analytical tool is it useless - with one exception - endgames).
mar
Posts: 2668
Joined: Fri Nov 26, 2010 2:00 pm
Location: Czech Republic
Full name: Martin Sedlak

Re: Success against chess engine was never so easy :)

Post by mar »

overlord wrote:Just check with Stockfish...at depth 14 it wants to play Nxf3 :)
Yes but it reconsiders in no time and that's my point - it's obvious that SF moved very fast and lost.
I doubt it was a ponder hit so maybe easy move (which I call easy blunder - elo hunters won't agree of course), but even so you claim to have played a 5 min game. And we're talking about move 32 so I guess SF still had plenty of time left.
I don't know what was your setup, whether SF played with book or not, how much time the engine spent per move, what were the clocks and what depth it reached.
It simply seems quite odd to me.
S.Taylor
Posts: 8514
Joined: Thu Mar 09, 2006 3:25 am
Location: Jerusalem Israel

Re: Success against chess engine was never so easy :)

Post by S.Taylor »

Cubeman wrote:[pgn]
[Event "Blitz:5'"]
[Site "Microsoft"]
[Date "2013.06.11"]
[Round "?"]
[White "Kvicala, Miroslav"]
[Black "Stockfish 3"]
[Result "1-0"]
[ECO "C02"]
[PlyCount "75"]

1. e4 e6 2. d4 d5 3. e5 c5 4. c3 Nc6 5. Nf3 Qb6 6. a3 c4 7. Qe2 Bd7 8. Nbd2 Be7 9. g3 Na5 10. Bg2 Nh6 11. O-O O-O 12. h3 Ba4 13. Nh2 Rfc8 14. Ng4 Nxg4 15. Qxg4 Bc2 16. f4 Qb5 17. h4 Bd3 18. Rf2 Qa4 19. Qf3 Bf5 20. Kh2 Rf8 21. g4 Bc2 22. g5 Nc6 23. Bh3 g6 24. Bg4 b5 25. h5 gxh5 26. Bxh5 b4 27. Bg4 Bd3 28. f5 bxc3 29. f6 Bc5 30. Qh3 Nxe5 31. Bf3 Bd6 32. Kg2 Nxf3 33. Qh6 Nh4+ 34. Kh3 Nf5 35. Rxf5 Bxf5+ 36. Kh4 Bg3+ 37. Kxg3 c2 38. Qg7# 1-0
[/pgn]
Liked this game, you played well, this line of French Advance looks dangerous for black to castle king side, and black missed chance to breakout with an earlier f6/f5.
I didn't think that todays top machines would allow such quick mates anymore. After 29.f6 i believe it is obvious, to any human eye, but various things before then, i would have thought that a todays top engine would have thwarted, unless it was moving instantly.
User avatar
Ajedrecista
Posts: 2163
Joined: Wed Jul 13, 2011 9:04 pm
Location: Madrid, Spain.

Re: Success against chess engine was never so easy :)

Post by Ajedrecista »

Hello Martin:
mar wrote:
Cubeman wrote: Liked this game, you played well, this line of French Advance looks dangerous for black to castle king side, and black missed chance to breakout with an earlier f6/f5.
I doubt StockFish would move into mate by commiting the blunder 32 ... Nxf3?? Something smells here and it's not StockFish :)
First of all, sorry for this long post. I tried this position:

[d]r4rk1/p4p1p/3bpP2/3pn1P1/q1pP4/P1pb1B1Q/1P1N1RK1/R1B5 b - - 3 32

With Multi PV on (Multi PV = 45 and Multi PV = 6) and surprisingly 32.- ..., Nxf3 was SF top choice until depth 15 or so, thinking that the best reply of white side is 33.- Nxf3 instead of 33.- Qh6 and mate in few moves. Here is an infinite analysis with Multi PV off:

Analysis by SF 3 w32 (using two cores) in an Intel Pentium D930 (3 GHz), with 1 GB of hash, running under Arena 2.0.1 GUI:
FEN: r4rk1/p4p1p/3bpP2/3pn1P1/q1pP4/P1pb1B1Q/1P1N1RK1/R1B5 b - - 3 32

Stockfish-3-32-ja:
1/2 00:00 280 280.000 +6,82 c3xd2 Bc1xd2 Ne5xf3 Rf2xf3
2/3 00:00 442 442.000 +6,82 c3xd2 Bc1xd2 Ne5xf3 Rf2xf3
3/4 00:00 2.172 2.172.000 +6,82 c3xd2 Bc1xd2 Ne5xf3 Rf2xf3
4/8 00:00 9.668 568.705 +5,49 Ne5g6 Nd2b1 c3xb2 Rf2xb2 Rf8b8
5/8 00:00 11.795 368.593 +5,69 Ne5xf3 Nd2xf3 Qa4d1 b2xc3 Bd3f5
6/8 00:00 18.610 387.708 +3,63 Ne5xf3 Nd2xf3 Bd3e4 Qh3h6 Be4xf3+ Rf2xf3 Qa4c2+ Rf3f2
7/15 00:00 37.412 393.810 +3,39 Ne5xf3 Nd2xf3 Rf8c8 b2xc3 Bd3f5 Qh3h4 Qa4d1 Nf3e5 Bf5e4+ Kg2h3 Bd6xe5 d4xe5
8/15 00:00 70.134 405.398 +3,91 Ne5xf3 Nd2xf3 Rf8b8 b2xc3 Qa4d1 Bc1b2 Bd3f1+ Rf2xf1 Rb8xb2+ Kg2h1 Qd1e2
9/18 00:00 100.359 456.177 +3,35 Ne5xf3 Nd2xf3 Rf8b8 b2xc3 Qa4d1 Qh3h6 Bd6f8 Qh6h1 Qd1xh1+ Kg2xh1 Rb8b3 Bc1d2 Bf8d6
10/19 00:00 120.440 451.086 +3,55 Ne5xf3 Nd2xf3 Rf8b8 b2xc3 Qa4d1 Qh3h6 Bd6f8 Qh6h1 Qd1xh1+ Kg2xh1 Rb8b3 Bc1d2 Rb3xa3 Ra1xa3 Bf8xa3
11/20 00:00 206.170 525.943 +3,19 Ne5xf3 Nd2xf3 Rf8b8 b2xc3 Bd3e4 Qh3h6 Bd6f8 Qh6h1 Qa4a5 Bc1d2 Rb8b3 Kg2g1 Qa5c7 Qh1h3 Bf8xa3
12/20 00:00 272.194 543.301 +3,19 Ne5xf3 Nd2xf3 Rf8b8 b2xc3 Bd3e4 Qh3h6 Bd6f8 Qh6h1 Qa4a5 Bc1d2 Rb8b3 Kg2g1 Qa5c7 Qh1h3 Bf8xa3
13/20 00:00 352.926 549.728 +3,51 Ne5xf3 Nd2xf3 Rf8b8 b2xc3 Qa4d1 Qh3h6 Bd6f8 Qh6h3 Bd3e4 Bc1b2 Qd1d3 g5g6 Be4xg6 Ra1h1 Bg6e4 Qh3g3+ Kg8h8 a3a4
14/26 00:01 657.861 600.786 +3,39 Ne5xf3 Nd2xf3 Rf8b8 b2xc3 Qa4d1 Qh3h6 Bd6f8 Qh6h3 Bd3e4 Bc1b2 Qd1d3 g5g6 Be4xg6 Nf3e5 Bg6e4+ Kg2h2 Qd3xh3+ Kh2xh3 Bf8d6 Ra1g1+ Kg8f8 Ne5d7+
15/27 00:01 887.434 610.339 +3,35 Ne5xf3 Nd2xf3 Rf8b8 b2xc3 Qa4d1 Qh3h6 Bd6f8 Qh6h3 Rb8b3 Bc1b2 Bd3f1+ Rf2xf1 Rb3xb2+ Kg2g1 Qd1d3 Rf1f2 Qd3xc3 Rf2xb2 Qc3xb2
16/27 00:03 2.486.998 674.165 +2,62 Kg8h8 d4xe5 c3xd2 e5xd6 d2d1Q Bf3xd1 Qa4xd1 Kg2h2 Ra8d8 Qh3f3 Qd1xf3 Rf2xf3 Rd8xd6 Bc1f4 Rd6c6 Rf3f2 d5d4 Kh2g3 Rc6c5 Rf2h2 e6e5 Bf4d2

17/28+ 00:04 2.709.664 674.549 +2,70 Kg8h8 d4xe5 c3xd2 e5xd6 d2d1Q Bf3xd1 Qa4xd1 Kg2h2 Ra8d8 Qh3f3 Qd1xf3 Rf2xf3 Rd8xd6 Bc1f4 Rd6c6 Rf3f2 d5d4 Kh2g3 Rc6c5 Rf2h2 e6e5 Bf4d2 Rf8g8
17/28+ 00:04 2.814.958 679.613 +2,78 Kg8h8 d4xe5 c3xd2 e5xd6 d2d1Q Bf3xd1 Qa4xd1 Kg2h2 Ra8d8 Qh3f3 Qd1xf3 Rf2xf3 Rd8xd6 Bc1f4 Rd6b6 Rf3f2 Rf8g8 Bf4e3 Rb6b3 a3a4 Bd3e4
17/30+ 00:04 2.992.928 679.130 +2,90 Kg8h8 d4xe5 c3xd2 e5xd6 d2d1Q Bf3xd1 Qa4xd1 Kg2h2 Ra8d8 Qh3f3 Qd1xf3 Rf2xf3 Rd8xd6 Bc1f4 Rd6b6 Rf3f2 Bd3e4 Bf4e3 Rb6b3 Be3xa7 Rf8g8
17/30 00:06 4.178.785 682.139 +2,90 Kg8h8 b2xc3 Ne5xf3 Nd2xf3 Bd3f5 Qh3h4 Qa4d1 Bc1b2 Qd1d3 Rf2d2 Qd3e4 Qh4xe4 Bf5xe4 a3a4 Bd6f4 Rd2e2 Rf8g8 Kg2f2 Ra8b8 Bb2a3 Rb8b3
18/30 00:08 5.450.501 686.547 +2,86 Kg8h8 b2xc3 Ne5xf3 Nd2xf3 Bd3f5 Qh3h4 Qa4d1 Bc1b2 Qd1d3 Rf2d2 Qd3e4 Qh4xe4 Bf5xe4 a3a4 Rf8g8 Kg2f2 Be4xf3 Kf2xf3 Rg8xg5 Rd2g2 Rg5f5+ Kf3e3 Rf5xf6
19/30 00:08 6.155.205 697.157 +2,86 Kg8h8 b2xc3 Ne5xf3 Nd2xf3 Bd3f5 Qh3h4 Qa4d1 Bc1b2 Qd1d3 Rf2d2 Qd3e4 Qh4xe4 Bf5xe4 a3a4 Rf8g8 Kg2f2 Be4xf3 Kf2xf3 Rg8xg5 Rd2g2 Rg5f5+ Kf3e3 Rf5xf6
20/30+ 00:09 6.794.941 694.637 +2,94 Kg8h8 b2xc3 Ne5xf3 Nd2xf3 Bd3f5 Qh3h4 Qa4d1 Bc1b2 Qd1d3 Rf2d2 Qd3e4 Qh4xe4 Bf5xe4 a3a4 Rf8g8 Kg2f2 Be4xf3 Kf2xf3 Rg8xg5 Rd2g2 Rg5f5+ Kf3e3 Rf5xf6
20/30+ 00:10 7.154.349 696.829 +3,03 Kg8h8 b2xc3 Ne5xf3 Nd2xf3 Bd3f5 Qh3h4 Qa4d1 Bc1b2 Qd1d3 Rf2d2 Qd3e4 Qh4xe4 Bf5xe4 a3a4 Rf8g8 Kg2f2 Be4xf3 Kf2xf3 Rg8xg5 Rd2g2 Rg5f5+ Kf3e3 Rf5xf6
20/31+ 00:13 9.473.771 692.881 +3,15 Kg8h8 b2xc3 Ne5g6 Kg2h1 Qa4a5 Bc1b2 Ra8b8 Nd2f1 Rb8xb2 Rf2xb2 Qa5xc3
20/40- 00:23 16.421.315 703.901 +2,78 Kg8h8 d4xe5 c3xd2 e5xd6 d2xc1Q Ra1xc1 Rf8g8 Qh3h6 Qa4d7 Bf3h5 d5d4 Rc1e1 Ra8f8
20/40 00:26 18.508.908 707.608 +3,31 Kg8h8 d4xe5 c3xd2 e5xd6 d2xc1Q Ra1xc1 Rf8g8 Kg2h1 Rg8xg5 Rf2g2 Rg5xg2 Qh3xg2 Ra8g8 Qg2h2 Bd3g6 Rc1g1 Qa4b3 Rg1g3 Qb3b6 Rg3g2 Qb6e3 Qh2g3 Rg8d8 d6d7 Qe3b6 Qg3h2 Rd8xd7 Qh2h6
21/40 00:28 19.973.751 713.705 +3,31 Kg8h8 d4xe5 c3xd2 e5xd6 d2xc1Q Ra1xc1 Rf8g8 Kg2h1 Rg8xg5 Rf2g2 Rg5xg2 Qh3xg2 Ra8g8 Qg2h2 Bd3g6 Rc1g1 Qa4b3 Rg1g3 Qb3b6 Rg3g2 Qb6e3 Qh2g3 Rg8d8 d6d7 Qe3b6 Qg3h2 Rd8xd7 Qh2h6
22/45- 00:33 24.214.893 726.541 +3,23 Kg8h8 d4xe5 c3xd2 e5xd6 d2xc1Q Ra1xc1 Rf8g8 Qh3h5 Rg8xg5+ Qh5xg5 Ra8g8 Qg5xg8+ Kh8xg8 Rf2d2 Qa4d7 Rd2xd3 c4xd3 Rc1c7 Qd7d8 Rc7xa7 Qd8b8 d6d7 Qb8xb2+ Kg2f1 Qb2b1+ Kf1f2 Qb1b6+ Kf2f1 Qb6d8 a3a4 Kg8f8 Kf1e1 h7h6 Ke1d2
22/45- 00:34 25.467.713 733.517 +3,15 Kg8h8 d4xe5 c3xd2 e5xd6 d2xc1Q Ra1xc1 Rf8g8 Qh3h5 Rg8xg5+ Qh5xg5 Ra8g8 Qg5xg8+ Kh8xg8 Rf2d2 Qa4d7 Rd2xd3 c4xd3 Rc1c7 Qd7d8 Rc7xa7 Qd8b8 d6d7 Qb8xb2+ Kg2f1 Qb2b1+ Kf1f2 Qb1b6+ Kf2f1 Qb6d8 a3a4 Kg8f8 Kf1e1 h7h6 Ke1d2
22/45- 00:36 26.845.784 739.920 +3,03 Kg8h8 d4xe5 c3xd2 e5xd6 d2xc1Q Ra1xc1 Rf8g8 Qh3h5 Rg8xg5+ Qh5xg5 Ra8g8 Qg5xg8+ Kh8xg8 Rf2d2 Qa4d7 Rd2xd3 c4xd3 Rc1c7 Qd7d8 Rc7xa7 Qd8b8 d6d7 Qb8xb2+ Kg2f1 Qb2b1+ Kf1f2 Qb1b6+ Kf2f1 Qb6d8 a3a4 Kg8f8 Kf1e1 h7h6 Ke1d2
22/45+ 00:49 36.233.196 735.445 +3,39 Kg8h8 d4xe5 Bd6xe5 Nd2b1 c3xb2 Bc1xb2 Be5xb2 Rf2xb2 Ra8b8 Rb2xb8 Qa4c2+ Kg2g1 Rf8xb8
22/45- 00:54 39.841.357 736.507 +2,84 Kg8h8 d4xe5 Bd6xe5 b2xc3 Be5xc3 Ra1a2 Bc3d4 Nd2f1 Bd4xf2 Ra2xf2 c4c3 Nf1e3 Ra8b8 Ne3g4 Bd3e4 Ng4e5 Be4xf3+ Qh3xf3 c3c2 Qf3e2 Rb8c8 Qe2d3 Qa4e4+ Qd3xe4 d5xe4 Kg2g3
22/45 00:56 41.418.783 734.688 +2,90 Kg8h8 d4xe5 Bd6xe5 b2xc3 Be5xc3 Ra1a2 Bc3d4 Nd2f1 Bd4xf2 Ra2xf2 c4c3 Nf1e3 c3c2 Qh3g4 Qa4b3 Qg4d4 Rf8c8 Rf2d2 Bd3b5 Rd2f2 a7a6 Bf3h5 Rc8c7 Bh5xf7 Rc7xf7 Rf2xc2 Qb3d3 Qd4xd3 Bb5xd3
23/45- 01:19 58.227.474 732.845 +2,82 Kg8h8 d4xe5 Bd6xe5 b2xc3 Be5xc3 Ra1a2 Bc3d4 Nd2f1 Bd4xf2 Ra2xf2 c4c3 Nf1e3 c3c2 Qh3g4 Qa4b3 g5g6
23/45+ 01:23 60.884.768 731.067 +2,98 Kg8h8 d4xe5 Bd6xe5 b2xc3 Be5xc3 Ra1a2 Bc3d4 Nd2f1 Bd4xf2 Ra2xf2 c4c3 Nf1e3 Rf8g8 Ne3g4 Bd3f5 Qh3h5 Bf5xg4 Bf3xg4 Ra8b8 Bg4d1 Qa4e4+ Kg2g1 Rb8b1
23/45 01:26 63.055.359 730.407 +3,03 Kg8h8 d4xe5 Bd6xe5 b2xc3 Be5xc3 Ra1a2 Bc3d4 Nd2f1 Bd4xf2 Ra2xf2 c4c3 Nf1e3 Rf8g8 Ne3g4 Bd3f5 Qh3h5 Bf5xg4 Bf3xg4 Ra8b8 Bg4d1 Qa4e4+ Bd1f3 Qe4b1 Rf2f1 Qb1c2+ Kg2g1 Qc2g6 Kg1f2 Qg6xh5 Bf3xh5
24/45 01:39 72.528.140 729.490 +3,03 Kg8h8 d4xe5 Bd6xe5 b2xc3 Be5xc3 Ra1a2 Bc3d4 Nd2f1 Bd4xf2 Ra2xf2 c4c3 Nf1e3 Rf8g8 Ne3g4 Bd3f5 Qh3h5 Bf5xg4 Bf3xg4 Ra8b8 Bg4d1 Qa4e4+ Bd1f3 Qe4b1 Rf2f1 Qb1c2+ Kg2g1 Qc2g6 Kg1f2 Qg6xh5 Bf3xh5
25/45 01:49 79.777.485 730.951 +3,03 Kg8h8 d4xe5 Bd6xe5 b2xc3 Be5xc3 Ra1a2 Bc3d4 Nd2f1 Bd4xf2 Ra2xf2 c4c3 Nf1e3 Rf8g8 Ne3g4 Bd3f5 Qh3h5 Bf5xg4 Bf3xg4 Ra8b8 Bg4d1 Qa4e4+ Bd1f3 Qe4b1 Rf2f1 Qb1c2+ Kg2g1 Qc2g6 Kg1f2 Qg6xh5 Bf3xh5
I stopped the analysis after two minutes. If SF did not run into depth, then the blunder is possible.

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

Once 32.- ..., Nxf3?? is played, here is the infinite analysis under the same conditions:

[d]r4rk1/p4p1p/3bpP2/3p2P1/q1pP4/P1pb1n1Q/1P1N1RK1/R1B5 w - - 0 33
FEN: r4rk1/p4p1p/3bpP2/3p2P1/q1pP4/P1pb1n1Q/1P1N1RK1/R1B5 w - - 0 33

Stockfish-3-32-ja:
1/1 00:00 182 182.000 -4,56 Nd2xf3 c3xb2 Bc1xb2
2/3 00:00 481 481.000 -4,88 Nd2xf3 Qa4d1 b2xc3
3/3 00:00 900 900.000 -4,88 Nd2xf3 Qa4d1 b2xc3
4/4 00:00 1.458 1.458.000 -5,33 Nd2xf3 Qa4d1 b2xc3 Bd3f5
5/9 00:00 5.978 373.625 -3,35 Nd2xf3 Bd3e4 Qh3h6 Be4xf3+ Rf2xf3 Qa4c2+ Rf3f2 c3xb2 Bc1xb2
6/9 00:00 8.656 541.000 -3,51 Nd2xf3 Rf8e8 b2xc3 Qa4d1 Bc1b2 Bd3b1
7/11 00:00 18.781 391.270 -3,23 Nd2xf3 Rf8c8 b2xc3 Bd3e4 g5g6 h7xg6 Qh3h6 Bd6f8
8/14 00:00 29.151 369.000 -3,19 Nd2xf3 Kg8h8 b2xc3 Qa4d1 Qh3h1 Qd1b3 Qh1h6 Rf8g8 Bc1d2 Bd3e4 Ra1h1 Qb3xa3
9/14 00:00 50.810 403.253 -3,19 Nd2xf3 Kg8h8 b2xc3 Qa4d1 Qh3h1 Qd1b3 Qh1h6 Rf8g8 Bc1d2 Bd3e4 Ra1h1 Qb3xa3 Kg2g1
10/14 00:00 102.312 544.212 -3,19 Nd2xf3 Rf8b8 b2xc3 Qa4d1 Qh3h6 Bd6f8 Qh6h1 Qd1xh1+ Kg2xh1 Rb8b3 Bc1b2
11/20 00:00 210.071 583.530 -2,70 Nd2xf3 Rf8b8 b2xc3 Qa4d1 Bc1b2 Bd3f1+ Rf2xf1 Rb8xb2+ Kg2h1 Qd1d3 Qh3h6 Bd6f8 Qh6h5 Qd3e2 Qh5h2 Qe2xh2+ Nf3xh2
12/22 00:00 260.872 595.598 -2,78 Nd2xf3 Rf8b8 b2xc3 Qa4d1 Bc1b2 Bd3f1+ Rf2xf1 Rb8xb2+ Kg2h1 Qd1d3 Qh3h6 Bd6f8 Qh6h5 Qd3e2 Qh5h2 Qe2xh2+ Nf3xh2 Rb2c2
13/24 00:00 364.613 554.966 -3,15 Nd2xf3 Rf8b8 b2xc3 Qa4d1 Bc1b2 Bd3f1+ Rf2xf1 Rb8xb2+ Kg2h1 Qd1d3 Qh3h6 Bd6f8 Qh6h5 Qd3e2 Qh5h2 Qe2xh2+ Nf3xh2 Ra8b8 Rf1f3 Bf8d6 Nh2g4
14/25 00:01 607.883 571.856 -3,55 Nd2xf3 Rf8b8 g5g6 Bd3xg6 b2xc3 Qa4d1 Qh3h6 Bd6f8 Qh6d2 Qd1xd2 Bc1xd2 Bg6e4 Kg2h3 Be4f5+ Kh3g3 Rb8b3
15/27 00:01 993.978 617.377 -3,39 Nd2xf3 Rf8b8 g5g6 Bd3xg6 b2xc3 Qa4d1 Qh3h6 Bd6f8 Qh6d2 Qd1xd2 Bc1xd2 Bf8d6 Nf3e5 Rb8b3 a3a4 Bg6e4+ Kg2h3 Ra8b8 Ra1g1+ Kg8f8
16/34+ 00:04 2.671.082 697.592 -2,65 Qh3h6 Nf3h4+ Kg2h3 Nh4f5 Rf2xf5 Bd3xf5+ Kh3h4

16/34+ 00:04 2.742.199 699.005 -2,24 Qh3h6 Nf3h4+ Kg2h3 Nh4f5 Rf2xf5 Bd3xf5+ Kh3h4
16/34+ 00:05 3.685.311 739.280 +101,52 Qh3h6 Nf3h4+ Kg2h3 Nh4f5 Rf2xf5 Bd3xf5+ Kh3h4
16/34 00:05 4.184.521 760.683 +M6 Qh3h6 Nf3h4+ Kg2h3 Nh4f5 Rf2xf5 Bd3xf5+ Kh3h4 Bd6g3+ Kh4xg3 c3xd2 Qh6g7+
17/34 00:06 4.474.199 773.815 +M6 Qh3h6 Nf3h4+ Kg2h3 Nh4f5 Rf2xf5 Bd3xf5+ Kh3h4 Bd6g3+ Kh4xg3 c3xd2 Qh6g7+
18/34 00:06 4.788.459 793.842 +M6 Qh3h6 Nf3h4+ Kg2h3 Nh4f5 Rf2xf5 Bd3xf5+ Kh3h4 Bd6g3+ Kh4xg3 c3xd2 Qh6g7+
19/34 00:06 5.228.424 820.016 +M6 Qh3h6 Nf3h4+ Kg2h3 Nh4f5 Rf2xf5 Bd3xf5+ Kh3h4 Bd6g3+ Kh4xg3 c3xd2 Qh6g7+
20/34 00:07 5.828.802 851.541 +M6 Qh3h6 Nf3h4+ Kg2h3 Nh4f5 Rf2xf5 Bd3xf5+ Kh3h4 Bd6g3+ Kh4xg3 c3xd2 Qh6g7+
Please remember that black had positive scores in the first analysis just because it is the side to move and SF thinks that black was better in that positon.

It would be interesting that SF team check this game. Is this blunder due to an excessive aggressive pruning?

Just checked the position after 32.- ..., Nxf3??; 33.- Qh6, Nh4+; 34.- Qxh4:

[d]r4rk1/p4p1p/3bpP2/3p2P1/q1pP3Q/P1pb4/1P1N1RK1/R1B5 b - - 0 34
FEN: r4rk1/p4p1p/3bpP2/3p2P1/q1pP3Q/P1pb4/1P1N1RK1/R1B5 b - - 0 34

Stockfish-3-32-ja:
1/2 00:00 338 338.000 +6,90 c3xd2 Bc1xd2 Bd3e4+ Kg2g1
2/4 00:00 536 536.000 +6,90 c3xd2 Bc1xd2 Bd3e4+ Kg2g1
3/5 00:00 1.334 1.334.000 +7,43 c3xd2 Bc1xd2 Bd3e4+ Kg2g1 Qa4b3
4/6 00:00 2.254 2.254.000 +7,07 c3xd2 Bc1xd2 Bd3e4+ Kg2g1 Qa4b3 Bd2c3
5/7 00:00 4.466 279.125 +6,90 c3xd2 Bc1xd2 Bd3e4+ Kg2g1 Qa4b3 Bd2f4 Bd6xf4 Qh4xf4
6/12 00:00 14.517 453.656 +6,50 c3xd2 Bc1xd2 Kg8h8 Kg2g1 e6e5 d4xe5 Bd6xe5 Bd2b4
7/12 00:00 18.307 381.395 +6,66 c3xd2 Bc1xd2 Kg8h8 Kg2g1 Bd3e4 Bd2f4 Bd6xf4 Qh4xf4 Qa4b3
8/14 00:00 31.116 393.873 +6,78 c3xd2 Bc1xd2 Kg8h8 Kg2g1 Qa4b3 Bd2b4 Bd6xb4 a3xb4 Bd3e4 b4b5 Qb3xb5
9/16 00:00 62.404 442.581 +7,15 c3xd2 Bc1xd2 Bd3e4+ Kg2f1 Kg8h8 Bd2f4 Bd6xf4 Qh4xf4 Qa4b3 Rf2e2 Qb3d3
10/21 00:00 105.975 519.485 +7,23 c3xd2 Bc1xd2 Bd3e4+ Kg2f1 Kg8h8 Bd2f4 Bd6xf4 Qh4xf4 Qa4b3 Kf1g1 Rf8g8 Rf2d2 Be4c2 Ra1e1 Qb3xb2 Rd2h2
11/21 00:00 113.057 513.895 +7,23 c3xd2 Bc1xd2 Bd3e4+ Kg2f1 Kg8h8 Bd2f4 Bd6xf4 Qh4xf4 Qa4b3 Kf1g1 Rf8g8 Rf2d2 Be4c2 Ra1e1 Qb3xb2 Rd2h2
12/23 00:00 218.329 580.662 +7,15 c3xd2 Bc1xd2 Bd3e4+ Kg2g1 Qa4b3 Bd2f4 Bd6xf4 Qh4xf4 Kg8h8 Ra1e1 Rf8g8 Rf2h2 Ra8b8 Re1c1 Qb3d3 Rh2d2 Qd3h3 Rd2h2
13/23 00:00 327.347 615.313 +7,11 c3xd2 Bc1xd2 Bd3e4+ Kg2g1 Qa4b3 Bd2f4 Bd6xf4 Qh4xf4 Kg8h8 Rf2h2 Rf8g8 a3a4 Ra8d8 Rh2d2 Qb3b8 Qf4e3 Qb8d6 Rd2h2 Rd8b8
14/23 00:00 352.765 609.265 +7,11 c3xd2 Bc1xd2 Bd3e4+ Kg2g1 Qa4b3 Bd2f4 Bd6xf4 Qh4xf4 Kg8h8 Rf2h2 Rf8g8 a3a4 Ra8d8 Rh2d2 Qb3b8 Qf4e3 Qb8d6 Rd2h2 Rd8b8
15/29 00:01 999.941 639.757 +7,43 c3xd2 Bc1xd2 Kg8h8 Kg2g1 Qa4b3 Bd2b4 Bd6xb4 a3xb4 Bd3e4 g5g6 Be4xg6 Ra1a3 Qb3xb4 Qh4h6 Rf8g8 Rf2g2
16/30 00:02 1.799.141 681.234 +7,79 c3xd2 Bc1xd2 Bd3e4+ Kg2g1 Qa4b3 Bd2f4 Bd6xf4 Qh4xf4 Kg8h8 Kg1h2 Rf8g8 Ra1g1 Ra8b8 g5g6 Be4xg6 Rg1g3 Bg6d3 Rf2g2 Bd3e4 Rg3xg8+ Rb8xg8 Rg2xg8+ Kh8xg8 Qf4g3+ Qb3xg3+ Kh2xg3 h7h5 Kg3f4
17/33 00:03 2.179.172 673.623 +7,71 c3xd2 Bc1xd2 Bd3e4+ Kg2g1 Qa4b3 Bd2f4 Bd6xf4 Qh4xf4 Kg8h8 Kg1h2 Rf8g8 Ra1g1 Ra8b8 Rf2e2 Rb8b7 g5g6 Be4xg6 Rg1g3 Qb3a2 Qf4c1 Bg6e4 Rg3xg8+ Kh8xg8 Qc1g5+ Kg8h8
18/33+ 00:03 2.548.589 682.353 +7,79 c3xd2 Bc1xd2 Bd3e4+ Kg2g1 Qa4b3 Bd2f4 Bd6xf4 Qh4xf4 Kg8h8 Kg1h2 Rf8g8 Ra1g1 Ra8b8 Rf2e2 Rb8b6 g5g6 Be4xg6 Qf4h4 Qb3d3 Re2f2 Rb6b3 Rg1g3 Qd3e4
18/33- 00:03 2.637.332 683.246 +7,63 c3xd2 Bc1xd2 Bd3e4+ Kg2g1 Qa4b3 Bd2f4 Bd6xf4 Qh4xf4 Kg8h8 Kg1h2 Rf8g8 Ra1g1 Ra8b8 Rg1a1 Rb8b7 Kh2g1 Rb7b8 Kg1h2
18/33+ 00:04 3.077.188 683.667 +7,87 c3xd2 Bc1xd2 Bd3e4+ Kg2g1 Qa4b3 Bd2f4 Bd6xf4 Qh4xf4 Kg8h8 Kg1h2 Rf8g8 Ra1g1 Ra8b8 Rg1a1 Rb8b7 Kh2g1 c4c3 b2xc3 Qb3xc3 Ra1d1 Rb7b3 Rf2h2 Rb3xa3
18/33 00:05 3.553.408 693.212 +8,00 c3xd2 Bc1xd2 Bd3e4+ Kg2g1 Qa4b3 Bd2f4 Bd6xf4 Qh4xf4 Kg8h8 Kg1h2 Rf8g8 Ra1g1 Ra8b8 Rf2e2 Rb8b7 g5g6 Be4xg6 Rg1g3 Qb3a2 Qf4c1 Rb7b3 Rg3xb3 Qa2xb3 Qc1c3 Bg6d3 Qc3xb3 c4xb3
19/34- 00:05 3.974.247 700.554 +7,91 c3xd2 Bc1xd2 Bd3e4+ Kg2g1 Qa4b3 Bd2f4 Bd6xf4 Qh4xf4 Kg8h8 Kg1h2 Rf8g8 Ra1g1 Ra8b8 Qf4c1 Qb3a2 Rg1g4 Rb8b3 g5g6
19/34- 00:05 4.025.143 698.082 +7,83 c3xd2 Bc1xd2 Bd3e4+ Kg2g1 Qa4b3 Bd2f4 Bd6xf4 Qh4xf4 Kg8h8 Kg1h2 Rf8g8 Ra1g1 Ra8b8 Qf4c1 Qb3a2 Rg1g4 Rb8b3 g5g6
19/34- 00:05 4.098.800 699.453 +7,71 c3xd2 Bc1xd2 Bd3e4+ Kg2g1 Qa4b3 Bd2f4 Bd6xf4 Qh4xf4 Kg8h8 Kg1h2 Rf8g8 Ra1g1 Ra8b8 Qf4c1 Qb3a2 Rg1g4 Rb8b3 g5g6
19/34 00:06 4.486.148 701.947 +7,67 c3xd2 Bc1xd2 Bd3e4+ Kg2g1 Qa4b3 Bd2f4 Bd6xf4 Qh4xf4 Kg8h8 Kg1h2 Rf8g8 Ra1g1 Ra8b8 Qf4c1 Qb3a2 Rg1g4 a7a5 Rg4g1 Rb8b3 g5g6 Be4xg6 Rf2g2 Rb3d3 Qc1h6 Rd3xd4 Kh2h1
20/34+ 00:07 5.186.315 709.191 +7,75 c3xd2 Bc1xd2 Bd3e4+ Kg2g1 Qa4b3 Bd2f4 Bd6xf4 Qh4xf4 Kg8h8 Kg1h2 Rf8g8 Ra1g1 Ra8b8 Qf4c1 Qb3a2 Rg1g4 a7a5 Rg4g1 Rb8b3 g5g6 Be4xg6 Rf2g2 Rb3d3 Qc1h6 Rd3xd4 Kh2h1 Qa2b3
20/34+ 00:07 5.412.425 711.225 +7,83 c3xd2 Bc1xd2 Bd3e4+ Kg2g1 Qa4b3 Bd2f4 Bd6xf4 Qh4xf4 Kg8h8 Kg1h2 Rf8g8 Ra1g1 Ra8b8 Qf4c1 Qb3a2 Rg1g4 a7a5 Rg4g1 Rb8b3 g5g6 Be4xg6 Rf2g2 Rb3d3 Qc1h6 Rd3xd4 Kh2h1 e6e5
20/36- 00:08 6.320.576 717.187 +7,59 c3xd2 Bc1xd2 Bd3e4+ Kg2g1 Qa4b3 Ra1e1 Bd6g3 Qh4h6 Bg3xf2+ Kg1f1 Qb3d3+ Re1e2 Be4g2+ Kf1xf2 Qd3f3+ Kf2g1 Qf3f1+ Kg1h2
20/36+ 00:10 7.409.654 723.952 +7,95 c3xd2 Bc1xd2 Bd3e4+ Kg2g1 Qa4b3 Ra1e1 Kg8h8 g5g6 Be4xg6 Rf2g2 Rf8b8 Bd2h6 Qb3f3 Re1e3 Qf3d1+ Re3e1 Qd1h5 Bh6g7+ Kh8g8
20/36 00:11 8.161.104 731.478 +8,04 c3xd2 Bc1xd2 Bd3e4+ Kg2g1 Qa4b3 Ra1e1 Kg8h8 g5g6 Be4xg6 Rf2g2 Rf8b8 Bd2h6 Qb3f3 Re1e3 Qf3d1+ Re3e1 Qd1h5 Qh4xh5 Bg6xh5 Bh6g7+ Kh8g8 Bg7h6+ Bh5g6 Re1f1 e6e5 Bh6e3 Rb8b3 d4xe5 Bd6xe5
21/36- 00:14 10.368.436 733.996 +7,95 c3xd2 Bc1xd2 Bd3e4+ Kg2g1 Qa4b3 Ra1e1 Kg8h8 Re1e3 Qb3xb2 g5g6 f7xg6 f6f7 Qb2a1+ Rf2f1 Qa1xd4 Bd2c3
21/36- 00:14 10.563.009 735.585 +7,87 c3xd2 Bc1xd2 Bd3e4+ Kg2g1 Qa4b3 Ra1e1 Kg8h8 Re1e3 Qb3xb2 g5g6 f7xg6 f6f7 Qb2a1+ Rf2f1 Qa1xd4 Bd2c3
21/36- 00:15 11.726.523 737.934 +7,57 c3xd2 Bc1xd2 Bd3e4+ Kg2g1 Qa4b3 Ra1e1 Kg8h8 Re1e3 Qb3xb2 g5g6 f7xg6 f6f7 Qb2a1+ Rf2f1 Qa1xd4 Bd2c3
21/36 00:18 13.456.304 743.031 +7,59 c3xd2 Bc1xd2 Bd3e4+ Kg2g1 Qa4b3 Ra1e1 Qb3g3+ Qh4xg3 Bd6xg3 Bd2b4 Rf8b8 Re1e3 a7a5 Bb4c3 Bg3c7 Re3h3 Be4f5 Rh3h5 Bc7d6 Rh5h2 Rb8b3 a3a4 Ra8b8 Rf2e2 Bf5d3 Re2f2 Bd3e4
22/36+ 00:20 15.172.790 748.091 +7,67 c3xd2 Bc1xd2 Bd3e4+ Kg2g1 Qa4b3 Ra1e1 Qb3g3+ Qh4xg3 Bd6xg3 Bd2b4 Rf8b8 Re1e3 a7a5 Bb4c3 Bg3c7 Re3h3 Rb8b3 Rh3h4 a5a4 Rf2f1 Bc7g3 Rh4h3 Bg3d6 Bc3b4 Rb3xh3 Bb4xd6 Rh3d3
22/36- 00:21 15.796.509 751.070 +7,51 c3xd2 Bc1xd2 Bd3e4+ Kg2g1 Qa4b3 Ra1e1 Qb3g3+ Qh4xg3 Bd6xg3 Bd2b4 Rf8b8 Re1e3 a7a5 Bb4c3 Bg3c7 a3a4 Kg8h8 Re3h3 Rb8g8 Rh3h5 Be4g6 Rh5h3 Bg6f5 Rh3h5 Bf5g4 Rh5h4 Bg4f5 Bc3d2 Ra8b8
22/46+ 00:24 18.234.318 755.795 +7,75 c3xd2 Bc1xd2 Bd3e4+ Kg2g1 Kg8h8 Rf2e2 Qa4c2
22/46+ 00:25 18.964.789 751.051 +7,93 c3xd2 Bc1xd2 Bd3e4+ Kg2g1 Kg8h8 Rf2e2 Qa4c2
22/46 00:29 22.219.031 747.234 +8,08 c3xd2 Bc1xd2 Bd3e4+ Kg2g1 Kg8h8 Bd2f4 Bd6xf4 Qh4xf4 Qa4b3 Kg1h2 Rf8g8 Ra1g1 Ra8b8 Qf4c1 Qb3a2 Rg1g4 a7a5 Qc1d2 Rb8b3 g5g6 Be4xg6 Rf2g2 Qa2a1 Rg4g3 Rb3xg3 Rg2xg3 Bg6e4 Rg3xg8+ Kh8xg8
23/46+ 00:35 26.616.656 752.733 +8,16 c3xd2 Bc1xd2 Bd3e4+ Kg2g1 Kg8h8 Bd2f4 Bd6xf4 Qh4xf4 Qa4b3 Kg1h2 Rf8g8 Ra1g1 Ra8b8 Qf4c1 Qb3a2 Rg1g4 a7a5 Qc1d2 Rb8b3 g5g6 Be4xg6 Rf2g2 Qa2a1 Rg4g3 Rb3xg3 Rg2xg3 Qa1f1
23/46- 00:39 29.827.215 762.337 +8,00 c3xd2 Qh4h6 Bd3e4+ Rf2f3 Be4xf3+ Kg2f2 d2d1N+ Kf2f1 Bf3g2+ Kf1g1 Bd6h2+ Kg1xh2
23/46- 00:44 30.752.479 762.842 +7,87 c3xd2 Qh4h6 Bd3e4+ Rf2f3 Be4xf3+ Kg2f2 d2d1N+ Kf2f1 Bf3g2+ Kf1g1 Bd6h2+ Kg1xh2
23/46- 00:44 32.127.847 764.657 +7,69 c3xd2 Qh4h6 Bd3e4+ Rf2f3 Be4xf3+ Kg2f2 d2d1N+ Kf2f1 Nd1e3+ Bc1xe3 Bf3g2+ Kf1g1
23/46- 00:49 39.038.099 786.646 +5,48 c3xd2 Qh4h6 Bd3e4+ Rf2f3 Be4xf3+ Kg2f2 d2d1N+ Kf2f1 Nd1e3+ Bc1xe3 Bf3g2+ Kf1g1
23/46 01:08 53.479.068 776.984 +4,60 Kg8h8 b2xc3 Qa4c2 Nd2xc4 Bd3e4+ Kg2g1 Qc2d1+ Rf2f1 Qd1e2 Nc4e3 Ra8c8 Rf1f2 Qe2e1+ Ne3f1 Qe1xc3 Bc1b2 Qc3d3 Ra1e1 Rc8c2 Nf1e3 Rc2xf2 Qh4xf2 Rf8b8 Qf2e2 Qd3xe2 Re1xe2 Rb8b3 a3a4 Bd6f4 Bb2c1 Bf4xg5

24/46 01:12 56.078.310 775.826 +4,60 Kg8h8 b2xc3 Qa4c2 Nd2xc4 Bd3e4+ Kg2g1 Qc2d1+ Rf2f1 Qd1e2 Nc4e3 Ra8c8 Rf1f2 Qe2e1+ Ne3f1 Qe1xc3 Bc1b2 Qc3d3 Ra1e1 Rc8c2 Nf1e3 Rc2xf2 Qh4xf2 Rf8b8 Qf2e2 Qd3xe2 Re1xe2 Rb8b3 a3a4 Bd6f4 Bb2c1 Bf4xg5
25/46+ 01:18 60.378.447 768.838 +4,68 Kg8h8 b2xc3 Qa4c2 Nd2xc4 Bd3e4+ Kg2g1 Qc2d1+ Rf2f1 Qd1e2 Nc4e3 Ra8b8 g5g6 Be4xg6 Qh4f2 Qe2h5 Qf2g2 Rf8g8
25/46+ 01:20 61.625.437 767.908 +4,76 Kg8h8 b2xc3 Qa4c2 Nd2xc4 Bd3e4+ Kg2g1 Qc2d1+ Rf2f1 Qd1e2 Nc4e3 Ra8b8 g5g6 Be4xg6 Qh4f2 Qe2h5 Qf2g2 Rf8g8
25/46+ 01:36 72.999.832 755.981 +5,07 Kg8h8 b2xc3 Qa4c2 Nd2xc4 Bd3e4+ Kg2g1 Qc2d1+ Rf2f1 Qd1e2 Nc4e3 Ra8b8 g5g6 Be4xg6 Qh4f2 Qe2h5 Qf2g2 Rf8g8
25/46 01:45 79.436.001 754.621 +4,80 Kg8h8 b2xc3 Qa4c2 Nd2f3 Qc2xc3 Bc1b2 Qc3b3 Ra1e1 Rf8g8 Bb2c1 Bd3e4 Re1xe4 d5xe4 Qh4xe4 Ra8b8 Rf2c2 Rb8b5 Rc2b2 Qb3d3 Qe4xd3 Rb5xb2+ Qd3d2 Rb2xd2+ Bc1xd2 Rg8c8 Bd2c3 Kh8g8 Kg2f2 Bd6xa3 Kf2e3 Ba3d6
26/46+ 01:50 82.997.729 753.128 +4,88 Kg8h8 b2xc3 Qa4c2 Nd2f3 Qc2xc3 Bc1b2 Qc3b3 Ra1e1 Rf8g8 Bb2c1 Bd3e4 Re1xe4 d5xe4 Qh4xe4 Ra8b8 Rf2c2 Rb8b5 Rc2b2 Qb3d3 Qe4xd3 Rb5xb2+ Qd3d2 Rb2xd2+ Bc1xd2 Rg8c8 Bd2c3 Kh8g8 Kg2f2 Bd6xa3 Kf2e3 Rc8b8
26/46+ 01:51 83.929.712 753.253 +4,96 Kg8h8 b2xc3 Qa4c2 Nd2f3 Qc2xc3 Bc1b2 Qc3b3 Ra1h1 Rf8g8 Rf2e2 c4c3 Bb2c1 Ra8b8 Qh4h6 Qb3c4 Rh1h3 c3c2 Nf3e5 Bd6xe5 d4xe5 Qc4g4+
26/46+ 01:56 87.411.758 752.026 +5,09 Kg8h8 b2xc3 Qa4c2 Nd2f3 Qc2xc3 Bc1b2 Qc3b3 Ra1e1 Rf8g8 Bb2c1 Bd3e4 Re1xe4 d5xe4 Qh4xe4 Ra8b8 Nf3e5 Bd6xe5 d4xe5 c4c3 Rf2e2 Rb8c8 Re2e1 Rc8c4 Qe4f3 Qb3c2+ Kg2f1 Qc2b1
26/46 02:04 93.879.639 753.200 +5,01 Kg8h8 b2xc3 Qa4c2 Nd2f3 Qc2xc3 Bc1b2 Qc3b3 Ra1e1 Rf8g8 Bb2c1 Bd3e4 Re1xe4 d5xe4 Qh4xe4 Ra8b8 Rf2c2 Rb8b5 Rc2b2 Qb3d3 Qe4xd3 Rb5xb2+ Qd3d2 Rb2xd2+ Bc1xd2 Rg8c8 Bd2c3 Kh8g8 Kg2f2 Bd6xa3 Kf2e3 Rc8b8 Nf3e5 Ba3c1+ Bc3d2 Bc1xd2+ Ke3xd2
27/46- 02:20 106.508.706 755.541 +4,92 Kg8h8 b2xc3 Qa4c2 Nd2f3 Qc2xc3 Bc1b2 Qc3b3 Ra1e1 Rf8g8 Bb2c1 Bd3e4 Re1xe4 d5xe4 Qh4xe4 Ra8b8 Rf2c2 Rb8b5 Rc2b2 Qb3d3 Qe4xd3 Rb5xb2+ Bc1xb2 c4xd3 Bb2c1 d3d2 Nf3xd2 Rg8xg5+ Kg2f3 Rg5f5+ Kf3e4 Rf5f4+ Ke4d3 Rf4xf6 Nd2c4 Rf6f3+ Kd3e4
27/46- 02:28 112.769.906 758.433 +4,84 Kg8h8 b2xc3 Qa4c2 Nd2f3 Qc2xc3 Bc1b2 Qc3b3 Ra1e1 c4c3 Bb2c1 Rf8g8 Bc1f4 Bd6xf4 Qh4xf4 Bd3e4 Re1xe4 d5xe4 Qf4xe4 Ra8c8 Rf2c2 Qb3xa3 Kg2f1
27/46- 02:40 122.114.033 759.793 +4,72 Kg8h8 b2xc3
27/46 03:04 138.958.369 753.157 +4,96 Kg8h8 b2xc3 Qa4c2 Nd2f3 Qc2xc3 Bc1b2 Qc3b3 Bb2c1
I stopped the analysis after 3.5 minutes. SF 3 fails to see 35.- Qh6 after 34.- ..., cxd2?? until depth 23! It is exactly the same pattern. If I play 35.- Qh6, SF realizes after a while that it is a checkmate in some moves. The same applies for Houdini.

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

But... surprise! Once 32.- ..., Nxf3?? is played, Houdini top choice is also 33.- Nxf3 (favouring black side with scores of +2) until depth 12, when changes its output to the correct white checkmates in six moves starting with 33.- Qh6.

Cyclone xTreme II (an old engine) was in the same error as Houdini, but it changed its mind at depth 10.

Ruffian 1.0.5 (an older engine) was in the same error too, but it changed its mind at depth 7, giving the correct score after 781454 nodes only.

Thinker 5.4D Inert MP moves 33.- Qh6 after 22 seconds, giving a checkmate score at depth 58:
FEN: r4rk1/p4p1p/3bpP1Q/3p2P1/q1pP4/P1pb1n2/1P1N1RK1/R1B5 b - - 1 33

MP-x86-Inert---Thinker:
Qh3-h6
58 00:22 85.884.503 3.903.841 +M105 h3h6 f3e1 g2h3 c3d2 c1d2
Thinker thinks that the best defence for black is 33.- ..., Ne1+ while the reality is that 33.- ..., Nh4+ is better in terms of a longer game (the result is the same: 1-0).

Quazar 0.4 w32 output is very similar to SF output, as expected:
FEN: r4rk1/p4p1p/3bpP2/3p2P1/q1pP4/P1pb1n1Q/1P1N1RK1/R1B5 w - - 0 33

Quazar_0.4_w32:
1 00:00 8 32 -4,33 Nd2xf3
2 00:00 44 176 -4,59 Nd2xf3 c3c2 Nf3e5
3 00:00 161 644 -4,26 Nd2xf3 c3c2 Nf3e5 Bd6xe5 d4xe5 Bd3e4+ Kg2g1
4 00:00 470 1.766 -4,26 Nd2xf3 c3xb2 Bc1xb2 Bd3f5 Qh3h6
5 00:00 1.359 5.109 -3,43 Nd2xf3 c3xb2 Bc1xb2 Rf8b8 Nf3e5 Bd6xe5 d4xe5 Bd3e4+ Kg2g3
6 00:00 2.946 10.446 -3,50 Nd2xf3 Bd3e4 Qh3h6 Be4xf3+ Rf2xf3 Qa4c2+ Rf3f2 Qc2e4+ Kg2f1 Qe4xd4 Qh6g7+
7 00:00 6.691 21.376 -2,00 Nd2xf3 h7h5 b2xc3 Qa4d1 Bc1b2 Qd1b3 Qh3xh5 Rf8b8
8 00:00 15.413 35.189 -2,58 Nd2xf3 Rf8d8 b2xc3 Qa4b3 Bc1d2 Bd6xa3 Nf3e5 Bd3e4+ Kg2g1 Ba3b2
9 00:00 16.561 37.810 -2,58 Nd2xf3 Rf8d8 b2xc3 Qa4b3 Bc1d2 Bd6xa3 Nf3e5 Bd3e4+ Kg2g1 Ba3b2
10 00:00 30.415 47.449 -3,13 Nd2xf3 Rf8b8 b2xc3 Rb8b3 g5g6 Bd3xg6 Nf3e5 Bd6xe5 d4xe5 h7h5 Kg2g1 Ra8b8 Qh3g3
11 00:01 57.773 52.047 -2,81 Nd2xf3 Rf8b8 b2xc3 Bd3e4 Bc1b2 Rb8b3 Qh3h6 Bd6f8 Qh6h2 Ra8b8 Bb2c1 Rb8b6
12 00:01 92.758 59.959 -3,27 Nd2xf3 Rf8b8 b2xc3 Bd3e4 Bc1b2 Rb8b3 Ra1e1 Ra8b8 Re1e2 Bd6xa3 Bb2xa3 Qa4xa3 Re2a2 Qa3d6 Qh3h6 Be4xf3+ Rf2xf3 Rb3b2+ Ra2xb2 Rb8xb2+ Rf3f2
13 00:01 115.855 63.343 -3,21 Nd2xf3 Rf8b8 b2xc3 Bd3e4 Qh3h6 Bd6f8 Qh6h2 Rb8b3 Bc1d2 Rb3xa3 Ra1xa3 Bf8xa3 Qh2h6 Ba3f8 Qh6h5 Bf8d6
14+ 00:02 145.041 66.777 -3,13 Qh3h6 Nf3h4+ Kg2h3 Nh4f5 Rf2xf5 Bd3xf5+ Kh3h4

14+ 00:02 150.151 68.624 -3,05 Qh3h6 Nf3h4+ Kg2h3 Nh4f5 Rf2xf5 Bd3xf5+ Kh3h4
14+ 00:03 354.591 99.969 +M8 Qh3h6 Nf3h4+ Kg2h3 Nh4f5 Rf2xf5 Bd3xf5+ Kh3h4
15 00:04 474.570 114.161 +M6 Qh3h6 Nf3h4+ Kg2h3 Nh4f5 Rf2xf5 Bd3xf5+ Kh3h4 Bd6g3+ Kh4xg3 c3xd2 Qh6g7+
16 00:04 579.530 131.981 +M6 Qh3h6 Nf3h4+ Kg2h3 Nh4f5 Rf2xf5 Bd3xf5+ Kh3h4 Bd6g3+ Kh4xg3 c3xd2 Qh6g7+
17 00:04 723.554 155.369 +M6 Qh3h6 Nf3h4+ Kg2h3 Nh4f5 Rf2xf5 Bd3xf5+ Kh3h4 Bd6g3+ Kh4xg3 c3xd2 Qh6g7+
18 00:05 932.007 182.388 +M6 Qh3h6 Nf3h4+ Kg2h3 Nh4f5 Rf2xf5 Bd3xf5+ Kh3h4 Bd6g3+ Kh4xg3 c3xd2 Qh6g7+
Here is Spike 1.2 Turin output:
FEN: r4rk1/p4p1p/3bpP2/3p2P1/q1pP4/P1pb1n1Q/1P1N1RK1/R1B5 w - - 0 33

Spike 1.2:
0 00:00 0 0 -10,63
1 00:00 681 42.562 -4,23 Nd2xf3 c3xb2 Bc1xb2
1 00:00 711 44.437 -4,23 Nd2xf3 c3xb2 Bc1xb2
2 00:00 795 49.687 -4,23 Nd2xf3 c3xb2 Bc1xb2
2 00:00 1.145 71.562 -4,23 Nd2xf3 c3xb2 Bc1xb2
3 00:00 1.530 95.625 -4,56 Nd2xf3 c3xb2 Bc1xb2 Ra8b8
3 00:00 2.881 180.062 -4,56 Nd2xf3 c3xb2 Bc1xb2 Ra8b8
4 00:00 5.353 172.677 -4,22 Nd2xf3 c3xb2 Bc1xb2 Bd3f5 Qh3h6
4 00:00 17.556 278.666 -4,22 Nd2xf3 c3xb2 Bc1xb2 Bd3f5 Qh3h6
5+ 00:00 23.066 366.126 -3,92 Nd2xf3 c3xb2 Bc1xb2 Bd3f5 Qh3h6
5+ 00:00 23.066 366.126 -3,92 Nd2xf3 c3xb2 Bc1xb2 Bd3f5 Qh3h6
5 00:00 31.308 333.063 -3,53 Nd2xf3 c3xb2 Bc1xb2 Bd3e4 Qh3h6 Be4xf3+ Rf2xf3 Qa4c2+ Rf3f2
5 00:00 56.646 401.744 -3,53 Nd2xf3 c3xb2 Bc1xb2 Bd3e4 Qh3h6 Be4xf3+ Rf2xf3 Qa4c2+ Rf3f2
6+ 00:00 77.849 414.090 -3,23 Nd2xf3 c3xb2 Qh3h6 Bd3f1+ Rf2xf1 Qa4c2+ Nf3d2
6+ 00:00 77.849 414.090 -3,23 Nd2xf3 c3xb2 Qh3h6 Bd3f1+ Rf2xf1 Qa4c2+ Nf3d2
6 00:00 130.492 416.907 -2,68 Nd2xf3 Kg8h8 b2xc3 Qa4b3 Qh3h6 Rf8g8
6+ 00:00 143.689 417.700 -1,17 Qh3h6 Nf3h4+ Kg2h3 Nh4f5 Rf2xf5 Bd3xf5+ Kh3h4 Bd6g3+ Kh4xg3

6+ 00:00 143.689 417.700 -1,17 Qh3h6 Nf3h4+ Kg2h3 Nh4f5 Rf2xf5 Bd3xf5+ Kh3h4 Bd6g3+ Kh4xg3
6 00:00 149.424 434.372 +M6 Qh3h6 Nf3h4+ Kg2h3 Nh4f5 Rf2xf5 Bd3xf5+ Kh3h4 Bd6g3+ Kh4xg3 c3xd2 Qh6g7+
6 00:00 150.212 436.662 +M6 Qh3h6 Nf3h4+ Kg2h3 Nh4f5 Rf2xf5 Bd3xf5+ Kh3h4 Bd6g3+ Kh4xg3 c3xd2 Qh6g7+
7 00:00 153.062 425.172 +M6 Qh3h6 Nf3h4+ Kg2h3 Nh4f5 Rf2xf5 Bd3xf5+ Kh3h4 Bd6g3+ Kh4xg3 c3xd2 Qh6g7+
7 00:00 155.480 431.888 +M6 Qh3h6 Nf3h4+ Kg2h3 Nh4f5 Rf2xf5 Bd3xf5+ Kh3h4 Bd6g3+ Kh4xg3 c3xd2 Qh6g7+
8 00:00 159.180 442.166 +M6 Qh3h6 Nf3h4+ Kg2h3 Nh4f5 Rf2xf5 Bd3xf5+ Kh3h4 Bd6g3+ Kh4xg3 c3xd2 Qh6g7+
8 00:00 167.544 446.784 +M6 Qh3h6 Nf3h4+ Kg2h3 Nh4f5 Rf2xf5 Bd3xf5+ Kh3h4 Bd6g3+ Kh4xg3 c3xd2 Qh6g7+
9 00:00 171.986 439.861 +M6 Qh3h6 Nf3h4+ Kg2h3 Nh4f5 Rf2xf5 Bd3xf5+ Kh3h4 Bd6g3+ Kh4xg3 c3xd2 Qh6g7+
9 00:00 197.053 449.892 +M6 Qh3h6 Nf3h4+ Kg2h3 Nh4f5 Rf2xf5 Bd3xf5+ Kh3h4 Bd6g3+ Kh4xg3 c3xd2 Qh6g7+
10 00:00 205.247 453.083 +M6 Qh3h6 Nf3h4+ Kg2h3 Nh4f5 Rf2xf5 Bd3xf5+ Kh3h4 Bd6g3+ Kh4xg3 c3xd2 Qh6g7+
10 00:00 238.460 462.131 +M6 Qh3h6 Nf3h4+ Kg2h3 Nh4f5 Rf2xf5 Bd3xf5+ Kh3h4 Bd6g3+ Kh4xg3 c3xd2 Qh6g7+
11 00:00 247.077 465.305 +M6 Qh3h6 Nf3h4+ Kg2h3 Nh4f5 Rf2xf5 Bd3xf5+ Kh3h4 Bd6g3+ Kh4xg3 c3xd2 Qh6g7+
11 00:00 291.092 465.747 +M6 Qh3h6 Nf3h4+ Kg2h3 Nh4f5 Rf2xf5 Bd3xf5+ Kh3h4 Bd6g3+ Kh4xg3 c3xd2 Qh6g7+
12 00:00 306.182 477.663 +M6 Qh3h6 Nf3h4+ Kg2h3 Nh4f5 Rf2xf5 Bd3xf5+ Kh3h4 Bd6g3+ Kh4xg3 c3xd2 Qh6g7+
12 00:00 383.930 491.587 +M6 Qh3h6 Nf3h4+ Kg2h3 Nh4f5 Rf2xf5 Bd3xf5+ Kh3h4 Bd6g3+ Kh4xg3 c3xd2 Qh6g7+
13 00:01 405.093 498.269 +M6 Qh3h6 Nf3h4+ Kg2h3 Nh4f5 Rf2xf5 Bd3xf5+ Kh3h4 Bd6g3+ Kh4xg3 c3xd2 Qh6g7+
13 00:01 533.790 509.828 +M6 Qh3h6 Nf3h4+ Kg2h3 Nh4f5 Rf2xf5 Bd3xf5+ Kh3h4 Bd6g3+ Kh4xg3 c3xd2 Qh6g7+
14 00:01 566.762 518.063 +M6 Qh3h6 Nf3h4+ Kg2h3 Nh4f5 Rf2xf5 Bd3xf5+ Kh3h4 Bd6g3+ Kh4xg3 c3xd2 Qh6g7+
14 00:01 642.610 527.161 +M6 Qh3h6 Nf3h4+ Kg2h3 Nh4f5 Rf2xf5 Bd3xf5+ Kh3h4 Bd6g3+ Kh4xg3 c3xd2 Qh6g7+
Now it is turn of an engine of year 2008... Zappa Mexico II w32 does not have problems!
FEN: r4rk1/p4p1p/3bpP2/3p2P1/q1pP4/P1pb1n1Q/1P1N1RK1/R1B5 w - - 0 33

Zappa Mexico II:
7/23 00:00 47.294 232.975 -1,56 Nd2xf3 Rf8b8 b2xc3 Bd3f5 Qh3h6 Bd6f8 Qh6h5 Bf8d6
7/23 00:00 148.167 249.860 +M6 Qh3h6 Nf3h4+ Kg2h3 Nh4f5 Rf2xf5 Bd3xf5+ Kh3h4 Bd6g3+ Kh4xg3 c3xd2 Qh6g7+

7/23 00:00 150.385 246.938 +M6 Qh3h6 Nf3h4+ Kg2h3 Nh4f5 Rf2xf5 Bd3xf5+ Kh3h4 Bd6g3+ Kh4xg3 c3xd2 Qh6g7+
8/23 00:00 150.440 247.028 +M6 Qh3h6 Nf3h4+ Kg2h3 Nh4f5 Rf2xf5 Bd3xf5+ Kh3h4 Bd6g3+ Kh4xg3 c3xd2 Qh6g7+
8/23 00:00 152.879 251.033 +M6 Qh3h6 Nf3h4+ Kg2h3 Nh4f5 Rf2xf5 Bd3xf5+ Kh3h4 Bd6g3+ Kh4xg3 c3xd2 Qh6g7+
9/23 00:00 160.375 238.653 +M6 Qh3h6 Nf3h4+ Kg2h3 Nh4f5 Rf2xf5 Bd3xf5+ Kh3h4 Bd6g3+ Kh4xg3 c3xd2 Qh6g7+
9/23 00:00 167.586 228.319 +M6 Qh3h6 Nf3h4+ Kg2h3 Nh4f5 Rf2xf5 Bd3xf5+ Kh3h4 Bd6g3+ Kh4xg3 c3xd2 Qh6g7+
10/23 00:00 170.906 232.842 +M6 Qh3h6 Nf3h4+ Kg2h3 Nh4f5 Rf2xf5 Bd3xf5+ Kh3h4 Bd6g3+ Kh4xg3 c3xd2 Qh6g7+
10/23 00:00 191.010 239.661 +M6 Qh3h6 Nf3h4+ Kg2h3 Nh4f5 Rf2xf5 Bd3xf5+ Kh3h4 Bd6g3+ Kh4xg3 c3xd2 Qh6g7+
11/23 00:00 193.514 238.318 +M6 Qh3h6 Nf3h4+ Kg2h3 Nh4f5 Rf2xf5 Bd3xf5+ Kh3h4 Bd6g3+ Kh4xg3 c3xd2 Qh6g7+
11/23 00:01 229.434 181.371 +M6 Qh3h6 Nf3h4+ Kg2h3 Nh4f5 Rf2xf5 Bd3xf5+ Kh3h4 Bd6g3+ Kh4xg3 c3xd2 Qh6g7+
12/23 00:01 239.496 184.654 +M6 Qh3h6 Nf3h4+ Kg2h3 Nh4f5 Rf2xf5 Bd3xf5+ Kh3h4 Bd6g3+ Kh4xg3 c3xd2 Qh6g7+
And finally, here is and old, almost forgotten gem: IPPOLIT 0.080d w32 from year 2009:
FEN: r4rk1/p4p1p/3bpP2/3p2P1/q1pP4/P1pb1n1Q/1P1N1RK1/R1B5 w - - 0 33

Ippolit0080d_w32:
1/9 00:00 65 4.000 -5,00 Kg2xf3
1/9 00:00 79 4.000 -2,16 Nd2xf3
2/9 00:00 226 14.000 -2,26 Nd2xf3 c3xb2 Bc1xb2
3/13 00:00 1.300 41.000 -2,28 Nd2xf3 c3xb2 Bc1xb2 Bd3f5
4/14 00:00 2.112 68.000 -2,28 Nd2xf3 c3xb2 Bc1xb2 Ra8b8 Nf3e5
5/16 00:00 4.183 134.000 -2,12 Nd2xf3 c3xb2 Bc1xb2 Bd3e4 Qh3h6 Be4xf3+ Rf2xf3
6/18 00:00 7.285 155.000 -2,02 Nd2xf3 c3xb2 Bc1xb2 Rf8b8 Qh3h6 Bd6f8 Qh6h2
7/18 00:00 10.280 218.000 -2,02 Nd2xf3 c3xb2 Bc1xb2 Rf8b8 Qh3h6 Bd6f8 Qh6h2 Bd3e4
8/20 00:00 14.939 237.000 -1,99 Nd2xf3 c3xb2 Bc1xb2 Rf8b8 Qh3h6 Bd6f8 Qh6h2 Bd3e4 Bb2c3
9/22 00:00 23.368 299.000 -1,91 Nd2xf3 c3xb2 Bc1xb2 Rf8b8 Qh3h6 Bd6f8 Qh6h2 Bd3e4 Bb2c3
9/22 00:00 27.357 350.000 -1,83 Nd2xf3 c3xb2 Bc1xb2 Rf8b8 Qh3h6 Bd6f8 Qh6h2 Bd3e4 Bb2c3
9/22 00:00 37.577 341.000 -1,53 Nd2xf3 c3xb2 Bc1xb2 Rf8b8 Qh3h6 Bd6f8 Qh6h2 Bd3e4 Bb2c3
9/22 00:00 57.601 408.000 -1,39 Nd2xf3 Rf8b8 b2xc3 Rb8b3 Bc1d2 Rb3xa3 Ra1xa3 Qa4xa3 Qh3h6 Bd6f8 Qh6h5
10/24 00:00 121.820 433.000 -1,33 Nd2xf3 Rf8b8 b2xc3 Rb8b3 Bc1d2 Rb3xa3 Ra1xa3 Qa4xa3 Qh3h6 Bd6f8 Qh6h2 Bd3e4
11/26 00:00 239.362 450.000 -1,28 Nd2xf3 Rf8b8 b2xc3 Rb8b3 Bc1d2 Ra8d8 Qh3h6 Bd6f8 Qh6h3 Bd3f5 Qh3h5
12/26 00:01 418.600 462.000 -1,33 Nd2xf3 Rf8b8 b2xc3 Rb8b3 Bc1d2 Bd3e4 Qh3h6 Bd6f8 Qh6h2 Qa4c6 Kg2g1 Rb3xa3 Ra1e1
13/31 00:02 1.150.203 474.000 -1,25 Qh3h6

13/31 00:03 1.362.669 481.000 -1,17 Qh3h6
13/36 00:22 11.268.759 493.000 +44,57 Qh3h6
13/36 00:23 11.623.157 496.000 +67,53 Qh3h6 Nf3e1+ Kg2g1 Ne1f3+ Nd2xf3 Qa4d1+ Kg1g2 Bd3f1+ Rf2xf1 Qd1c2+ Bc1d2 Qc2g6 Qh6xg6+ h7xg6 Bd2xc3 Rf8b8 Rf1h1 Rb8b3 a3a4 Ra8b8 Nf3d2 Rb3xb2 Bc3xb2
13/36 00:24 12.251.732 504.000 +231,12 Qh3h6 Nf3e1+ Kg2g1 Ne1f3+ Nd2xf3 Qa4d1+ Kg1g2 Bd3f1+ Rf2xf1 Qd1c2+ Bc1d2 Qc2g6 Qh6xg6+ h7xg6 Bd2xc3 Rf8b8 Rf1h1 Rb8b3 a3a4 Ra8b8 Nf3d2 Rb3xb2 Bc3xb2
13/36 00:25 12.891.390 504.000 +M6 Qh3h6 Nf3h4+ Kg2h3 Nh4f5 Rf2xf5 Bd3xf5+ Kh3h4 Bd6g3+ Kh4xg3 c3xd2 Qh6g7+
14/41 00:30 15.293.712 502.000 +M6 Qh3h6 Nf3h4+ Kg2h3 Nh4f5 Rf2xf5 Bd3xf5+ Kh3h4 Bd6g3+ Kh4xg3 c3xd2 Qh6g7+
15/54 00:43 21.657.688 496.000 +M6 Qh3h6 Nf3h4+ Kg2h3 Nh4f5 Rf2xf5 Bd3xf5+ Kh3h4 Bd6g3+ Kh4xg3 c3xd2 Qh6g7+
15/54 01:00 29.947.021 494.000 +M6 Qh3h6 Nf3h4+ Kg2h3 Nh4f5 Rf2xf5 Bd3xf5+ Kh3h4 Bd6g3+ Kh4xg3 c3xd2 Qh6g7+
I stopped this analysis after a minute.

It looks like older engines have less problems with this position, so my question is the following one: this blindness about Nxf3 and Qh6 could be due to modern search algorithms, LMR, IPPOLIT influence, etc.? Thanks in advance.

It is needless to say that hash was cleared between analysis and analysis.

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

Other thing is the TC for SF (maybe fixed depth 1? :P) or going backwards(trial and error) in some moves. I do not want to discuss because I do not like speculations. Good game by the way.

Regards from Spain.

Ajedrecista.
overlord
Posts: 198
Joined: Sun Jun 03, 2012 6:46 pm
Location: Trinec, Czech Republic

Re: Success against chess engine was never so easy :)

Post by overlord »

5 min. game,2 cores of mobile phone Samsung Galaxy Note II,ponder on,native book,full strength.Average depth in middle is 14-15 ply.In endgame around 20.
mar
Posts: 2668
Joined: Fri Nov 26, 2010 2:00 pm
Location: Czech Republic
Full name: Martin Sedlak

Re: Success against chess engine was never so easy :)

Post by mar »

overlord wrote:5 min. game,2 cores of mobile phone Samsung Galaxy Note II,ponder on,native book,full strength.Average depth in middle is 14-15 ply.In endgame around 20.
Ok that explains everything. I apologize - didn't know you played on a mobile phone.
User avatar
Don
Posts: 5106
Joined: Tue Apr 29, 2008 4:27 pm

Re: Success against chess engine was never so easy :)

Post by Don »

mar wrote:
overlord wrote:5 min. game,2 cores of mobile phone Samsung Galaxy Note II,ponder on,native book,full strength.Average depth in middle is 14-15 ply.In endgame around 20.
Ok that explains everything. I apologize - didn't know you played on a mobile phone.
I don't know about stockfish 3 but all other stockfish versions are very weak for a given depth. For example critter, houdini or komodo would crush stockfish if you played fixed depth games.

But Stockfish iterates very quickly and get to incredible depths, easily going a ply or more deeper than these other programs in the same amount of time. Due to the speed I was assume their search is far more selective and thus I could see it making blunders other programs might not, especially on a mobile phone platform.
Capital punishment would be more effective as a preventive measure if it were administered prior to the crime.
overlord
Posts: 198
Joined: Sun Jun 03, 2012 6:46 pm
Location: Trinec, Czech Republic

Re: Success against chess engine was never so easy :)

Post by overlord »

Don,

I absolutely agree with you. Stockfish evaluation under ply 20 (at least, is very suspicious). In free style, I start trusting Stockfish at plies around 25 and higher. In case of Houdini it is something about ply 20 (but of course, to destroy human being ply around 16 is more then enough).
User avatar
fern
Posts: 8755
Joined: Sun Feb 26, 2006 4:07 pm

Re: Man vs machine - how to beat chess engine

Post by fern »

Hi:
I am somewhat flabbergasted by your advices or perhaps i did not understand well.

You talk of horizon effect, which was a sensible issue 15 years ago, when engines did not go beyond, say, 7 or 8 moves. ow they calculate in half a minute 21 ply or more. Certainly that is an "horizon" far beyond the one we get.

In fact the only way to draw an engine is to play antichess , which is a way of not playing chess. I have never understood the rationale of a style of playing equivalent to rest from the game all his interest. Just to get half a point. I would do that only if a s a prize I get a night with Scarlett Johansen.

My best
fern
lkaufman
Posts: 6279
Joined: Sun Jan 10, 2010 6:15 am
Location: Maryland USA
Full name: Larry Kaufman

Re: Success against chess engine was never so easy :)

Post by lkaufman »

Another point is the the French defense is obviously the worst possible choice for an engine against a human. It is the most likely to lead to closed positions, in which the engine's calculating ability is of limited value. In a real human vs computer match the French would never be selected.