Any engines solve this one?
Moderators: hgm, Harvey Williamson, bob
Forum rules
This textbox is used to restore diagrams posted with the [d] tag before the upgrade.
This textbox is used to restore diagrams posted with the [d] tag before the upgrade.

 Posts: 7
 Joined: Wed Aug 30, 2017 11:26 pm
Any engines solve this one?
1. Nd4 with at least +2 or equivalent winning score.
Re: Any engines solve this one?
Cfishdev sees nothing in the position after Nd4:
info depth 62 seldepth 95 multipv 1 score cp 0 nodes 521067471148 nps 35892319 hashfull 1000 tbhits 0 time 14517520 pv f3h5 d3e2 h5g6 c4d5 b6d5 e2d3 g6h5 h1g1 d5f4 g3f4 g7g6 d3e2 h5h4 g1g4 h4h3 g4g3 h3h4 d4f3 h4f4 d1d4 f4h6 d4h4 h6g7 g3h3 a3e7 h4h7 g7h7 h3h7 g8h7 c2d2 b7b5 f3g5 e7g5 d2g5 c8b7 b1b2 a7a6 g5f4 h7g8 h2h4 f8d8 h4h5 g6h5 f4g5 g8f8 g5h6 f8g8 e2h5 d8d7 h6g5 g8f8 g5f6 f8g8 f6h4 g8g7 h5g4 a8c8 h4f6 g7g8 f6g5 g8f8 g5h6 f8e8 h6h8 e8e7 h8f6 e7e8 f6h8
Re: Any engines solve this one?
Seems like Cfish struggles to find Nd4 Qh5 g4!zullil wrote: ↑Fri Jul 12, 2019 9:49 amCfishdev sees nothing in the position after Nd4:
info depth 62 seldepth 95 multipv 1 score cp 0 nodes 521067471148 nps 35892319 hashfull 1000 tbhits 0 time 14517520 pv f3h5 d3e2 h5g6 c4d5 b6d5 e2d3 g6h5 h1g1 d5f4 g3f4 g7g6 d3e2 h5h4 g1g4 h4h3 g4g3 h3h4 d4f3 h4f4 d1d4 f4h6 d4h4 h6g7 g3h3 a3e7 h4h7 g7h7 h3h7 g8h7 c2d2 b7b5 f3g5 e7g5 d2g5 c8b7 b1b2 a7a6 g5f4 h7g8 h2h4 f8d8 h4h5 g6h5 f4g5 g8f8 g5h6 f8g8 e2h5 d8d7 h6g5 g8f8 g5f6 f8g8 f6h4 g8g7 h5g4 a8c8 h4f6 g7g8 f6g5 g8f8 g5h6 f8e8 h6h8 e8e7 h8f6 e7e8 f6h8
 Eelco de Groot
 Posts: 4103
 Joined: Sun Mar 12, 2006 1:40 am
 Location: Groningen
Re: Any engines solve this one?
Certainly interesting guys, if that is correct, is it certain that 2. g4 is the main line intended after 1. Nd4 Qh5. Seeing nothing yet with Multi PV overnight. Currently at 0.10 for 2. g4 fifth choice.
Engine: Kaissa IV NoContempt (512 MB)
by T. Romstad, M. Costalba, J. Kiiski, G. Linscott
49 499:25 0.00 2.Be2 Qg6 3.Bd3 Qh5 (43.285.771.565) 1444
49 499:25 0.00 2.Rdg1 Nxc4 3.Be2 Qg6 4.Bd3 (43.285.771.565) 1444
49 499:25 0.00 2.Nb5 Bc5 3.cxd5 Nxd5 4.Be2 Qg6
5.Bd3 (43.285.771.565) 1444
49 499:25 0.00 2.cxd5 Nxd5 3.Be2 Qh3 4.Bf1 Qh5 (43.285.771.565) 1444
49 499:25 0.00 2.c5 Nc4 3.Qb3 Bxc5 4.Be2 Qg6+ 5.Bd3 Qh5 (43.285.771.565) 1444
48 499:25 0.10 2.g4 Qh4 3.cxd5 Nxd5 4.Bg3 Qh3 5.Be4 b5
6.Ka1 Bd7 7.Qb3 Bc5 8.f3 Rad8 9.Rhe1 b4
10.Bxd5 Bxd4 11.Rxd4 exd5 12.cxb4 Be6
13.Qd3 Rc8 14.Qf1 Qh6 15.Qe2 (43.285.771.565) 1444
But doing iteration 49 now of that move.
Engine: Kaissa IV NoContempt (512 MB)
by T. Romstad, M. Costalba, J. Kiiski, G. Linscott
49 499:25 0.00 2.Be2 Qg6 3.Bd3 Qh5 (43.285.771.565) 1444
49 499:25 0.00 2.Rdg1 Nxc4 3.Be2 Qg6 4.Bd3 (43.285.771.565) 1444
49 499:25 0.00 2.Nb5 Bc5 3.cxd5 Nxd5 4.Be2 Qg6
5.Bd3 (43.285.771.565) 1444
49 499:25 0.00 2.cxd5 Nxd5 3.Be2 Qh3 4.Bf1 Qh5 (43.285.771.565) 1444
49 499:25 0.00 2.c5 Nc4 3.Qb3 Bxc5 4.Be2 Qg6+ 5.Bd3 Qh5 (43.285.771.565) 1444
48 499:25 0.10 2.g4 Qh4 3.cxd5 Nxd5 4.Bg3 Qh3 5.Be4 b5
6.Ka1 Bd7 7.Qb3 Bc5 8.f3 Rad8 9.Rhe1 b4
10.Bxd5 Bxd4 11.Rxd4 exd5 12.cxb4 Be6
13.Qd3 Rc8 14.Qf1 Qh6 15.Qe2 (43.285.771.565) 1444
But doing iteration 49 now of that move.
Debugging is twice as hard as writing the code in the first
place. Therefore, if you write the code as cleverly as possible, you
are, by definition, not smart enough to debug it.
 Brian W. Kernighan
place. Therefore, if you write the code as cleverly as possible, you
are, by definition, not smart enough to debug it.
 Brian W. Kernighan

 Posts: 7
 Joined: Wed Aug 30, 2017 11:26 pm
Re: Any engines solve this one?
Yes, 2. g4 is the idea. Engines will see 2... Qxg4 3. Qd2 as good for White but may struggle to see 2... Qh4 3. Nf3 Qxg4 4. Rhg1. Even after forcing 1. Nd4 it is a difficult position for engines, that makes the original position very, very difficult!Eelco de Groot wrote: ↑Fri Jul 12, 2019 12:30 pmCertainly interesting guys, if that is correct, is it certain that 2. g4 is the main line intended after 1. Nd4 Qh5.
The game had a beautiful (but not all best moves) finish: 1. Nd4 Qh5 2. g4 Qh4 3. Nf3 Qxg4 4. Rhg1 Qxf3 5. Bxh7+ Kh8 6. Rxg7 Kxg7 7. Rg1+ Kh8 8. Bh6 f5 9. exf6 Qxf6 10. Bg7+ Qxg7 11. Rxg7 Kxg7 12. Qg6+ Kh8 13. Bg8 Rxg8 14. Qh6#
Re: Any engines solve this one?
Who played this game ?
Re: Any engines solve this one?
With a bit of learning, after 1. Nd4 Qh5 2. g4 Qh4 3. Nf3 Qxg4 4. Rhg1 Qxf3 5. Bxh7+ Kh8 6. Rxg7 Kxg7, SF eval rise to +5 (since 3...Qxg4).King Solomon wrote: ↑Fri Jul 12, 2019 12:44 pmYes, 2. g4 is the idea. Engines will see 2... Qxg4 3. Qd2 as good for White but may struggle to see 2... Qh4 3. Nf3 Qxg4 4. Rhg1. Even after forcing 1. Nd4 it is a difficult position for engines, that makes the original position very, very difficult!Eelco de Groot wrote: ↑Fri Jul 12, 2019 12:30 pmCertainly interesting guys, if that is correct, is it certain that 2. g4 is the main line intended after 1. Nd4 Qh5.
The game had a beautiful (but not all best moves) finish: 1. Nd4 Qh5 2. g4 Qh4 3. Nf3 Qxg4 4. Rhg1 Qxf3 5. Bxh7+ Kh8 6. Rxg7 Kxg7 7. Rg1+ Kh8 8. Bh6 f5 9. exf6 Qxf6 10. Bg7+ Qxg7 11. Rxg7 Kxg7 12. Qg6+ Kh8 13. Bg8 Rxg8 14. Qh6#
On black 3rd move, SF like more :
Code: Select all
29/47 01:09 406.710k 5.819k +2,59 3. ... Qh3 4.Ng5 Qxg4 5.Be3 Be7 6.Rdg1 Qh4 7.Bxh7+ Kh8 8.Bd3 dxc4 9.Be2 f5 10.exf6 Rxf6 11.Rg4 Qh5 12.Rg3 Qg6 13.Ne4 Qh7 14.Rhg1 Rf7 15.Bg4 Kg8 16.Rh3 Qg6 17.Rgg3 Nd5 18.Bd4 Nf4 19.Bh5 Qxh5 20.Rxh5 Nxh5 21.Rg6 e5 22.Bxe5 Bf5 23.Rg1 Rff8 24.Bd4
29/61 01:09 406.710k 5.819k +2,59 3. ... Qd8 4.Bxh7+ Kh8 5.c5 Nd7 6.Ng5 Nxc5 7.Ka1 Qc7 8.Rde1 Qa5 9.Re3 Qa4 10.Rh3 Qxc2 11.Bxc2+ Kg8 12.Rd1 b6 13.Nh7 Nd7 14.Nxf8 Nxf8 15.Bg5 Ba6 16.Bd3 Bxd3 17.Rhxd3 Ng6 18.f4 Re8 19.c4 Be7 20.Bxe7 Nxe7 21.cxd5 Nxd5 22.f5 Rc8 23.Kb2
29/55 01:09 406.710k 5.819k +2,38 3. ... Qe7 4.Bxh7+ Kh8 5.Ng5 Qc7 6.Ka1 Nxc4 7.Rb1 g6 8.h4 f6 9.Bxg6 fxe5 10.Bg3 Bd6 11.Qd3 Qg7 12.h5 e4 13.Nxe4 Be5 14.Rbc1 Bxg3 15.Nxg3 Bd7 16.g5 Qe5 17.Ne4 Bc6 18.Nf6 Qxg5 19.Nh7 Qd2 20.Qxd2 Nxd2 21.Nxf8 Rxf8 22.Rce1 Ne4 23.Bxe4 dxe4 24.Ref1 Kh7 25.Kb1
Code: Select all
30/47 00:31 209.077k 6.669k +6,98 1. ... Qg2 2.h4 Bd7 3.Rdg1 Ba4 4.Qe2 Qxg1+ 5.Rxg1 dxc4 6.Be4 Nd7 7.Qxc4 Nc5 8.Nc2 b5 9.Qd4 Bxc2+ 10.Bxc2 Rfc8 11.Bd2 Rab8 12.Rd1 Rd8 13.Qe3 Na4 14.Be1 Rxd1+ 15.Bxd1 Nc5 16.Bc2 b4 17.c4 Rc8 18.Qf4 Rd8 19.Bd2 h6
30/49 00:31 209.077k 6.669k +4,40 1. ... Qg4 2.h4 Qh5 3.g4 Qxg4 4.Qd2 Qh5 5.Rdg1 f6 6.Bh6 g6 7.exf6 Rxf6 8.Bg5 Rf3 9.Nxf3 Qxf3 10.h5 e5 11.hxg6 hxg6 12.Be3 Bf5 13.Qc2 Bxd3 14.Qxd3 Qe4 15.Qxe4 dxe4 16.Bxb6 axb6 17.Rxg6+ Kf7 18.Rxb6 Rf8 19.Rxb7+ Ke6 20.Rb6+ Kf5 21.Rb5 e3 22.fxe3 Ke4 23.c5 Kxe3
30/36 00:31 209.077k 6.669k +2,55 1. ... Qh5 2.g4 Qh4 3.Nf3 Qh3 4.Ng5 Qxg4 5.Be3 Be7 6.Rhg1 Qh4 7.Nxh7 Nxc4 8.Qe2 g6 9.Nxf8 Kxf8 10.Bxc4 Qxc4 11.Qxc4 dxc4 12.h4 b5 13.h5 Bb7 14.hxg6 fxg6 15.Kb2 Bf3 16.Rd4 Bh5 17.Rd7 a6 18.Rc7 Ke8
Re: Any engines solve this one?
Code: Select all
Analysis by CyberNezhdev:
1.Nd4
White has a decisive advantage: + (2.55 ++) Depth: 32/51 00:00:55 414MN
1.Nd4
White has a decisive advantage: + (3.16 ++) Depth: 32/51 00:00:55 415MN
1.Nd4 Qh5 2.g4 Qxg4 3.Qd2 Nxc4 4.Bxc4 dxc4 5.Rhg1 Qh5 6.f3 Be7 7.Bg5 Bxg5 8.Rxg5 Qh6 9.Rdg1 g6 10.Nb5 Bd7 11.Rxg6+ fxg6 12.Qxh6 Bxb5 13.h4 Rxf3 14.h5 Be8 15.Qg5 Kh8 16.Qg2 Rf7 17.Qh3 Rg7 18.h6 Rf7 19.Qxe6
White is winning: + (3.51) Depth: 32/51 00:00:55 419MN
1.Nd4 Qh5 2.g4 Qxg4 3.Qd2 Be7 4.h4 Nxc4 5.Bxc4 dxc4 6.h5 Rd8 7.Ka1 b5 8.Rdg1 Rxd4 9.cxd4 Bb4 10.Rxg4 Bxd2 11.Bxd2 Bb7 12.Rhg1 g6 13.Rh4 Bd5 14.Rb1 Rb8 15.Bb4 Rb7 16.Rf4 Kg7 17.Kb2 c3+ 18.Kxc3 Bxa2 19.Rg1 Bd5 20.hxg6 Rc7+ 21.Kb2 hxg6
White is clearly winning: + (4.49) Depth: 35/61 00:01:29 677MN
(, 12.07.2019)

 Posts: 9782
 Joined: Wed Mar 08, 2006 7:57 pm
 Location: Redmond, WA USA
 Contact:
Re: Any engines solve this one?
First move for white, bm is projected as Nd4, but Stockfish stubbornly insists on cxd5, worth a measly quarter of a pawn:carldaman wrote: ↑Fri Jul 12, 2019 6:07 pm
We clearly need better engines than St'Fish.Code: Select all
Analysis by CyberNezhdev: 1.Nd4 White has a decisive advantage: + (2.55 ++) Depth: 32/51 00:00:55 414MN 1.Nd4 White has a decisive advantage: + (3.16 ++) Depth: 32/51 00:00:55 415MN 1.Nd4 Qh5 2.g4 Qxg4 3.Qd2 Nxc4 4.Bxc4 dxc4 5.Rhg1 Qh5 6.f3 Be7 7.Bg5 Bxg5 8.Rxg5 Qh6 9.Rdg1 g6 10.Nb5 Bd7 11.Rxg6+ fxg6 12.Qxh6 Bxb5 13.h4 Rxf3 14.h5 Be8 15.Qg5 Kh8 16.Qg2 Rf7 17.Qh3 Rg7 18.h6 Rf7 19.Qxe6 White is winning: + (3.51) Depth: 32/51 00:00:55 419MN 1.Nd4 Qh5 2.g4 Qxg4 3.Qd2 Be7 4.h4 Nxc4 5.Bxc4 dxc4 6.h5 Rd8 7.Ka1 b5 8.Rdg1 Rxd4 9.cxd4 Bb4 10.Rxg4 Bxd2 11.Bxd2 Bb7 12.Rhg1 g6 13.Rh4 Bd5 14.Rb1 Rb8 15.Bb4 Rb7 16.Rf4 Kg7 17.Kb2 c3+ 18.Kxc3 Bxa2 19.Rg1 Bd5 20.hxg6 Rc7+ 21.Kb2 hxg6 White is clearly winning: + (4.49) Depth: 35/61 00:01:29 677MN (, 12.07.2019)
acd 36; acs 203; bm cxd5; c3 "Nd4"; ce 24; pm cxd5; pv cxd5 Nxd5 Bg5 Bc5 Bxh7+ Kh8 Bd3 Bd7 h4 Ba4 Qxa4 Qxe2 Bxe2 Nxc3+ Kc2 Nxa4 Kb3 b5 Bxb5 Nb6 Rd2 a6 Rc1 Bb4 Kxb4 axb5 Be7 Rfb8 Bd6 Rb7 Kb3 Nc4 Rdd1 Rba7 a3 Kh7 f4 Ra4 Bb4 Na5+ Bxa5 R8xa5 Rc7 Rxa3+ Kb4 R3a4+ Kc5 b4+ Kc4 b3+ Kxb3 Ra3+ Kc4 R3a4+ Kd3 Rd5+ Ke2 Ra2+ Ke1 Rxd1+ Kxd1 Rg2 Rxf7 Rxg3 h5 Rh3 Ke2 Rxh5 Kf3 Kg6 Rb7 Rh1;
First response for black after forcing Nd4, Sockfish agrees with Qh5 but thinks it is a drawn position:
acd 36; acs 23; bm Qh5; c3 "Qh5"; ce 0; pm Qh5; pv Qh5 Be2 Qg6 Be3 Be7 Qxg6 fxg6 Bg4 Kf7 cxd5 Nxd5 Rd3 Bd7 Rhd1 Rac8 Nxe6 Nxe3 Ng5+ Kg8 Rxd7 Nxd1 Be6+ Kh8 Rxe7 Rxc3 Rd7 Nxf2 Kb2 Rd3 Nf7+ Kg8 Nd6+ Kh8 Bb3 Re3 Nf7+ Kg8 Nd8+ Kh8;
Second move for white after Qh5 Stockfish wants to play h4, not g4 and still thinks it is drawn:
acd 36; acs 35; bm h4; c3 "g4"; ce 0; pm h4; pv h4 dxc4 Be4 Bd7 g4 Qxg4 Qd2 Bc6 Nxc6 bxc6 Rdg1 Qh5 Rg5 Qh6 Rg4 Qh5;
Second move for black, after forcing g4, we see some kind of a problem. Suggested move for black is Qxg4, but Stockfish wants Qh4 with a plus score for black:
acd 36; acs 71; am Qxg4; bm Qh4; c3 "Qxg4"; ce 15; pm Qh4; pv Qh4 cxd5 Nxd5 Bg3 Qh3 c4 Nb4 Qe2 Nxd3 Rxd3 Qg2 Qf1 Qxf1+ Rxf1 Bc5 Nb5 b6 Nd6 Ba6 Rfd1 Rfd8 Bh4 Rdb8 f4 b5 cxb5 Bxd6 Rxd6 Bxb5 Rd8+ Rxd8 Rxd8+ Rxd8 Bxd8 h5 gxh5 Kh7 h3 Kh6 Be7 Kxh5 Kc2 Bf1 Kc3 Bxh3 Bc5;
After forcing Qxg4 white is going to win. But black is not going to play Qxg4 on the previous ply, but instead Qh4.
acd 36; acs 207; bm Qd2; c3 "Qd2"; ce 218; pm Qd2; pv Qd2 Qh5 Rhg1 g6 Bg5 Nxc4 Qf4 f6 Bxf6 Rf7 h4 Bd7 Nf3 Bf8 Rg5 Bh6 Rxh5 Bxf4 Bxg6 Rxf6 Bxh7+ Kf7 exf6 Kxf6 Bd3 Rc8 Nd4 Be8 Rh8 Nd6 Ne2 Be5 Rh6+ Ke7 Rh7+ Bf7 f4 Bxc3 Nxc3 Rxc3 Kb2 Rc8 h5 a5 Bg6 a4 Bxf7 Nxf7 Rg1 Rc4;
Taking ideas is not a vice, it is a virtue. We have another word for this. It is called learning.
But sharing ideas is an even greater virtue. We have another word for this. It is called teaching.
But sharing ideas is an even greater virtue. We have another word for this. It is called teaching.
Re: Any engines solve this one?
Great one!
Most engines in 3MV mode have Nd4 at third best from very start, yet none with default parameters found 2.g4 after 1...Qh5 within 15 minutes at my trials at 24 threads of 3x12GHz CPU except Houdini 6.03 Tactical (probably).
This one takes at a single try when having 1.Nd4 up at about 12 minutes still 15 more minutes for showing this one point of the position in output line, till then always only 1.Nd4 with slowly climbing eval, so it can't be said for sure when exactly 2.g4 is found, even if I'd guess it might be already at the point of rising eval.
Peter.