One more issue with king safety/attack for Stockfish

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

Moderators: bob, hgm, Harvey Williamson

Forum rules
This textbox is used to restore diagrams posted with the [d] tag before the upgrade.
Vinvin
Posts: 4378
Joined: Thu Mar 09, 2006 8:40 am
Full name: Vincent Lejeune

One more issue with king safety/attack for Stockfish

Post by Vinvin » Tue Jul 16, 2019 11:28 pm

From the game Vitiugov - Grischuk (Riga Grand Prix 2019).

https://www.chessbomb.com/arena/2019-ri ... _Alexander

Stockfish wants to play the mistake 20.Nxe4? with a big advantage (+3) for a long time because it has trouble to see 20...Qxg6! for black.
Houdini 1.5 and Critter 1.6a find Qxg6 is less than 0.1 second.
[d]2kr3r/p5p1/2pq4/n1b1pb2/4N1p1/3PB1P1/PPP1BPK1/R3QR2 b - - 0 20

Code: Select all

Stockfish_19071209_x64_modern:
Found 511 tablebases
...
 34/55	01:17	 612 703 556	7 888 142	+4,09	20. ... Bxe4+ 21.dxe4 Nb7 22.Bxg4+ Kb8 23.Rd1 Qe7 24.Rxd8+ Rxd8 25.Rh1 Rf8 26.Rh7 Bb6 27.Qc3 Bxe3 28.Qxe3 a5 29.Rh5 g6 30.Rh6 Qc5 31.Qd2 Qc4 32.Bf3 Qf7 33.Qe3 Qg7 34.Be2 Rd8 35.Bg4 Rd6 36.c3 c5 37.Qg5 Ka7 38.b3 Ka6 39.Bf5 Kb6
 35/54+	01:27	 690 987 911	7 875 942	+4,18	20. ... Bxe4+ 21.dxe4
 35/60-	01:29	 707 439 196	7 877 328	+4,09	20. ... Bxe4+
 35/60	01:32	 725 748 551	7 884 286	+4,12	20. ... Bxe4+ 21.dxe4 Nb7 22.Bxg4+ Kc7 23.Rd1 Qe7 24.Rxd8 Rxd8 25.Rh1 Re8 26.Rh7 Bb6 27.Bh6 Kb8 28.Qd2 Rd8 29.Qc3 Qf6 30.Be3 Kc7 31.Rh5 Nd6 32.Bxb6+ axb6 33.Qxe5 Re8 34.Qxf6 gxf6 35.Rh7+ Kb8 36.f3 Nc4 37.Kh3 Nxb2 38.Bd7 Rd8 39.Bxc6 Nc4 40.Bd5 Ne3 41.c4 Kc8
 36/57-	02:00	 936 629 206	7 801 277	+4,03	20. ... Qg6
 36/57-	02:00	 936 747 385	7 801 287	+3,92	20. ... Qg6
 36/57-	02:00	 936 900 658	7 801 329	-2,91	20. ... Qg6
 36/57	02:00	 938 603 260	7 800 113	-3,89	20. ... Qg6 21.Rh1 Bxe4+ 22.dxe4 Rxh1 23.Qxh1 Bxe3 24.Qh4 Qxe4+ 25.Kf1 Rf8 26.Qxg4+ Qxg4 27.Bxg4+ Kc7 28.Ke2 Bd4 29.f3 Bxb2 30.Rb1 Bd4 31.a4 Rh8 32.Kd3 Nb7 33.a5 Bf2 34.Bf5 Nxa5 35.Be4 Nb7 36.Ke2 Bd4 37.Bd3 Nc5
 37/42-	02:07	 986 277 842	7 763 644	-3,99	20. ... Qg6
 37/46-	02:10	1 010 990 683	7 769 201	-4,08	20. ... Qg6
 37/61-	02:27	1 154 003 847	7 813 689	-5,04	20. ... Qg6

zullil
Posts: 5668
Joined: Mon Jan 08, 2007 11:31 pm
Location: PA USA
Full name: Louis Zulli

Re: One more issue with king safety/attack for Stockfish

Post by zullil » Wed Jul 17, 2019 12:01 am

Stockfish-dev with default settings (16 MB hash and single-threaded, deterministic search) needs about 172 million nodes to find Qg6:

info depth 35 seldepth 51 multipv 1 score cp -365 lowerbound nodes 171929033 nps 1976877 hashfull 1000 tbhits 0 time 86970 pv d6g6
info depth 34 currmove d6g6 currmovenumber 1
info depth 35 seldepth 51 multipv 1 score cp -351 lowerbound nodes 172165946 nps 1976624 hashfull 1000 tbhits 0 time 87101 pv d6g6
info depth 33 currmove d6g6 currmovenumber 1
info depth 35 seldepth 51 multipv 1 score cp -331 lowerbound nodes 172335433 nps 1976550 hashfull 1000 tbhits 0 time 87190 pv d6g6
info depth 32 currmove d6g6 currmovenumber 1
info depth 35 seldepth 51 multipv 1 score cp -305 lowerbound nodes 172431167 nps 1976424 hashfull 1000 tbhits 0 time 87244 pv d6g6
info depth 31 currmove d6g6 currmovenumber 1
info depth 35 seldepth 51 multipv 1 score cp -269 lowerbound nodes 172584594 nps 1976212 hashfull 1000 tbhits 0 time 87331 pv d6g6
info depth 30 currmove d6g6 currmovenumber 1
info depth 35 seldepth 51 multipv 1 score cp -222 lowerbound nodes 172787832 nps 1975892 hashfull 1000 tbhits 0 time 87448 pv d6g6
info depth 29 currmove d6g6 currmovenumber 1
info depth 35 seldepth 51 multipv 1 score cp -161 lowerbound nodes 173334357 nps 1975297 hashfull 1000 tbhits 0 time 87751 pv d6g6
info depth 28 currmove d6g6 currmovenumber 1
info depth 35 seldepth 51 multipv 1 score cp -83 lowerbound nodes 173562031 nps 1974988 hashfull 1000 tbhits 0 time 87880 pv d6g6
info depth 27 currmove d6g6 currmovenumber 1
info depth 35 seldepth 51 multipv 1 score cp 16 lowerbound nodes 173910556 nps 1974595 hashfull 1000 tbhits 0 time 88074 pv d6g6
info depth 26 currmove d6g6 currmovenumber 1
info depth 35 seldepth 51 multipv 1 score cp 144 lowerbound nodes 174250511 nps 1974263 hashfull 1000 tbhits 0 time 88261 pv d6g6
info depth 25 currmove d6g6 currmovenumber 1
info depth 35 seldepth 55 multipv 1 score cp 305 lowerbound nodes 177152544 nps 1976134 hashfull 1000 tbhits 0 time 89646 pv d6g6
info depth 24 currmove d6g6 currmovenumber 1
info depth 24 currmove c5e3 currmovenumber 2
info depth 24 currmove f5e4 currmovenumber 3
info depth 24 currmove d6d5 currmovenumber 4
info depth 24 currmove d6f8 currmovenumber 5
info depth 24 currmove f5d7 currmovenumber 6
info depth 24 currmove d6e6 currmovenumber 7
info depth 24 currmove d6d7 currmovenumber 8
info depth 24 currmove c5d4 currmovenumber 9
info depth 24 currmove c5b6 currmovenumber 10
info depth 24 currmove d6f6 currmovenumber 11
info depth 24 currmove h8h3 currmovenumber 12
info depth 24 currmove g7g6 currmovenumber 13
info depth 24 currmove a5b7 currmovenumber 14
info depth 24 currmove d8d7 currmovenumber 15
info depth 24 currmove c8b7 currmovenumber 16
info depth 24 currmove h8h1 currmovenumber 17
info depth 24 currmove d6h6 currmovenumber 18
info depth 24 currmove d8f8 currmovenumber 19
info depth 24 currmove c8b8 currmovenumber 20
info depth 24 currmove c8d7 currmovenumber 21
info depth 24 currmove f5e6 currmovenumber 22
info depth 24 currmove a5c4 currmovenumber 23
info depth 24 currmove d6c7 currmovenumber 24
info depth 24 currmove d6e7 currmovenumber 25
info depth 24 currmove h8h2 currmovenumber 26
info depth 24 currmove h8h5 currmovenumber 27
info depth 24 currmove a5b3 currmovenumber 28
info depth 24 currmove h8e8 currmovenumber 29
info depth 24 currmove h8h4 currmovenumber 30
info depth 24 currmove d8e8 currmovenumber 31
info depth 24 currmove h8f8 currmovenumber 32
info depth 24 currmove d6d4 currmovenumber 33
info depth 24 currmove c8c7 currmovenumber 34
info depth 24 currmove h8h7 currmovenumber 35
info depth 24 currmove f5g6 currmovenumber 36
info depth 24 currmove g7g5 currmovenumber 37
info depth 24 currmove d6b8 currmovenumber 38
info depth 24 currmove f5h7 currmovenumber 39
info depth 24 currmove h8g8 currmovenumber 40
info depth 24 currmove a7a6 currmovenumber 41
info depth 24 currmove c5a3 currmovenumber 42
info depth 24 currmove c5b4 currmovenumber 43
info depth 24 currmove d8g8 currmovenumber 44
info depth 24 currmove h8h6 currmovenumber 45
info depth 24 currmove d6d3 currmovenumber 46
info depth 35 seldepth 55 multipv 1 score cp 408 nodes 180919306 nps 1974885 hashfull 1000 tbhits 0 time 91610 pv d6g6 f1h1 h8h1 e1h1 c5e3 h1h4 d8f8 f2e3 f5e4 g2g1 e4f3 e2f3 g4f3 g1f2 a5b7 a1f1 g6e6 f1h1 e5e4 h4h3 c8d7 d3d4 b7d6 h3e6 d7e6 h1h5 g7g6 h5g5


Crafty-25.2 (16 MB hash and 1 thread) finds Qg6 in less than 10 seconds and in fewer than 66 million nodes. :D

Code: Select all

Crafty v25.2 (20 threads)

White(1): mt = 0
parallel threads disabled.
White(1): hash = 16M
hash table memory = 16M bytes (1M entries).
White(1): setb 2kr3r/p5p1/2pq4/n1b1pb2/4N1p1/3PB1P1/PPP1BPK1/R3QR2 b - - 0 20
Black(1): d

       +---+---+---+---+---+---+---+---+
    8  |   | . |<K>|<R>|   | . |   |<R>|
       +---+---+---+---+---+---+---+---+
    7  |<P>|   | . |   | . |   |<P>|   |
       +---+---+---+---+---+---+---+---+
    6  |   | . |<P>|<Q>|   | . |   | . |
       +---+---+---+---+---+---+---+---+
    5  |<N>|   |<B>|   |<P>|<B>| . |   |
       +---+---+---+---+---+---+---+---+
    4  |   | . |   | . |-N-| . |<P>| . |
       +---+---+---+---+---+---+---+---+
    3  | . |   | . |-P-|-B-|   |-P-|   |
       +---+---+---+---+---+---+---+---+
    2  |-P-|-P-|-P-| . |-B-|-P-|-K-| . |
       +---+---+---+---+---+---+---+---+
    1  |-R-|   | . |   |-Q-|-R-| . |   |
       +---+---+---+---+---+---+---+---+
         a   b   c   d   e   f   g   h

Black(1): st = 10
search time set to 10.00.
Black(1): go
        time surplus   0.00  time limit 10.00 (10.00)
        depth     time       score   variation (1)
         19     1.12/10.00    2.12   1. ... Bxe4+ 2. dxe4 Bxe3 3. Bxg4+ Kb7
                                     4. Qxe3 Qd4 5. Qxd4 Rxd4 6. Rfd1 Rhd8
                                     7. Rxd4 Rxd4 8. Rd1 Nc4 9. Rxd4 exd4
                                     10. b3 Ne5 11. Be6 c5 12. Bd5+ Kc7
         19->   1.27/10.00    2.12   1. ... Bxe4+ 2. dxe4 Bxe3 3. Bxg4+ Kb7
                                     4. Qxe3 Qd4 5. Qxd4 Rxd4 6. Rfd1 Rhd8
                                     7. Rxd4 Rxd4 8. Rd1 Nc4 9. Rxd4 exd4
                                     10. b3 Ne5 11. Be6 c5 12. Bd5+ Kc7
         20     2.00/10.00    2.18   1. ... Bxe4+ 2. dxe4 Bxe3 3. Bxg4+ Kc7
                                     4. Qxe3 Qd4 5. Qxd4 Rxd4 6. Rfd1 Rhd8
                                     7. b3 Rd2 8. Rxd2 Rxd2 9. c4 Kd6 10. Rd1
                                     Rxd1 11. Bxd1 Kc5 12. Bg4
         20->   2.16/10.00    2.18   1. ... Bxe4+ 2. dxe4 Bxe3 3. Bxg4+ Kc7
                                     4. Qxe3 Qd4 5. Qxd4 Rxd4 6. Rfd1 Rhd8
                                     7. b3 Rd2 8. Rxd2 Rxd2 9. c4 Kd6 10. Rd1
                                     Rxd1 11. Bxd1 Kc5 12. Bg4
         21     2.71/10.00    2.15   1. ... Bxe4+ 2. dxe4 Bxe3 3. Bxg4+ Kc7
                                     4. Qxe3 Qd4 5. Qxd4 Rxd4 6. Rfd1 Rhd8
                                     7. b3 Rd2 8. Rxd2 Rxd2 9. c4 Kd6 10. Rd1
                                     Rxd1 11. Bxd1 Kc5 12. Bg4 Kd4
         21->   2.85/10.00    2.15   1. ... Bxe4+ 2. dxe4 Bxe3 3. Bxg4+ Kc7
                                     4. Qxe3 Qd4 5. Qxd4 Rxd4 6. Rfd1 Rhd8
                                     7. b3 Rd2 8. Rxd2 Rxd2 9. c4 Kd6 10. Rd1
                                     Rxd1 11. Bxd1 Kc5 12. Bg4 Kd4
         22     4.57/10.00    2.19   1. ... Bxe4+ 2. dxe4 Bxe3 3. Bxg4+ Kc7
                                     4. Qxe3 Qd4 5. Qxd4 Rxd4 6. Rfd1 Nc4 7. c3
                                     Rdd8 8. b3 Nd6 9. Kf3 Nb5 10. Rac1 Rxd1
                                     11. Rxd1 Nxc3 12. Rd7+ Kb6 13. a3 g5
        time=10.09(100%)  nodes=65877778(65.9M)  fh1=87%  pred=0  nps=6.5M
        chk=1.6M  qchk=2.6M  fp=23.7M  mcp=7.4M  50move=1
        LMReductions:  1/1.2M  2/764.3K  3/662.7K  4/430.3K  5/63.9K  6/2.2K
        null-move (R):  3/2.0M  4/206.2K  5/9.1K  6/178
Black(1): Qg6
              time used:  10.09

zullil
Posts: 5668
Joined: Mon Jan 08, 2007 11:31 pm
Location: PA USA
Full name: Louis Zulli

Re: One more issue with king safety/attack for Stockfish

Post by zullil » Wed Jul 17, 2019 12:35 am

Vinvin wrote:
Tue Jul 16, 2019 11:28 pm
From the game Vitiugov - Grischuk (Riga Grand Prix 2019).

https://www.chessbomb.com/arena/2019-ri ... _Alexander

MikeB
Posts: 3467
Joined: Thu Mar 09, 2006 5:34 am
Location: Pen Argyl, Pennsylvania

Re: One more issue with king safety/attack for Stockfish

Post by MikeB » Wed Jul 17, 2019 12:41 am

You are misreading the output, SF is considering Bxe4 with a huge disadvantage - notice how the sign flips when it sees Qg6 ( on your output) .

dev Burnzie is no different , still takes about a billion nodes... lc0 like score % output ( note: there is no chance to misread the output when using lc0 like output- it's always from the side to move)

Code: Select all

dep	score	nodes	time	(not shown:  tbhits	knps	seldep)
 36	+88.65 	3.41G	2:20.30	Qg6 Rh1 Rxh1 Qxh1 Bxe3 Qh4 Bb6 Kg1 Nb7 a4 Rf8 Rf1 Bd8 Qh1 Nd6 Nxd6+ Qxd6 b4 Kb8 a5 Qxb4 Qxc6 Bxa5 Qd5 Bc7 c4 Qb3 Ra1 Qb2 Ra5 Qxe2 Rb5+ Bb6 Rxb6+ axb6 Qd6+ Kb7 Qxf8 Bxd3 Qe7+ Kc6 Qe6+ Kc5 Qd5+ Kb4 Qb5+ Kc3 Qxb6 
 36	+88.90!	2.94G	2:00.98	Qg6! 
 36	+88.28?	2.71G	1:51.29	Qg6 Rh1? 
 36	+88.55?	2.68G	1:49.80	Qg6 Rh1? 
 35	+88.82 	2.28G	1:33.63	Qg6 Rh1 Rxh1 Qxh1 Bxe3 Qh4 Bb6 Ng5 Rd4 c3 Rd7 Bxg4 Bd8 f4 exf4 gxf4 Rxd3 Qh5 Qxh5 Bxh5 Nc4 Bf7 Nxb2 Be6+ Bxe6 Nxe6 Bf6 Rh1 Rxc3 Rh7 Nd3 f5 Rc2+ Kg3 Re2 Kf3 Re1 Rh8+ Kb7 Rg8 Ne5+ Kf2 Ra1 Nxg7 Rxa2+ Kf1 Nd3 Ne6 Rf2+ Kg1 Rxf5 Rg3  
 35	+87.94!	2.09G	1:25.90	Qg6! 
 34	+86.27 	1.70G	1:09.37	Qg6 Rh1 Rxh1 Qxh1 Bxe3 Kg1 Bb6 a4 Nb7 b4 Nd6 a5 Bd4 c3 Bxe4 dxe4 Bxc3 Rc1 Nxe4 Qh4 Kc7 Qxg4 Qxg4 Bxg4 Bxb4 Rc2 Rd2 Rxd2 Nxd2 f4 exf4 gxf4 Bxa5 Kf2 Bb6+ Ke2 Nc4 Be6 Ne3 Bb3 Nd5 f5 a5 Kd2 
 34	+85.44!	1.51G	1:01.65	Qg6! 
 33	+83.70 	1.33G	0:54.04	Qg6 Rh1 Rxh1 Qxh1 Bxe3 Kg1 Bb6 a4 Nb7 Nd2 Rf8 Rf1 Kb8 Nc4 Qe8 Nxb6 Rh8 Qg2 axb6 f3 Qg6 Kf2 Be6 Ke1 Bd5 Qf2 gxf3 Bxf3 Rf8 Qg2 Qg5 Bxd5 Qc1+ Ke2 Rxf1 Qxf1 Qxc2+ Kf3 cxd5 Ke3 Nc5 Qf8+ Ka7 Qxg7+ Ka6 Qxe5 
 33	+82.21!	1.28G	0:51.75	Qg6! 
 33	+80.15!	1.25G	0:50.52	Qg6! 
 33	+78.47!	1.22G	0:49.64	Qg6! 
 32	+76.70 	1.10G	0:44.42	Qg6 Rh1 Rxh1 Qxh1 Bxe3 Kg1 Bb6 b4 Nc4 Qh4 Nd6 Nc5 Nb7 Na4 Be6 Nxb6+ axb6 a4 Nd6 Qe7 Qf6 Qxf6 gxf6 Rd1 
 32	+66.92!	1.08G	0:43.90	Qg6! 
 32	+54.96!	1.08G	0:43.81	Qg6! 
 32	+45.04!	1.08G	0:43.76	Qg6! 
 32	+37.48!	1.08G	0:43.72	Qg6! 
 32	+31.96!	1.08G	0:43.68	Qg6! 
 32	+28.00!	1.08G	0:43.66	Qg6! 
 32	+25.17!	1.08G	0:43.65	Qg6! 
 32	+23.15!	1.08G	0:43.63	Qg6! 
 32	+21.72!	1.08G	0:43.62	Qg6! 
 32	+20.73!	1.08G	0:43.59	Qg6! 
 32	+20.06!	1.07G	0:43.44	Qg6! 
 32	+19.62!	927.7M	0:37.54	Bxe4+! 
 31	+19.20 	909.2M	0:36.81	Bxe4+ dxe4 Nb7 Bxg4+ Kc7 Rd1 Qe7 Rxd8 Rxd8 Rh1 a5 Rh5 Rf8 Qc3 Bxe3 Qxe3 Re8 Rg5 Qf6 Bh5 Rf8 Rg6 Qe7 Bg4 Rf6 Rg5 Rf8 f4 exf4 gxf4 Qf6 Rf5 Qe7 Rxf8 Qxf8 e5 Qc5 Qe4 Qd5 Qxd5 cxd5 f5 
 31	+19.05!	900.2M	0:36.44	Bxe4+! 
 30	+18.65 	832.7M	0:33.69	Bxe4+ dxe4 Bb6 Bxg4+ Kb8 Rd1 Qg6 Rxd8+ Rxd8 Bf5 Qf6 Rh1 Nb7 Rh5 Rg8 Qc3 Bxe3 Qxe3 Rf8 Qc3 c5 Qe3 a6 a3 Qe7 Rh7 Ka7 b4 Rd8 Bg4 Kb8 Qc3 Rf8 Rh5 
 30	+18.82!	823.6M	0:33.32	Bxe4+! 
 29	+18.43 	621.3M	0:25.19	Bxe4+ dxe4 Bb6 Bxg4+ Kb7 Qc3 Bxe3 Qxe3 Ka8 Rad1 Qf6 Rxd8+ Rxd8 Rh1 Nb7 Qc3 Rf8 Bf5 c5 Rh5 Kb8 Qe3 a5 Rh7 Re8 a3 Rd8 Qc3 Qe7 Qb3 Qf6 Qe3 Ka7 b4 Qe7 bxa5 Nxa5 Qe2 
 29	+18.82?	484.1M	0:19.70	Bxe4+ dxe4? 
 28	+20.62 	365.2M	0:14.89	Bxe4+ dxe4 Bb6 Bxg4+ Kb7 Qc3 Qg6 Bf5 Qf7 b3 c5 Qxe5 Nc6 Qe6 Qf6 Qxf6 gxf6 Rh1 c4 Bxb6 Rxh1 Rxh1 Kxb6 bxc4 Kc5 Rh7 Ne5 f4 Nxc4 Rc7+ Kd4 g4 Nd6 Kf3 a5 Ra7 Nxf5 exf5 
 28	+20.46!	352.2M	0:14.36	Bxe4+! 
 28	+19.80!	346.9M	0:14.14	Bxe4+! 
 28	+19.38!	346.3M	0:14.11	Bxe4+! 
 27	+18.98 	318.0M	0:12.95	Bxe4+ dxe4 Bb6 Bxg4+ Kb8 Rd1 Qc7 Qc3 Bxe3 Qxe3 Nb7 Qc3 Rd6 Rxd6 Qxd6 Rd1 Qf6 Bf5 Rd8 Rh1 Rf8 Rh7 c5 Rh5 Ka8 Qe3 Kb8 Bg4 Kc7 Be2 Kb8 Rh7 Qg6 Rh1 Qb6 Qg5 Qf6 Qxf6 Rxf6 Rh8+ Kc7 Re8 
 27	+20.92!	222.8M	0:09.05	Bxe4+! 
 26	+20.49 	173.7M	0:07.04	Bxe4+ dxe4 Bxe3 Bxg4+ Kb8 Qxe3 Nb7 Rad1 Qf6 Rxd8+ Rxd8 Rh1 Nd6 Qc5 Nb7 Qc3 Rf8 Bf5 Nd6 Rh5 Kb7 Qb4+ Ka8 Qc5 Nb7 Qc4 Kb8 Qc3 c5 Qe3 a5 Bg4 Ka7 Rf5 Qd6 Rxf8 Qxf8 Qc3 Qd6 Be2 
 25	+20.95 	125.6M	0:05.12	Bxe4+ dxe4 Bxe3 Bxg4+ Kb8 Qxe3 Nb7 Rad1 Qg6 Bf5 Qf6 Rh1 Rxh1 Rxh1 Nd6 Rh5 Ka8 Qc5 Nf7 Bg4 Rd2 Rf5 Qd6 Qxd6 Nxd6 Rxe5 Nc4 Rg5 Rxc2 e5 Rxb2 e6 Ne3+ Kf3 Nxg4 Rxg4 Rb7 
 24	+21.55 	84.1M  	0:03.46	Bxe4+ dxe4 Bxe3 Bxg4+ Kb8 Qxe3 Qg6 Bf5 Qf6 Rh1 Nb7 Rxh8 Rxh8 Rd1 Rd8 Rxd8+ Qxd8 Qc3 Qf6 a4 a5 b4 axb4 Qxb4 Ka7 a5 Qd8 Qc4 Qc7 a6 Nd6 Qc5+ Kxa6 Qxe5 
 24	+22.14!	82.8M  	0:03.40	Bxe4+! 
 24	+21.69!	82.6M  	0:03.40	Bxe4+! 
 23	+21.25 	79.2M  	0:03.26	Bxe4+ dxe4 Bxe3 Bxg4+ Kb8 Qxe3 Qg6 Bf5 Qf6 Rh1 Nb7 Rxh8 Rxh8 Rd1 Rd8 Rh1 Nd6 Rh5 Nb5 Qb3 Ka8 Qc4 Nd4 c3 Nxf5 exf5 Rf8 Qe6 Kb7 g4 Kc7 Qxf6 gxf6 Rh7+ Kb6 Kg3 a5 
 23	+21.66!	77.2M  	0:03.18	Bxe4+! 
 22	+21.22 	66.5M  	0:02.74	Bxe4+ dxe4 Bb6 Bxg4+ Kb7 Qe2 Kb8 Rh1 Rxh1 Rxh1 Nb7 Bg5 Rf8 f3 Qc5 c3 a5 Qd2 Qb5 a4 Qb3 Qe2 Ka7 Bd7 Qxa4 Rd1 Qb5 Qxb5 cxb5 Bxb5 Rh8 
 21	+25.42 	12.9M  	0:00.52	Bxe4+ dxe4 Bxe3 Bxg4+ Kb8 Qxe3 Qd4 Qe2 Nb7 Rad1 Qa4 Rxd8+ Rxd8 a3 Qb5 Qxb5 cxb5 f4 Nd6 fxe5 Nc4 Rf2 Re8 e6 Ne3+ Kf3 Nxg4 Kxg4 Rxe6 Rf8+ Kc7 Rf7+ Kd6 
 20	+24.50 	11.3M  	0:00.46	Bxe4+ dxe4 Bxe3 Bxg4+ Kc7 Qxe3 Qd4 Qf3 Rdf8 Qe2 Qc4 Qe3 Qd4 Qe1 Nb7 Rd1 Qxb2 Rd7+ Kb8 Rxg7 Rhg8 Rxg8 Rxg8 Bf5 Qxc2 Rh1 Rf8 Rh7 Qxa2 
 19	+26.21 	7.33M  	0:00.30	Bxe4+ dxe4 Bxe3 Bxg4+ Kb8 Qxe3 Qd4 Qe2 Qc4 Rad1 Qxe2 Bxe2 Nb7 f4 Rde8 Kf3 Nc5 f5 Kc7 b4 Nd7 Rd3 Rh2 Ra3 Kb8 Bd3 Reh8 Ra6 Kb7 Ra5 Ka8 c3 Kb7 Rf2 
 18	+25.26 	7.11M  	0:00.29	Bxe4+ dxe4 Bxe3 Bxg4+ Kc7 Qxe3 Qd4 Qf3 Rhf8 Bf5 Nc4 Qg4 Kb6 Rad1 Qc5 c3 Qb5 b4 Rd2 Rxd2 Nxd2 Rd1 Qd3 a3 
 17	+26.34 	5.52M  	0:00.22	Bxe4+ dxe4 Bxe3 Bxg4+ Kb8 Qxe3 Qd4 Qe2 Qd2 Rad1 Qxe2 Bxe2 Nb7 Rxd8+ Rxd8 Kf3 Rh8 Ke3 Kc7 f4 Rh3 fxe5 Nd8 Rf3 Kd7 g4 Rh2 Rf2 Rxf2 Kxf2 
 16	+29.24 	2.89M  	0:00.12	Bxe4+ dxe4 Bxe3 Bxg4+ Kc7 Qxe3 Qd4 Qe2 Qc4 Rad1 Qxe2 Bxe2 Nb7 f4 Rxd1 Rxd1 Nc5 Kf3 a5 fxe5 Rh5 e6 Nxe6 Ke3 
 15	+25.70 	2.60M  	0:00.11	Bxe4+ dxe4 Bxe3 Bxg4+ Kc7 Qxe3 Qd4 Qf3 Rdf8 Qe2 Qc4 Rad1 Qxe2 Bxe2 Nb7 f4 exf4 Rxf4 Rxf4 gxf4 Nd6 Bd3 Nc8 
 14	+28.16 	1.39M  	0:00.06	Bxe4+ dxe4 Bxe3 Bxg4+ Kc7 Qxe3 Qd4 Qe2 Qc4 Qf3 Rdf8 Qe3 Qd4 Qxd4 exd4 Be2 Re8 b4 Nb7 Rad1 
 13	+27.96 	864632	0:00.04	Bxe4+ dxe4 Bxe3 Bxg4+ Kc7 Qxe3 Qd4 Qe2 Qc4 Rad1 Qxe2 Bxe2 Rxd1 Rxd1 Nb7 b4 Nd6 f3 Nf7 a3 Ng5 c3 Rh6 a4 Rd6 Rxd6 Kxd6 
 12	+28.16 	544579	0:00.02	Bxe4+ dxe4 Bxe3 Bxg4+ Kc7 Qxe3 Qd4 Qe2 Qc4 Rad1 Qxe2 Bxe2 c5 Rfe1 Rd4 b3 Nc6 c3 Rxd1 Rxd1 
 11	+28.83 	371298	0:00.02	Bxe4+ dxe4 Bxe3 Bxg4+ Kc7 Qxe3 Qd4 Qe2 Qc4 Rh1 Qxe2 Bxe2 Rxh1 Rxh1 Rd2 
 10	+28.77 	260788	0:00.01	Bxe4+ dxe4 Bxe3 Bxg4+ Kc7 Qxe3 Qd4 Qxd4 Rxd4 
  9	+25.36 	160461	0:00.01	Bxe4+ dxe4 Bxe3 Bxg4+ Kc7 Qxe3 Qd4 Qxd4 exd4 e5 c5 Rfe1 Nc6 e6 
  8	+26.43 	75995  	0:00.00	Bxe4+ dxe4 Bxe3 Bxg4+ Kb8 Qxe3 Qd4 Rad1 Qxe3 Rxd8+ Rxd8 fxe3 
  7	+25.08 	70498  	0:00.00	Bxe4+ dxe4 Bxe3 Bxg4+ Kb8 Qxe3 Qd4 Rad1 Qxe3 fxe3 
  6	+27.11 	48003  	0:00.00	Bxe4+ dxe4 Bxe3 Bxg4+ Kb8 Qxe3 Qb4 a3 Qd2 Qxd2 
  5	+33.52 	30297  	0:00.00	Qd5 Bxc5 Bxe4+ dxe4 Qxc5 b4 
  4	+65.55 	10727  	0:00.00	Bxe4+ dxe4 Bb4 Bxg4+ Kb8 Bxa7+ Kxa7 
  3	+32.79 	6583    	0:00.00	Bxe4+ dxe4 Nb7 Bxg4+ Kb8 
  2	+29.93 	4400    	0:00.00	Bxe4+ dxe4 Bxe3 
  1	+31.22 	2301    	0:00.00	Bxe4+ dxe4 
  0	# 

zullil
Posts: 5668
Joined: Mon Jan 08, 2007 11:31 pm
Location: PA USA
Full name: Louis Zulli

Re: One more issue with king safety/attack for Stockfish

Post by zullil » Wed Jul 17, 2019 12:57 am

MikeB wrote:
Wed Jul 17, 2019 12:41 am
You are misreading the output, SF is considering Bxe4 with a huge disadvantage - notice how the sign flips when it sees Qg6 ( on your output) .
No, he's not misreading. He's referring to Stockfish's search of the position just prior to the one he posted.

Point your engine at this:


MikeB
Posts: 3467
Joined: Thu Mar 09, 2006 5:34 am
Location: Pen Argyl, Pennsylvania

Re: One more issue with king safety/attack for Stockfish

Post by MikeB » Wed Jul 17, 2019 1:03 am

zullil wrote:
Wed Jul 17, 2019 12:57 am
MikeB wrote:
Wed Jul 17, 2019 12:41 am
You are misreading the output, SF is considering Bxe4 with a huge disadvantage - notice how the sign flips when it sees Qg6 ( on your output) .
No, he's not misreading. He's referring to Stockfish's search of the position just prior to the one he posted.

Point your engine at this:

Correct - his comment was the move before. My bad.

zullil
Posts: 5668
Joined: Mon Jan 08, 2007 11:31 pm
Location: PA USA
Full name: Louis Zulli

Re: One more issue with king safety/attack for Stockfish

Post by zullil » Wed Jul 17, 2019 1:07 am

MikeB wrote:
Wed Jul 17, 2019 1:03 am
zullil wrote:
Wed Jul 17, 2019 12:57 am
MikeB wrote:
Wed Jul 17, 2019 12:41 am
You are misreading the output, SF is considering Bxe4 with a huge disadvantage - notice how the sign flips when it sees Qg6 ( on your output) .
No, he's not misreading. He's referring to Stockfish's search of the position just prior to the one he posted.

Point your engine at this:

Correct - his comment was the move before. My bad.
And Stockfish-dev gets to depth 38 believing that 20. Nxe4 is a winner (+4.01) rather than the loser that it is!

info depth 38 seldepth 61 multipv 1 score cp 401 nodes 480162938 nps 2101158 hashfull 1000 tbhits 0 time 228523 pv c3e4 f5e4 d3e4 c5b6 e2g4 c8b7 e1e2 d6g6 g4f5 g6h5 e2h5 h8h5 g3g4 h5h8 e3g5 d8f8 a1d1 g7g6 d1d7 b7a6 f5g6 h8g8 g5e7 g8g6 e7f8 g6g4 g2h3 g4e4 f8h6 e4c4 c2c3 c4a4 a2a3 a4e4 f2f3 e4e2 d7d2 e2d2 h6d2 a5c4 d2c1 c4d6 h3g4 a6b5 b2b3

Something does seem amiss with the fish.

MikeB
Posts: 3467
Joined: Thu Mar 09, 2006 5:34 am
Location: Pen Argyl, Pennsylvania

Re: One more issue with king safety/attack for Stockfish

Post by MikeB » Wed Jul 17, 2019 1:33 am

zullil wrote:
Wed Jul 17, 2019 1:07 am
MikeB wrote:
Wed Jul 17, 2019 1:03 am
zullil wrote:
Wed Jul 17, 2019 12:57 am
MikeB wrote:
Wed Jul 17, 2019 12:41 am
You are misreading the output, SF is considering Bxe4 with a huge disadvantage - notice how the sign flips when it sees Qg6 ( on your output) .
No, he's not misreading. He's referring to Stockfish's search of the position just prior to the one he posted.

Point your engine at this:

Correct - his comment was the move before. My bad.
And Stockfish-dev gets to depth 38 believing that 20. Nxe4 is a winner (+4.01) rather than the loser that it is!

info depth 38 seldepth 61 multipv 1 score cp 401 nodes 480162938 nps 2101158 hashfull 1000 tbhits 0 time 228523 pv c3e4 f5e4 d3e4 c5b6 e2g4 c8b7 e1e2 d6g6 g4f5 g6h5 e2h5 h8h5 g3g4 h5h8 e3g5 d8f8 a1d1 g7g6 d1d7 b7a6 f5g6 h8g8 g5e7 g8g6 e7f8 g6g4 g2h3 g4e4 f8h6 e4c4 c2c3 c4a4 a2a3 a4e4 f2f3 e4e2 d7d2 e2d2 h6d2 a5c4 d2c1 c4d6 h3g4 a6b5 b2b3

Something does seem amiss with the fish.

dev-Burnzie - single thread - 1GB hash, turns slightly negative at the 700M node mark

Code: Select all

dep	score	nodes	time	(not shown:  tbhits	knps	seldep)
 36	+49.96 	714.7M	7:31.45	Bg5 exd3 Bxd3 Rdf8 Rd1 Bd4 Qe2 Qg6 Bh4 Nb7 Ne4 Kb8 c3 Bb6 b4 Bd8 Rh1 Bb6 
 36	+51.85!	664.5M	6:58.36	Bg5! 
 36	+51.03?	628.7M	6:34.79	Bg5 exd3? 
 36	+52.06?	620.5M	6:29.63	Bg5 exd3? 
 36	+52.71?	613.0M	6:24.87	Bg5 exd3? 
 35	+53.37 	597.8M	6:14.93	Bg5 exd3 Bxd3 Rdf8 Rd1 Bd4 Qe2 Qg6 Bxf5+ Qxf5 Ne4 Nb7 c3 Bb6 Be7 Qf3+ Qxf3 gxf3+ Kg1 Rf5 Bh4 Rfh5 Rfe1 Bd8 Re3 Bxh4 gxh4 Rxh4 Ng3 e4 Rxe4 Rh2 Re8+ Kc7 Re7+ Kb6 Rxg7 Rg2+ Kf1 Rhh2 Rd2 a5 Rg8 Ka7 b4 Rh1+ Nxh1 Rxg8 
 35	+53.45?	359.6M	3:43.07	Nxe4 Qg6? 
 35	+61.09?	343.7M	3:33.18	Nxe4 Qg6? 
 35	+66.70?	336.9M	3:28.91	Nxe4 Qg6? 
 35	+70.76?	331.3M	3:25.40	Nxe4 Qg6? 
 35	+73.66?	325.1M	3:21.52	Nxe4 Qg6? 
 35	+75.75?	317.3M	3:16.74	Nxe4 Qg6? 
 35	+77.23?	314.6M	3:15.08	Nxe4 Qg6? 
 35	+78.25?	312.3M	3:13.68	Nxe4 Qg6? 
 35	+78.94?	309.8M	3:12.19	Nxe4 Qg6? 
 35	+79.38?	280.5M	2:53.98	Nxe4 Bxe4+? 
 34	+79.81 	230.4M	2:22.76	Nxe4 Bxe4+ dxe4 Bb6 Bxg4+ Kb7 Qc3 Qg6 Bf5 Qf7 b3 Bxe3 Qxe3 Ka8 Qc3 Qc7 Rh1 Rxh1 Rxh1 Nb7 Rh7 Kb8 Qe3 Nd6 Qc5 Nb7 Qc3 Re8 Bg6 Rf8 b4 Rf6 Bf5 Nd8 Rh1 Rh6 Rxh6 gxh6 a4 Kb7 
 34	+79.73!	198.6M	2:03.20	Nxe4! 
 34	+79.00?	181.3M	1:52.44	Nxe4 Bxe4+? 
 34	+79.67?	179.9M	1:51.52	Nxe4 Bxe4+? 
 34	+80.10?	163.8M	1:41.74	Nxe4 Bxe4+? 
 33	+80.51 	152.0M	1:34.38	Nxe4 Bxe4+ dxe4 Bb6 Bxg4+ Kb7 Qc3 Bxe3 Qxe3 Nc4 Qb3+ Nb6 Rad1 Qf6 Rxd8 Rxd8 a4 a5 Rh1 Qg6 Bf3 Qd6 Rh7 Qe7 Bg4 Ra8 Bf5 Kc7 Qf3 Nd7 Bxd7 Qxd7 Qc3 Kb6 Qxe5 Re8 Rxg7 Rxe5 Rxd7 Rxe4 b3 c5 c3 Re6 
 33	+79.78!	130.3M	1:20.91	Nxe4! 
 33	+79.00?	118.6M	1:13.79	Nxe4 Bxe4+? 
 33	+79.43?	117.0M	1:12.83	Nxe4 Bxe4+? 
 32	+79.86 	114.4M	1:11.18	Nxe4 Bxe4+ dxe4 Bb6 Bxg4+ Kb7 Qc3 Bxe3 Qxe3 Nc4 Qb3+ Nb6 Rad1 Qe7 Rxd8 Rxd8 Rh1 Rf8 Qc3 Qf7 Qc5 Nd7 Qb4+ Nb6 Rf1 Rd8 Rd1 Rxd1 Bxd1 Qd7 Be2 Kc7 Qa5 Kd6 Qc3 Qc7 Qd3+ Ke7 Qe3 Qd6 Bg4 Kf7 
 32	+79.81!	101.9M	1:03.44	Nxe4! 
 32	+79.03?	96.1M  	0:59.94	Nxe4 Bxe4+? 
 32	+79.46?	95.6M  	0:59.63	Nxe4 Bxe4+? 
 31	+79.89 	91.0M  	0:56.75	Nxe4 Bxe4+ dxe4 Bb6 Bxg4+ Kb8 Rd1 Qg6 Rxd8+ Rxd8 Bf5 Qf6 Rh1 Nb7 Rh5 Qf7 Rg5 Qe7 Qc3 Bxe3 Qxe3 a5 Rh5 Kc7 Bg4 Rf8 Qa7 Qc5 Qxc5 Nxc5 Rxe5 Kd6 f4 Rh8 Rg5 Nxe4 Rxg7 Nf6 Bf3 a4 
 31	+78.45!	79.5M  	0:49.59	Nxe4! 
 31	+77.69?	78.4M  	0:48.94	Nxe4 Bxe4+? 
 31	+78.39?	76.6M  	0:47.84	Nxe4 Bxe4+? 
 31	+78.83?	76.3M  	0:47.64	Nxe4 Bxe4+? 
 30	+79.27 	71.0M  	0:44.24	Nxe4 Bxe4+ dxe4 Bb6 Bxg4+ Kb8 Rd1 Qg6 Rxd8+ Rxd8 Bf5 Qd6 Rh1 Nc4 Bg5 Rf8 b3 Na3 c3 Qc7 Rh7 a5 Bh6 Rf7 Rh8+ Kb7 Bc1 Nb5 Be3 Nd6 Bg4 Bxe3 Qxe3 Re7 f3 c5 Qd3 Ka7 a4 Re8 Rxe8 Nxe8 Qd7 Qxd7 Bxd7 
 29	+78.34 	55.9M  	0:34.80	Nxe4 Bxe4+ dxe4 Bb6 Bxg4+ Kb8 Rd1 Qg6 Rxd8+ Rxd8 Bf5 Qf7 Qc3 Nc4 Bxb6 axb6 b3 Na3 Qxe5+ Qc7 Qxc7+ Kxc7 e5 Rd2 c4 Kd8 Rh1 Nc2 Bxc2 Rxc2 Rh7 Rxa2 Rxg7 Re2 g4 b5 cxb5 cxb5 Kg3 Rxe5 g5 Ke8 Kg4 Rd5 
 29	+79.27!	54.8M  	0:34.15	Nxe4! 
 29	+78.47?	51.9M  	0:32.32	Nxe4 Bxe4+? 
 29	+78.92?	48.6M  	0:30.23	Nxe4 Bxe4+? 
 28	+79.35 	47.1M  	0:29.32	Nxe4 Bxe4+ dxe4 Bxe3 Bxg4+ Kb7 Qxe3 Nc4 Qb3+ Nb6 Rad1 Qf6 Rxd8 Rxd8 Rh1 Rf8 Bf5 Qf7 Rh7 Qxb3 cxb3 Rf7 Rh8 Re7 g4 Kc7 Kf3 Kd6 g5 Kc7 Kg3 Kd6 Bg4 c5 g6 
 28	+79.51!	39.4M  	0:24.55	Nxe4! 
 28	+77.17?	33.4M  	0:20.88	Nxe4 Bxe4+? 
 28	+78.67!	31.6M  	0:19.72	Nxe4! 
 28	+77.89?	29.0M  	0:18.15	Nxe4 Bxe4+? 
 28	+78.34?	28.6M  	0:17.92	Nxe4 Bxe4+? 
 27	+78.78 	27.0M  	0:16.89	Nxe4 Bxe4+ dxe4 Bb6 Bxg4+ Kb7 Qc3 Qg6 Bf5 Qf7 b3 c5 Qxe5 Nc6 Qe6 Qf6 Qxf6 gxf6 Rh1 Nd4 Rxh8 Rxh8 Rd1 Rd8 Rh1 Nxc2 Rh7+ Kc6 Bf4 Nd4 Rf7 Nxf5 exf5 c4 Rxf6+ Kd5 bxc4+ Kxc4 Rc6+ Kb5 
 27	+78.89!	24.6M  	0:15.32	Nxe4! 
 27	+76.52?	21.5M  	0:13.35	Nxe4 Bxe4+? 
 27	+78.06!	21.1M  	0:13.11	Nxe4! 
 27	+77.26?	19.0M  	0:11.83	Nxe4 Bxe4+? 
 27	+77.71?	18.9M  	0:11.78	Nxe4 Bxe4+? 
 26	+78.17 	17.6M  	0:10.93	Nxe4 Bxe4+ dxe4 Bb6 Bxg4+ Kb8 Rd1 Qg6 Rxd8+ Rxd8 Bf5 Qf6 Qc3 Nb7 Rh1 Bxe3 Qxe3 Nd6 Rh5 Ka8 Qc5 Nf7 Bg4 Qd6 Qxd6 Rxd6 Rh7 Nh6 Bf3 Rd2 Rxg7 Rxc2 b4 
 26	+76.94!	16.4M  	0:10.22	Nxe4! 
 26	+76.14?	15.5M  	0:09.65	Nxe4 Bxe4+? 
 26	+76.61?	15.2M  	0:09.47	Nxe4 Bxe4+? 
 25	+77.08 	12.6M  	0:07.81	Nxe4 Bxe4+ dxe4 Bb6 Bxg4+ Kb8 Qc3 Bxe3 Qxe3 Nb7 Rad1 Qg6 Bf5 Qf6 Rxd8+ Rxd8 Rh1 Nd6 Bg4 Qg6 Bf3 Nb5 c3 Qf6 Bg4 Qe7 Rd1 Rxd1 Bxd1 Kc7 f4 a5 Bg4 exf4 gxf4 
 25	+76.55!	10.3M  	0:06.41	Nxe4! 
 25	+75.78?	8.65M  	0:05.35	Nxe4 Bxe4+? 
 25	+76.52?	8.60M  	0:05.32	Nxe4 Bxe4+? 
 25	+76.99?	8.54M  	0:05.28	Nxe4 Bxe4+? 
 24	+77.46 	8.30M  	0:05.13	Nxe4 Bxe4+ dxe4 Bb6 Bxg4+ Kb8 Rd1 Qg6 Bf5 Rxd1 Qxd1 Qf6 Qe2 Nb7 Rd1 Rd8 Rxd8+ Qxd8 Bxb6 axb6 Qc4 Qc7 Qg8+ Ka7 Qe8 Nd8 Bg4 Kb8 b4 Ka7 Qg6 Nb7 a3 
 24	+76.52!	6.55M  	0:04.06	Nxe4! 
 24	+74.89!	5.82M  	0:03.61	Nxe4! 
 23	+73.18 	3.75M  	0:02.34	Nxe4 Bxe4+ dxe4 Bxe3 Bxg4+ Kb8 Qxe3 Qd4 Qe2 Qc4 Rad1 Qxe2 Bxe2 Nb7 f4 Nc5 fxe5 Rde8 Rf7 Rxe5 Rxg7 Nxe4 Bf3 Ng5 Bg4 Rc5 Rd2 Ne4 Re2 Nf6 Be6 Nd5 Rf7 Re8 Rh7 Rb5 
 22	+74.24 	2.94M  	0:01.85	Nxe4 Bxe4+ dxe4 Bxe3 Bxg4+ Kb8 Qxe3 Qd4 Qe2 Qc4 Rad1 Qxe2 Bxe2 Nb7 f4 Nc5 fxe5 Rxd1 Bxd1 Re8 Rf7 Rxe5 Bf3 a5 g4 Nxe4 Rd7 Ng5 Rxg7 Nxf3 Kxf3 
 21	+74.70 	1.97M  	0:01.24	Nxe4 Bxe4+ dxe4 Bxe3 Bxg4+ Kb8 Qxe3 Qd4 Qe2 Qd2 Rad1 Qxe2 Bxe2 Nb7 f4 Rxd1 Rxd1 Re8 Bf3 Kc7 Rh1 Nd6 Rh7 Re7 f5 Rd7 g4 c5 Kg3 Kc6 g5 Rb7 
 20	+72.40 	1.27M  	0:00.81	Nxe4 Bxe4+ dxe4 Bxe3 Bxg4+ Kb8 Qxe3 Qd4 Qe2 Qd2 Rad1 Qxe2 Bxe2 Nb7 f4 Rxd1 Rxd1 Re8 Bf3 Kc7 Rh1 Rd8 Rh7 Rd2+ Kf1 Rxc2 Rxg7+ Kb6 f5 Nd6 f6 Rc1+ Ke2 Rc2+ Kd3 Rxb2 
 19	+73.98 	991383	0:00.64	Nxe4 Bxe4+ dxe4 Bxe3 Bxg4+ Kb8 Qxe3 Qd4 Qe2 Qd2 Rh1 Qxe2 Bxe2 Nb7 Rxh8 Rxh8 f4 Kc7 fxe5 Re8 Rf1 Rxe5 Rf7+ Kb6 Rxg7 Nd6 Bd3 Nxe4 
 18	+72.63 	519469	0:00.34	Nxe4 Bxe4+ dxe4 Bxe3 Bxg4+ Kb8 Qxe3 Qd4 Qe2 Qd2 Rad1 Qxe2 Bxe2 Nb7 f4 exf4 gxf4 Rxd1 Rxd1 Nc5 e5 Rf8 Kg3 Ne6 Bg4 Nxf4 
 17	+72.63 	385565	0:00.25	Nxe4 Bxe4+ dxe4 Bxe3 Bxg4+ Kb8 Qxe3 Qd4 Qe2 Qd2 Rad1 Qxe2 Bxe2 Nb7 f4 exf4 gxf4 Rxd1 Rxd1 Nc5 e5 Rf8 
 16	+71.44 	253716	0:00.17	Nxe4 Bxe4+ dxe4 Bxe3 Bxg4+ Kb8 Qxe3 Qh6 Qxh6 Rxh6 Rad1 Rdh8 Kf3 Nc4 Rd7 Rf6+ Ke2 g6 Bf3 a5 Rh1 Rxh1 Bxh1 Nxb2 
 15	+71.17 	189562	0:00.13	Nxe4 Bxe4+ dxe4 Bxe3 Bxg4+ Kb8 Qxe3 Qd4 Qe2 Qd2 Rad1 Qxe2 Bxe2 Nb7 Rh1 Rxh1 Rxd8+ Nxd8 Kxh1 Ne6 Bg4 Ng5 
 14	+71.54 	139099	0:00.09	Nxe4 Bxe4+ dxe4 Bxe3 Bxg4+ Kb8 Qxe3 Qd4 Rfe1 Qxe3 Rxe3 Nb7 Rc3 c5 Rd1 Rxd1 Bxd1 
 13	+70.35 	80849  	0:00.06	Nxe4 Bxe4+ dxe4 Bxe3 Bxg4+ Kc7 Qxe3 Qd4 Qxd4 Rxd4 Rad1 Rhd8 b3 Rxd1 Rxd1 Rxd1 Bxd1 Nb7 f4 Nd6 fxe5 
 12	+71.34 	50663  	0:00.04	Nxe4 Bxe4+ dxe4 Bxe3 Bxg4+ Kb7 Qxe3 Qd4 Qxd4 Rxd4 b3 Rd2 Rac1 Kc7 Rcd1 Rxc2 Rd7+ Kb6 
 11	+73.09 	31300  	0:00.02	Nxe4 Bxe4+ dxe4 Bxe3 Bxg4+ Kb7 Qxe3 Qg6 Bf5 Qh6 Qxh6 gxh6 Rfd1 Nc4 Rxd8 
 10	+67.96 	16822  	0:00.01	Nxe4 Bxe4+ dxe4 Bxe3 Bxg4+ Kb8 fxe3 Nc4 Rd1 Qg6 Rxd8+ Rxd8 Bf5 
  9	+65.29 	9528    	0:00.01	Nxe4 Bxe4+ dxe4 Bxe3 Bxg4+ Kb8 fxe3 Nc4 Rd1 Qg6 Rxd8+ Rxd8 Bf5 
  8	+71.30 	4335    	0:00.00	Nxe4 Bxe4+ dxe4 Bxe3 Bxg4+ Kb7 Qxe3 Qd2 Rad1 Qxc2 Rxd8 Rxd8 
  7	+83.79 	1090    	0:00.00	Bxc5 Qxc5 dxe4 
  6	+65.03 	918      	0:00.00	Bxc5 Qxc5 dxe4 Be6 a3 
  5	+58.48 	746      	0:00.00	Bxc5 Qxc5 dxe4 Be6 a3 
  4	+67.21 	299      	0:00.00	Nxe4 Bxe4+ dxe4 Nb7 Bxg4+ Kb8 
  3	+70.07 	196      	0:00.00	Nxe4 Bxe4+ dxe4 Bxe3 
  2	+68.78 	132      	0:00.00	Nxe4 Bxe4+ dxe4 
  1	+74.86 	70        	0:00.00	Nxe4 
  0	#  

carldaman
Posts: 1721
Joined: Sat Jun 02, 2012 12:13 am

Re: One more issue with king safety/attack for Stockfish

Post by carldaman » Wed Jul 17, 2019 2:34 am

Nezh finds Qg6 quickly and sticks with it.

Code: Select all

20...Qg6 
  Black is clearly winning: -+ (-6.01 ++)  Depth: 28/41   00:00:05  41592kN
With White on move 20, it soon recognizes Black is better and wants to play Rh1, with a disadvantage, instead of the clearly losing Nxe4.

OneTrickPony
Posts: 93
Joined: Mon Apr 29, 2013 11:29 pm

Re: One more issue with king safety/attack for Stockfish

Post by OneTrickPony » Wed Jul 17, 2019 6:43 am

Great position!
It's scary how long it takes for SF to realize black is winning after 20.Nxe4. This is the kind of position it's supposed to be strong in.
Leela (both the official net and Jhortos J13-390 net) sees 20...Qg6 as winning instantly (on 2080ti) and it wants to play 20.dxe4 or 20.Bg5 instantly but it doesn't show convincingly winning eval for black for 20.Nxe4 (maybe that's how search in Leela works that evals for moves that are not preferred isn't very accurate).

Post Reply