Fascinating Study

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

Moderator: Ras

OWL
Posts: 8
Joined: Sat Dec 26, 2020 10:20 am
Full name: Daan

Fascinating Study

Post by OWL »

[d]3q2k1/P7/3P3P/5p2/1P4p1/4p1P1/P3B3/6K1 b - - 0 30
Black to move and win, my computer does not find it! This from an analysis from a game.
Top
User avatar
Ajedrecista
Posts: 2201
Joined: Wed Jul 13, 2011 9:04 pm
Location: Madrid, Spain.

Re: Fascinating study.

Post by Ajedrecista »

Hello Daan:
OWL wrote: ↑Sun Feb 07, 2021 8:11 pm [d]3q2k1/P7/3P3P/5p2/1P4p1/4p1P1/P3B3/6K1 b - - 0 30
Black to move and win, my computer does not find it! This from an analysis from a game.
SF 12 SSE4.1+POPCNT with two cores and NNUE disabled finds ..., Kf7 here with winning eval, using Multi-PV = 7. Analysis with full 3, 4 and 5-man Syzygy TB under Arena 2.0.1 GUI with 256 MB of hash running in an Intel i5-650:

Code: Select all

FEN: 3q2k1/P7/3P3P/5p2/1P4p1/4p1P1/P3B3/6K1 b - - 0 1

Stockfish_20090216_x64_modern:
[...]
---------------------------------------------------------------------------
 37/33	00:43	 117.450.044	2.677.778	 0,00	Qd8c8 b4b5 Kg8h7 d6d7 Qc8g8 b5b6 Qg8d5 b6b7 Qd5d2 Kg1f1 Qd2c1+ Kf1g2 Qc1b2 Kg2g1
 37/32	00:43	 117.450.044	2.677.778	 0,00	Kg8h8 d6d7 Kh8h7 b4b5 Qd8a5 Kg1f1 Kh7xh6 b5b6 Qa5d5 Kf1g1 Kh6g6 a2a4 Kg6h5 b6b7 Qd5d2 Kg1f1 Qd2c1+ Kf1g2 Qc1b2 a7a8Q Qb2xe2+ Kg2h1 Qe2d1+ Kh1h2 Qd1d2+ Kh2g1 Qd2c1+ Kg1h2 Qc1b2+ Kh2g1
 37/20	00:43	 117.450.044	2.677.778	 0,00	Qd8f8 d6d7 Kg8h7 b4b5 Kh7xh6 b5b6 Kh6g6 b6b7 Qf8b4 d7d8Q Qb4e1+ Kg1h2 Qe1xe2+ Kh2h1 Qe2e1+ Kh1h2 Qe1f2+ Kh2h1 Qf2f1+
 37/25	00:43	 117.450.044	2.677.778	 0,00	Kg8f8 h6h7 Kf8g7 d6d7 Kg7xh7 b4b5 Qd8a8 b5b6 Qa8d5 Kg1h2 Qd5xa2 d7d8Q Qa2xe2+ Kh2h1 Qe2f1+ Kh1h2 Qf1h3+ Kh2g1 Qh3xg3+ Kg1h1 Qg3f3+ Kh1g1
 37/25	00:43	 117.450.044	2.677.778	 0,00	Kg8f7 h6h7 Kf7g7 d6d7 Kg7xh7 b4b5 Qd8a8 b5b6 Qa8d5 Kg1h2 Qd5xa2 d7d8Q Qa2xe2+ Kh2h1 Qe2f1+ Kh1h2
 37/22	00:43	 117.450.044	2.677.778	 0,00	Qd8a8 d6d7 Kg8h7 b4b5 Kh7xh6 b5b6 Qa8d5 a2a3 Kh6h5 b6b7 Qd5d2 b7b8Q Qd2e1+ Kg1h2 Qe1xe2+ Kh2g1 Qe2e1+ Kg1g2 Qe1f2+ Kg2h1 Qf2e1+
 37/57	00:43	 117.450.044	2.677.778	+0,73	Kg8h7 d6d7 Qd8h8 b4b5 Qh8a1+ Kg1h2 Qa1xa2 Kh2h1 Qa2xe2 a7a8Q Qe2d1+ Kh1h2 Qd1d2+ Qa8g2 Qd2xd7 b5b6 Qd7b5 b6b7 Kh7xh6 Qg2a2 Kh6g7 Qa2a1+ Kg7g6 Qa1e1 Qb5b2+ Kh2g1 e3e2 Kg1f2 Qb2xb7 Qe1xe2 Kg6f6 Qe2e3 Qb7e4 Qe3e2 Qe4d4+ Kf2g2 Qd4d5+ Kg2f2 Qd5e4
---------------------------------------------------------------------------
 38/33	02:15	 387.551.156	2.865.803	 0,00	Qd8c8 b4b5 Kg8h7 d6d7 Qc8g8 b5b6 Qg8d5 b6b7 Qd5d2 Kg1f1 Qd2c1+ Kf1g2 Qc1b2 Kg2g1
 38/32	02:15	 387.551.156	2.865.803	 0,00	Kg8h8 d6d7 Kh8h7 b4b5 Qd8a5 Kg1f1 Kh7xh6 b5b6 Qa5d5 Kf1g1 Kh6g6 a2a4 Kg6h5 b6b7 Qd5d2 Kg1f1 Qd2c1+ Kf1g2 Qc1b2 a7a8Q Qb2xe2+ Kg2h1 Qe2d1+ Kh1h2 Qd1d2+ Kh2g1 Qd2c1+ Kg1h2 Qc1b2+ Kh2g1
 38/20	02:15	 387.551.156	2.865.803	 0,00	Qd8f8 d6d7 Kg8h7 b4b5 Kh7xh6 b5b6 Kh6g6 b6b7 Qf8b4 d7d8Q Qb4e1+ Kg1h2 Qe1xe2+ Kh2h1 Qe2e1+ Kh1h2 Qe1f2+ Kh2h1 Qf2e1+
 38/33	02:15	 387.551.156	2.865.803	 0,00	Kg8f8 h6h7 Kf8g7 d6d7 Kg7xh7 b4b5 Qd8a8 b5b6 Qa8d5 Kg1h2 Qd5xa2 d7d8Q Qa2xe2+ Kh2h1 Qe2f1+ Kh1h2 Qf1h3+ Kh2g1 Qh3xg3+ Kg1h1 Qg3f3+ Kh1g1 Qf3f2+ Kg1h1 e3e2 Qd8e7+ Kh7g6 Qe7e8+ Kg6f6 Qe8f8+ Kf6g5 Qf8e7+
 38/24	02:15	 387.551.156	2.865.803	 0,00	Qd8a8 d6d7 Kg8h7 b4b5 Kh7xh6 b5b6 Qa8d5 a2a3 Kh6h5 b6b7 Qd5d2 b7b8Q Qd2e1+ Kg1h2 Qe1xe2+ Kh2g1 Qe2e1+ Kg1g2 Qe1f2+ Kg2h1 Qf2f1+ Kh1h2 Qf1e2+
 38/57	02:15	 387.551.156	2.865.803	 0,00	Kg8h7 d6d7 Qd8h8 b4b5 Qh8a1+ Kg1h2 Qa1xa2 Kh2g1 Qa2xe2 a7a8Q Qe2f2+ Kg1h1 Qf2e1+ Kh1h2 Qe1d2+ Qa8g2 Qd2xd7 Qg2f1 Qd7e6 b5b6 e3e2 Qf1e1 Qe6xh6+ Kh2g2 Qh6c6+ Kg2f2 Qc6xb6+ Kf2xe2 Qb6c6 Ke2f1 Kh7g6 Qe1e3 Kg6f7 Qe3e5 Qc6f3+ Kf1g1 Qf3e4 Qe5c3 Kf7e6 Qc3b3+ Qe4d5 Qb3e3+ Ke6d6 Qe3b6+ Kd6e5 Qb6e3+ Qd5e4 Kg1f2 Ke5e6 Qe3c3
 38/56	02:15	 387.551.156	2.865.803	+4,48	Kg8f7 d6d7 Kf7e7 Be2b5 f5f4 g3xf4 e3e2 Bb5xe2 Qd8a8 h6h7 Qa8xa7+ Kg1g2 Qa7a8+ Kg2g3 Qa8h1 Be2xg4 Qh1xh7 b4b5 Qh7d3+ Kg3h4 Qd3xb5 Bg4e6 Qb5b4 Kh4g5 Qb4d2 Be6b3 Qd2xd7 f4f5 Qd7d2+ Kg5g6 Qd2f4 Bb3e6 Qf4e3 Kg6h5 Ke7f6
There were 1,353,019 TB hits at depth 38. I have not checked the full mainline.

Regards from Spain.

Ajedrecista.
Top
Vinvin
Posts: 5316
Joined: Thu Mar 09, 2006 9:40 am
Full name: Vincent Lejeune

Re: Fascinating Study

Post by Vinvin »

OWL wrote: ↑Sun Feb 07, 2021 8:11 pm 3q2k1/P7/3P3P/5p2/1P4p1/4p1P1/P3B3/6K1 b - - 0 30
Black to move and win, my computer does not find it! This from an analysis from a game.
3 runs with Crystal3.1 on an old i7-6*4 GHz. Time to find a rising eval : 5m52s, 2m18s and 4m18s. Average : 4 min 9 sec.

Code: Select all

Crystal3.1-sse4-popcnt:
Found 511 tablebases
NNUE evaluation using nn-62ef826d1a6d.nnue enabled
...
 34/44	01:26	 403 301 438	4 678 888	 0,00	30. ... Kh7 31.d7 Kxh6 32.b5 Qa5 33.Kh2 Kg7 34.b6 Qxa2 35.d8Q Qxe2+ 36.Kh1 Qf1+ 37.Kh2
 35/44	04:34	1 385 303 821	5 041 685	 0,00	30. ... Kh7 31.d7 Kxh6 32.b5 Qa5 33.Kh2 Kg7 34.b6 Qxa2 35.d8Q Qxe2+ 36.Kh1 Qf1+ 37.Kh2
 36/44-	05:32	1 686 457 764	5 079 676	-0,08	30. ... Kf7
 36/44-	05:59	1 826 077 089	5 084 115	-0,16	30. ... Kf7
 36/44-	29:03	9 163 809 973	5 255 834	-6,37	30. ... Kf7

Code: Select all

Crystal3.1-sse4-popcnt:
Found 511 tablebases
NNUE evaluation using nn-62ef826d1a6d.nnue enabled
...
 31/20	01:18	 424 259 467	5 422 400	 0,00	30. ... Kf7 31.h7 Ke6 32.Bc4+ Kxd6 33.Bg8 Qb6 34.h8Q e2+ 35.Kg2 Qb7+ 36.Kh2 e1Q 37.Qd4+ Kc7 38.Qg7+ Kb6 39.Qd4+
 32/28-	02:18	 740 302 289	5 343 483	-0,03	30. ... Kf7
 32/41-	02:24	 772 812 365	5 356 745	-0,08	30. ... Kf7
 32/62-	07:49	2 547 528 052	5 422 545	-4,87	30. ... Kf7

Code: Select all

Crystal3.1-sse4-popcnt:
Found 511 tablebases
NNUE evaluation using nn-62ef826d1a6d.nnue enabled
...
 32/21	01:50	 600 782 132	5 451 199	 0,00	30. ... Kf7 31.h7 Ke6 32.Bc4+ Kxd6 33.Bg8 Qb6 34.h8Q e2+ 35.Kg2 Qb7+ 36.Kh2 e1Q 37.Qd4+ Kc7 38.Bd5 Qe2+ 39.Kg1 Qe1+
 33/21-	04:58	1 585 654 392	5 313 997	-0,03	30. ... Kf7
 33/21-	05:20	1 710 687 284	5 329 310	-0,08	30. ... Kf7
 33/25-	10:41	3 441 686 926	5 361 050	-2,03	30. ... Kf7
Top
Vinvin
Posts: 5316
Joined: Thu Mar 09, 2006 9:40 am
Full name: Vincent Lejeune

Re: Fascinating Study

Post by Vinvin »

OWL wrote: ↑Sun Feb 07, 2021 8:11 pm 3q2k1/P7/3P3P/5p2/1P4p1/4p1P1/P3B3/6K1 b - - 0 30
Black to move and win, my computer does not find it! This from an analysis from a game.
3 runs with C-Fish (CF_EXT_x64_040121_ELTO_AVX_sparse) on an old i7-6*4 GHz. Time to find a rising eval : 1m19s, 16s and 27s. Average : 41 sec.

Code: Select all

CF_EXT_x64_040121_ELTO_AVX_sparse:
Hybrid NNUE evaluation using nn-62ef826d1a6d.nnue enabled.
...
 47/28	00:48	 334 452 058	6 955 145	 0,00	30. ... Kh7 31.d7 Qh8 32.b5 Qa1+ 33.Kh2 Qxa2 34.Kg1 Qxe2 35.a8Q Qe1+ 36.Kh2 Qd2+ 37.Qg2 Qd1 38.Qb2 Qxd7 39.Qe5 Qd2+ 40.Kg1 Qf2+ 41.Kh1 Qf1+ 42.Kh2 Qe2+ 43.Kg1
 48/27	00:58	 407 246 781	6 905 296	 0,00	30. ... Kh7 31.d7 Qh8 32.b5 Qa1+ 33.Kh2 Qxa2 34.Kg1 Qxe2 35.a8Q Qe1+ 36.Kh2 Qd2+ 37.Qg2 Qd1 38.Qb2 Qxd7 39.Qe5 Qd2+ 40.Kg1 Qf2+ 41.Kh1 Qf1+ 42.Kh2 Qe2+ 43.Kg1
 49/28-	01:19	 561 047 018	7 030 664	-0,08	30. ... Kf7
 49/28-	01:29	 632 859 319	7 046 960	-0,16	30. ... Kf7
 49/70-	04:16	1 940 298 611	7 563 425	-6,37	30. ... Kf7
 49/70	06:24	3 116 950 775	8 112 622	-7,78	30. ... Kf7 31.h7 Kg6 32.d7 f4 33.Bd3+ Kh5 34.gxf4 Kh4 35.Kf1 Kg3 36.Ke2 Kxf4 37.Bb5 Qf8 38.Kd3 g3 39.Kc2 e2 40.a8R e1N+ 41.Kb3 Qxa8 42.Bc6 Qf8 43.b5 g2 44.Bxg2 Nxg2 45.a4 Kf5 46.b6 Ne3 47.b7 Ke6 48.d8R Qxd8 49.Kc3 Nd5+ 50.Kc2 Qh8 51.b8Q Qxb8
 50/45+	06:42	3 264 846 804	8 115 291	-7,63	30. ... Kf7 31.h7
 50/45+	07:46	3 809 761 954	8 159 014	-7,59	30. ... Kf7 31.h7
 50/45+	10:19	5 129 032 865	8 273 754	-6,25	30. ... Kf7 31.h7

Code: Select all

CF_EXT_x64_040121_ELTO_AVX_sparse:
Hybrid NNUE evaluation using nn-62ef826d1a6d.nnue enabled.
...
 38/19	00:13	 87 836 935	6 664 916	 0,00	30. ... Kf7 31.h7 Qh8 32.b5 Qa1+ 33.Kg2 Qxa2 34.h8Q Qxe2+ 35.Kh1 Qe1+ 36.Kg2 Qd2+ 37.Kg1 Qd1+ 38.Kh2 Qc2+ 39.Kg1
 39/20-	00:16	 110 986 227	6 742 784	-0,08	30. ... Kf7
 39/20-	00:17	 115 464 846	6 759 841	-0,16	30. ... Kf7
 39/44-	00:34	 246 513 809	7 106 806	-4,87	30. ... Kf7
 39/50	00:57	 443 732 083	7 759 181	-5,76	30. ... Kf7 31.h7 Kg6 32.Bd3 Qh8 33.Bf1 Kxh7 34.d7 Kg6 35.a4 f4 36.gxf4 g3 37.Bd3+ Kg7 38.Kg2 Qh2+ 39.Kf3 Qf2+ 40.Ke4 Qg2+ 41.Kxe3 Qa8 42.Bb5 Qxa7+ 43.Kf3 Qf2+ 44.Ke4 Qg2+ 45.Ke3 Qd5 46.Ke2 Qe4+ 47.Kd2 Qd4+ 48.Bd3 Qxd7 49.Kc2 Qxa4+ 50.Kc3
 40/26+	01:04	 502 407 013	7 819 077	-5,68	30. ... Kf7 31.h7
 40/32-	02:11	 898 376 729	6 821 024	-5,76	30. ... Kf7

Code: Select all

CF_EXT_x64_040121_ELTO_AVX_sparse:
Hybrid NNUE evaluation using nn-62ef826d1a6d.nnue enabled.
...
 41/53	00:17	 125 887 930	7 161 675	 0,00	30. ... Kh7 31.d7 Qh8 32.b5 Qa1+ 33.Kh2 Qxa2 34.Kg1 Qxe2 35.a8Q Qf2+ 36.Kh1 Qe1+ 37.Kh2 Qd2+ 38.Qg2 Qxd7 39.b6 Qd2 40.Qxd2 exd2 41.b7 d1Q 42.b8Q Qd2+ 43.Kg1 Qe3+ 44.Kf1 Qd3+ 45.Kg1 Kxh6 46.Qe5 Qd1+ 47.Kf2 Qd2+ 48.Qe2 Qd5 49.Ke1 Qh1+ 50.Kd2 Qc6 51.Qd3 Kg6 52.Qe3 Kf7 53.Qe5 Qf3 54.Qd6 Qe4 55.Qd7+ Kf6 56.Qd8+ Ke6
 42/34-	00:27	 192 952 911	7 133 721	-0,08	30. ... Kf7
 42/34-	00:28	 200 635 188	7 128 373	-0,16	30. ... Kf7
 42/61-	01:00	 457 967 332	7 527 281	-4,87	30. ... Kf7
 42/61	01:34	 756 103 700	7 965 442	-6,33	30. ... Kf7 31.d7 Ke7 32.a4 Kxd7 33.a5 Qh8 34.b5 Qa1+ 35.Kh2 Qxa5 36.h7 Qd8 37.Bc4 Qh8 38.Kg1 Qa1+ 39.Kh2 Kd6 40.b6 Qb2+ 41.Kg1 Qc1+ 42.Kh2 Qd2+ 43.Kg1 Qe1+ 44.Kh2 Qf2+ 45.Kh1 e2 46.Bxe2 Qe1+ 47.Kh2 Qxe2+ 48.Kg1 Qe3+ 49.Kg2 Qf3+ 50.Kg1 Qxg3+ 51.Kf1 Qf4+ 52.Ke1 Qb4+ 53.Ke2
 43/38+	01:55	 932 681 762	8 079 433	-5,99	30. ... Kf7 31.h7
 43/38-	03:03	1 520 059 261	8 277 611	-6,33	30. ... Kf7
 43/50-	12:18	6 919 601 042	9 374 768	-8,93	30. ... Kf7
Top

Return to “Computer Chess Club: General Topics”