will Leela ever solve this ?

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

Moderators: hgm, Rebel, chrisw

Joost Buijs
Posts: 1563
Joined: Thu Jul 16, 2009 10:47 am
Location: Almere, The Netherlands

Re: will Leela ever solve this ?

Post by Joost Buijs »

CMCanavessi wrote: ↑Sun Jun 10, 2018 9:39 pm
Spliffjiffer wrote: ↑Sun Jun 10, 2018 3:30 pm [d]8/8/8/3qk3/3p4/R1PK4/2PN4/8 w - - 0 1

Ra5, game over :-)
Finds it in 6 seconds:

Code: Select all

FEN: 8/8/8/3qk3/3p4/R1PK4/2PN4/8 w - - 0 1

Lc0 ID 25 (TEST):
 1/2	00:00,010	 2	200	-1,57	1.cxd4+ Qxd4+
 1/3	00:00,011	 3	272	-0,72	1.cxd4+ Qxd4+ 2.Ke2
 1/4	00:00,012	 6	500	 0,00	1.cxd4+ Qxd4+ 2.Ke2 Qg4+
 1/5	00:00,013	 15	1k	-0,07	1.cxd4+ Qxd4+ 2.Ke2 Qc5 3.Re3+
 1/6	00:00,015	 33	2k	-0,10	1.cxd4+ Qxd4+ 2.Ke2 Qc5 3.Re3+ Kf5
 1/7	00:00,017	 65	4k	-0,01	1.cxd4+ Qxd4+ 2.Ke2 Qc5 3.Re3+ Kd5 4.c4+
 1/8	00:00,018	 97	5k	+0,13	1.cxd4+ Qxd4+ 2.Ke2 Qg4+ 3.Nf3+ Kd6 4.Rd3+ Kc5
 1/9	00:00,020	 148	7k	+0,03	1.cxd4+ Qxd4+ 2.Ke2 Qg4+ 3.Nf3+ Kd6 4.Rd3+ Kc5 5.Ke3
 1/9	00:00,022	 187	9k	+0,68	1.c4 Qc5 2.Ra8 Qb4 3.Nf3+ Kf4 4.Nxd4
 1/10	00:00,026	 282	11k	+0,73	1.c4 Qc5 2.Ra8 Qb4 3.Nf3+ Kf4 4.Nxd4 Qb1
 1/11	00:00,035	 504	14k	+0,38	1.c4 Qg2 2.Ra5+ Kd6 3.Ne4+ Kc7 4.Rf5 Qh3+
 1/12	00:00,076	 2k	21k	+0,16	1.c4 Qg2 2.Ra5+ Kf4 3.c5 Qd5 4.Nb3 Ke5 5.c6 Kd6
 1/13	00:00,099	 2k	23k	+0,01	1.c4 Qg2 2.Ra5+ Kf4 3.c5 Qg6+ 4.Kxd4 Qxc2 5.Nc4 Qe4+ 6.Kc3 Qe1+
 1/13	00:00,111	 3k	24k	+0,20	1.Nc4+ Kf4 2.cxd4 Qf3+ 3.Kd2 Qd5 4.Kc3 Ke4 5.Nd2+ Ke3 6.Nc4+ Ke4 7.Nd2+
 1/14	00:00,123	 3k	24k	+0,25	1.Nc4+ Kf4 2.cxd4 Qf3+ 3.Kd2 Qd5 4.Kc3 Ke4 5.Nd2+ Ke3 6.Nc4+ Ke4 7.Nd2+
 1/15	00:00,143	 4k	25k	+0,23	1.Nc4+ Kf4 2.cxd4 Qf3+ 3.Kd2 Qd5 4.Kc3 Ke4 5.Nd2+ Ke3 6.Nc4+ Ke4 7.Nd2+ Ke3
 1/16	00:00,178	 5k	27k	+0,22	1.Nc4+ Kf4 2.cxd4 Qf3+ 3.Kd2 Qd5 4.Kc3 Ke4 5.Nd2+ Ke3 6.Nc4+ Ke4 7.Nd2+ Ke3 8.Nc4+
 1/17	00:00,223	 6k	29k	+0,25	1.Nc4+ Kf4 2.cxd4 Qf3+ 3.Kd2 Qd5 4.Kc3 Ke4 5.Nd2+ Ke3 6.Nc4+ Ke4 7.Nd2+ Ke3 8.Nc4+ Ke4
 1/19	00:00,396	 13k	32k	+0,25	1.Nc4+ Kf4 2.cxd4 Qf3+ 3.Kd2 Qd5 4.Kc3 Ke4 5.Nd2+ Ke3 6.Nc4+ Ke4 7.Nd2+ Ke3 8.Nc4+ Ke4
 1/20	00:00,776	 28k	36k	+0,30	1.Nc4+ Kf4 2.cxd4 Qf3+ 3.Kd2 Qf2+ 4.Kd3 Qf3+ 5.Kd2 Qd5 6.Kc3 Ke4 7.Nd2+ Ke3 8.Nc4+ Ke4 9.Nd2+ Ke3 10.Nc4+ Kf4
 2/20	00:00,859	 31k	36k	+0,31	1.Nc4+ Kf4 2.cxd4 Qf3+ 3.Kd2 Qf2+ 4.Kd3 Qf3+ 5.Kd2 Qd5 6.Kc3 Ke4 7.Nd2+ Ke3 8.Nc4+ Ke4 9.Nd2+ Ke3 10.Nc4+ Kf4
 2/21	00:01,094	 41k	37k	+0,35	1.Nc4+ Kf4 2.cxd4 Qf3+ 3.Kd2 Qf2+ 4.Kd3 Qf3+ 5.Kd2 Qd5 6.Kc3 Ke4 7.Nd2+ Ke3 8.Nc4+ Ke4 9.Nd2+ Ke3 10.Nc4+ Kf4
 2/22	00:02,602	 103k	39k	+0,43	1.Nc4+ Kf4 2.cxd4 Qf3+ 3.Kd2 Qf2+ 4.Kc3 Ke4 5.Nd2+ Kf4 6.Nb3 Qe1+ 7.Kc4 Qe2+ 8.Kc3 Qe1+ 9.Kb2 Qb4 10.Ra8 Ke4
 2/23	00:04,326	 185k	43k	+0,49	1.Nc4+ Kf4 2.cxd4 Qf3+ 3.Kd2 Qf2+ 4.Kc3 Qe1+ 5.Kb3 Qd1 6.Kc3 Qe1+ 7.Kb3 Qb1+ 8.Kc3 Qe1+
 2/29	00:05,777	 274k	47k	+0,49	1.Nc4+ Kf4 2.cxd4 Qf3+ 3.Kd2 Qf2+ 4.Kc3 Qe1+ 5.Kb3 Qd1 6.Kc3 Qe1+ 7.Kb3 Qb1+ 8.Kc3 Qe1+
 2/29	00:06,320	 310k	49k	+20,27	1.Ra5 Qxa5 2.Nc4+ Kd5 3.Nxa5 dxc3 4.Kxc3 Kc5 5.Nb3+ Kb5 6.Kd4 Kb4 7.c4 Kxb3 8.c5 Kb4 9.c6 Kb5 10.c7 Kc6 11.c8Q+ Kd6 12.Qb8+ Kc6 13.Kc4 Kd7
 2/31	00:07,178	 367k	51k	+20,74	1.Ra5 Qxa5 2.Nc4+ Kd5 3.Nxa5 dxc3 4.Kxc3 Kc5 5.Nb3+ Kb5 6.Kd4 Kb4 7.c4 Kxb3 8.c5 Kb4 9.c6 Kb5 10.c7 Kc6 11.c8Q+ Kd6 12.Qb8+ Kc6 13.Kc4 Kd7
 2/32	00:07,680	 400k	52k	+20,99	1.Ra5 Qxa5 2.Nc4+ Kd5 3.Nxa5 dxc3 4.Kxc3 Kc5 5.Nb3+ Kb5 6.Kd4 Kb4 7.c4 Kxb3 8.c5 Kb4 9.c6 Kb5 10.c7 Kc6 11.c8Q+ Kd6 12.Qb8+ Kc6 13.Kc4 Kd7 14.Kd5
 2/32	00:12,684	 735k	58k	+22,38	1.Ra5 Qxa5 2.Nc4+ Kd5 3.Nxa5 dxc3 4.Kxc3 Kc5 5.Nb3+ Kb5 6.Kd4 Kb4 7.Nd2 Kb5 8.c3 Kc6 9.c4 Kd6 10.c5+ Kc6 11.Nc4 Kc7 12.Kd5 Kd7 13.c6+ Kc7
 2/32	00:17,688	 1.073k	61k	+23,32	1.Ra5 Qxa5 2.Nc4+ Kd5 3.Nxa5 dxc3 4.Kxc3 Kc5 5.Nb3+ Kb5 6.Kd4 Kb4 7.Na1 Kb5 8.c3 Kc6 9.c4 Kd6 10.Nb3 Kc6 11.c5 Kc7 12.Kd5 Kd7 13.Nd4 Kc7 14.c6 Kc8
 2/33	00:19,591	 1.206k	62k	+23,72	1.Ra5 Qxa5 2.Nc4+ Kd5 3.Nxa5 dxc3 4.Kxc3 Kc5 5.Nb3+ Kb5 6.Kd4 Kb4 7.Na1 Kb5 8.c3 Kc6 9.c4 Kd6 10.Nb3 Kc6 11.c5 Kc7 12.Kd5 Kd7 13.Nd4 Kc7 14.c6 Kc8 15.Nb5
 2/34	00:19,828	 1.223k	62k	+23,77	1.Ra5 Qxa5 2.Nc4+ Kd5 3.Nxa5 dxc3 4.Kxc3 Kc5 5.Nb3+ Kb5 6.Kd4 Kb4 7.Na1 Kb5 8.c3 Kc6 9.c4 Kd6 10.Nb3 Kc6 11.c5 Kc7 12.Kd5 Kd7 13.Nd4 Kc7 14.c6 Kc8 15.Nb5
 2/34	00:22,167	 1.382k	62k	+24,16	1.Ra5 Qxa5 2.Nc4+ Kd5 3.Nxa5 dxc3 4.Kxc3 Kc5 5.Nb3+ Kb5 6.Kd4 Kb4 7.Na1 Kb5 8.Kd5 Kb4 9.c4 Kc3 10.c5 Kb4 11.c6 Kb5 12.c7 Kb6 13.c8Q Ka7 14.Nc2 Kb6 15.Qb8+ Ka6 16.Kc5 Ka5 17.Qa7+
Wow! That is fast! Most a-b engines will find it within a millisecond, it is obvious that tactics is not one of Leela's strong points.
Top
User avatar
CMCanavessi
Posts: 1142
Joined: Thu Dec 28, 2017 4:06 pm
Location: Argentina

Re: will Leela ever solve this ?

Post by CMCanavessi »

Joost Buijs wrote: ↑Mon Jun 11, 2018 1:32 pm
CMCanavessi wrote: ↑Sun Jun 10, 2018 9:39 pm
Spliffjiffer wrote: ↑Sun Jun 10, 2018 3:30 pm [d]8/8/8/3qk3/3p4/R1PK4/2PN4/8 w - - 0 1

Ra5, game over :-)
Finds it in 6 seconds:

Code: Select all

FEN: 8/8/8/3qk3/3p4/R1PK4/2PN4/8 w - - 0 1

Lc0 ID 25 (TEST):
 1/2	00:00,010	 2	200	-1,57	1.cxd4+ Qxd4+
 1/3	00:00,011	 3	272	-0,72	1.cxd4+ Qxd4+ 2.Ke2
 1/4	00:00,012	 6	500	 0,00	1.cxd4+ Qxd4+ 2.Ke2 Qg4+
 1/5	00:00,013	 15	1k	-0,07	1.cxd4+ Qxd4+ 2.Ke2 Qc5 3.Re3+
 1/6	00:00,015	 33	2k	-0,10	1.cxd4+ Qxd4+ 2.Ke2 Qc5 3.Re3+ Kf5
 1/7	00:00,017	 65	4k	-0,01	1.cxd4+ Qxd4+ 2.Ke2 Qc5 3.Re3+ Kd5 4.c4+
 1/8	00:00,018	 97	5k	+0,13	1.cxd4+ Qxd4+ 2.Ke2 Qg4+ 3.Nf3+ Kd6 4.Rd3+ Kc5
 1/9	00:00,020	 148	7k	+0,03	1.cxd4+ Qxd4+ 2.Ke2 Qg4+ 3.Nf3+ Kd6 4.Rd3+ Kc5 5.Ke3
 1/9	00:00,022	 187	9k	+0,68	1.c4 Qc5 2.Ra8 Qb4 3.Nf3+ Kf4 4.Nxd4
 1/10	00:00,026	 282	11k	+0,73	1.c4 Qc5 2.Ra8 Qb4 3.Nf3+ Kf4 4.Nxd4 Qb1
 1/11	00:00,035	 504	14k	+0,38	1.c4 Qg2 2.Ra5+ Kd6 3.Ne4+ Kc7 4.Rf5 Qh3+
 1/12	00:00,076	 2k	21k	+0,16	1.c4 Qg2 2.Ra5+ Kf4 3.c5 Qd5 4.Nb3 Ke5 5.c6 Kd6
 1/13	00:00,099	 2k	23k	+0,01	1.c4 Qg2 2.Ra5+ Kf4 3.c5 Qg6+ 4.Kxd4 Qxc2 5.Nc4 Qe4+ 6.Kc3 Qe1+
 1/13	00:00,111	 3k	24k	+0,20	1.Nc4+ Kf4 2.cxd4 Qf3+ 3.Kd2 Qd5 4.Kc3 Ke4 5.Nd2+ Ke3 6.Nc4+ Ke4 7.Nd2+
 1/14	00:00,123	 3k	24k	+0,25	1.Nc4+ Kf4 2.cxd4 Qf3+ 3.Kd2 Qd5 4.Kc3 Ke4 5.Nd2+ Ke3 6.Nc4+ Ke4 7.Nd2+
 1/15	00:00,143	 4k	25k	+0,23	1.Nc4+ Kf4 2.cxd4 Qf3+ 3.Kd2 Qd5 4.Kc3 Ke4 5.Nd2+ Ke3 6.Nc4+ Ke4 7.Nd2+ Ke3
 1/16	00:00,178	 5k	27k	+0,22	1.Nc4+ Kf4 2.cxd4 Qf3+ 3.Kd2 Qd5 4.Kc3 Ke4 5.Nd2+ Ke3 6.Nc4+ Ke4 7.Nd2+ Ke3 8.Nc4+
 1/17	00:00,223	 6k	29k	+0,25	1.Nc4+ Kf4 2.cxd4 Qf3+ 3.Kd2 Qd5 4.Kc3 Ke4 5.Nd2+ Ke3 6.Nc4+ Ke4 7.Nd2+ Ke3 8.Nc4+ Ke4
 1/19	00:00,396	 13k	32k	+0,25	1.Nc4+ Kf4 2.cxd4 Qf3+ 3.Kd2 Qd5 4.Kc3 Ke4 5.Nd2+ Ke3 6.Nc4+ Ke4 7.Nd2+ Ke3 8.Nc4+ Ke4
 1/20	00:00,776	 28k	36k	+0,30	1.Nc4+ Kf4 2.cxd4 Qf3+ 3.Kd2 Qf2+ 4.Kd3 Qf3+ 5.Kd2 Qd5 6.Kc3 Ke4 7.Nd2+ Ke3 8.Nc4+ Ke4 9.Nd2+ Ke3 10.Nc4+ Kf4
 2/20	00:00,859	 31k	36k	+0,31	1.Nc4+ Kf4 2.cxd4 Qf3+ 3.Kd2 Qf2+ 4.Kd3 Qf3+ 5.Kd2 Qd5 6.Kc3 Ke4 7.Nd2+ Ke3 8.Nc4+ Ke4 9.Nd2+ Ke3 10.Nc4+ Kf4
 2/21	00:01,094	 41k	37k	+0,35	1.Nc4+ Kf4 2.cxd4 Qf3+ 3.Kd2 Qf2+ 4.Kd3 Qf3+ 5.Kd2 Qd5 6.Kc3 Ke4 7.Nd2+ Ke3 8.Nc4+ Ke4 9.Nd2+ Ke3 10.Nc4+ Kf4
 2/22	00:02,602	 103k	39k	+0,43	1.Nc4+ Kf4 2.cxd4 Qf3+ 3.Kd2 Qf2+ 4.Kc3 Ke4 5.Nd2+ Kf4 6.Nb3 Qe1+ 7.Kc4 Qe2+ 8.Kc3 Qe1+ 9.Kb2 Qb4 10.Ra8 Ke4
 2/23	00:04,326	 185k	43k	+0,49	1.Nc4+ Kf4 2.cxd4 Qf3+ 3.Kd2 Qf2+ 4.Kc3 Qe1+ 5.Kb3 Qd1 6.Kc3 Qe1+ 7.Kb3 Qb1+ 8.Kc3 Qe1+
 2/29	00:05,777	 274k	47k	+0,49	1.Nc4+ Kf4 2.cxd4 Qf3+ 3.Kd2 Qf2+ 4.Kc3 Qe1+ 5.Kb3 Qd1 6.Kc3 Qe1+ 7.Kb3 Qb1+ 8.Kc3 Qe1+
 2/29	00:06,320	 310k	49k	+20,27	1.Ra5 Qxa5 2.Nc4+ Kd5 3.Nxa5 dxc3 4.Kxc3 Kc5 5.Nb3+ Kb5 6.Kd4 Kb4 7.c4 Kxb3 8.c5 Kb4 9.c6 Kb5 10.c7 Kc6 11.c8Q+ Kd6 12.Qb8+ Kc6 13.Kc4 Kd7
 2/31	00:07,178	 367k	51k	+20,74	1.Ra5 Qxa5 2.Nc4+ Kd5 3.Nxa5 dxc3 4.Kxc3 Kc5 5.Nb3+ Kb5 6.Kd4 Kb4 7.c4 Kxb3 8.c5 Kb4 9.c6 Kb5 10.c7 Kc6 11.c8Q+ Kd6 12.Qb8+ Kc6 13.Kc4 Kd7
 2/32	00:07,680	 400k	52k	+20,99	1.Ra5 Qxa5 2.Nc4+ Kd5 3.Nxa5 dxc3 4.Kxc3 Kc5 5.Nb3+ Kb5 6.Kd4 Kb4 7.c4 Kxb3 8.c5 Kb4 9.c6 Kb5 10.c7 Kc6 11.c8Q+ Kd6 12.Qb8+ Kc6 13.Kc4 Kd7 14.Kd5
 2/32	00:12,684	 735k	58k	+22,38	1.Ra5 Qxa5 2.Nc4+ Kd5 3.Nxa5 dxc3 4.Kxc3 Kc5 5.Nb3+ Kb5 6.Kd4 Kb4 7.Nd2 Kb5 8.c3 Kc6 9.c4 Kd6 10.c5+ Kc6 11.Nc4 Kc7 12.Kd5 Kd7 13.c6+ Kc7
 2/32	00:17,688	 1.073k	61k	+23,32	1.Ra5 Qxa5 2.Nc4+ Kd5 3.Nxa5 dxc3 4.Kxc3 Kc5 5.Nb3+ Kb5 6.Kd4 Kb4 7.Na1 Kb5 8.c3 Kc6 9.c4 Kd6 10.Nb3 Kc6 11.c5 Kc7 12.Kd5 Kd7 13.Nd4 Kc7 14.c6 Kc8
 2/33	00:19,591	 1.206k	62k	+23,72	1.Ra5 Qxa5 2.Nc4+ Kd5 3.Nxa5 dxc3 4.Kxc3 Kc5 5.Nb3+ Kb5 6.Kd4 Kb4 7.Na1 Kb5 8.c3 Kc6 9.c4 Kd6 10.Nb3 Kc6 11.c5 Kc7 12.Kd5 Kd7 13.Nd4 Kc7 14.c6 Kc8 15.Nb5
 2/34	00:19,828	 1.223k	62k	+23,77	1.Ra5 Qxa5 2.Nc4+ Kd5 3.Nxa5 dxc3 4.Kxc3 Kc5 5.Nb3+ Kb5 6.Kd4 Kb4 7.Na1 Kb5 8.c3 Kc6 9.c4 Kd6 10.Nb3 Kc6 11.c5 Kc7 12.Kd5 Kd7 13.Nd4 Kc7 14.c6 Kc8 15.Nb5
 2/34	00:22,167	 1.382k	62k	+24,16	1.Ra5 Qxa5 2.Nc4+ Kd5 3.Nxa5 dxc3 4.Kxc3 Kc5 5.Nb3+ Kb5 6.Kd4 Kb4 7.Na1 Kb5 8.Kd5 Kb4 9.c4 Kc3 10.c5 Kb4 11.c6 Kb5 12.c7 Kb6 13.c8Q Ka7 14.Nc2 Kb6 15.Qb8+ Ka6 16.Kc5 Ka5 17.Qa7+
Wow! That is fast! Most a-b engines will find it within a millisecond, it is obvious that tactics is not one of Leela's strong points.
"Most" a-b engines will still lose to Leela (with the exception of the very top, for now...), so who cares about the time it takes :mrgreen:
Follow my tournament and some Leela gauntlets live at http://twitch.tv/ccls
Top
glennsamuel32
Posts: 136
Joined: Sat Dec 04, 2010 5:31 pm
Location: 223

Re: will Leela ever solve this ?

Post by glennsamuel32 »

CMCanavessi wrote: ↑Mon Jun 11, 2018 2:29 pm
Joost Buijs wrote: ↑Mon Jun 11, 2018 1:32 pm
CMCanavessi wrote: ↑Sun Jun 10, 2018 9:39 pm

Finds it in 6 seconds:

Code: Select all

FEN: 8/8/8/3qk3/3p4/R1PK4/2PN4/8 w - - 0 1

Lc0 ID 25 (TEST):
 1/2	00:00,010	 2	200	-1,57	1.cxd4+ Qxd4+
 1/3	00:00,011	 3	272	-0,72	1.cxd4+ Qxd4+ 2.Ke2
 1/4	00:00,012	 6	500	 0,00	1.cxd4+ Qxd4+ 2.Ke2 Qg4+
 1/5	00:00,013	 15	1k	-0,07	1.cxd4+ Qxd4+ 2.Ke2 Qc5 3.Re3+
 1/6	00:00,015	 33	2k	-0,10	1.cxd4+ Qxd4+ 2.Ke2 Qc5 3.Re3+ Kf5
 1/7	00:00,017	 65	4k	-0,01	1.cxd4+ Qxd4+ 2.Ke2 Qc5 3.Re3+ Kd5 4.c4+
 1/8	00:00,018	 97	5k	+0,13	1.cxd4+ Qxd4+ 2.Ke2 Qg4+ 3.Nf3+ Kd6 4.Rd3+ Kc5
 1/9	00:00,020	 148	7k	+0,03	1.cxd4+ Qxd4+ 2.Ke2 Qg4+ 3.Nf3+ Kd6 4.Rd3+ Kc5 5.Ke3
 1/9	00:00,022	 187	9k	+0,68	1.c4 Qc5 2.Ra8 Qb4 3.Nf3+ Kf4 4.Nxd4
 1/10	00:00,026	 282	11k	+0,73	1.c4 Qc5 2.Ra8 Qb4 3.Nf3+ Kf4 4.Nxd4 Qb1
 1/11	00:00,035	 504	14k	+0,38	1.c4 Qg2 2.Ra5+ Kd6 3.Ne4+ Kc7 4.Rf5 Qh3+
 1/12	00:00,076	 2k	21k	+0,16	1.c4 Qg2 2.Ra5+ Kf4 3.c5 Qd5 4.Nb3 Ke5 5.c6 Kd6
 1/13	00:00,099	 2k	23k	+0,01	1.c4 Qg2 2.Ra5+ Kf4 3.c5 Qg6+ 4.Kxd4 Qxc2 5.Nc4 Qe4+ 6.Kc3 Qe1+
 1/13	00:00,111	 3k	24k	+0,20	1.Nc4+ Kf4 2.cxd4 Qf3+ 3.Kd2 Qd5 4.Kc3 Ke4 5.Nd2+ Ke3 6.Nc4+ Ke4 7.Nd2+
 1/14	00:00,123	 3k	24k	+0,25	1.Nc4+ Kf4 2.cxd4 Qf3+ 3.Kd2 Qd5 4.Kc3 Ke4 5.Nd2+ Ke3 6.Nc4+ Ke4 7.Nd2+
 1/15	00:00,143	 4k	25k	+0,23	1.Nc4+ Kf4 2.cxd4 Qf3+ 3.Kd2 Qd5 4.Kc3 Ke4 5.Nd2+ Ke3 6.Nc4+ Ke4 7.Nd2+ Ke3
 1/16	00:00,178	 5k	27k	+0,22	1.Nc4+ Kf4 2.cxd4 Qf3+ 3.Kd2 Qd5 4.Kc3 Ke4 5.Nd2+ Ke3 6.Nc4+ Ke4 7.Nd2+ Ke3 8.Nc4+
 1/17	00:00,223	 6k	29k	+0,25	1.Nc4+ Kf4 2.cxd4 Qf3+ 3.Kd2 Qd5 4.Kc3 Ke4 5.Nd2+ Ke3 6.Nc4+ Ke4 7.Nd2+ Ke3 8.Nc4+ Ke4
 1/19	00:00,396	 13k	32k	+0,25	1.Nc4+ Kf4 2.cxd4 Qf3+ 3.Kd2 Qd5 4.Kc3 Ke4 5.Nd2+ Ke3 6.Nc4+ Ke4 7.Nd2+ Ke3 8.Nc4+ Ke4
 1/20	00:00,776	 28k	36k	+0,30	1.Nc4+ Kf4 2.cxd4 Qf3+ 3.Kd2 Qf2+ 4.Kd3 Qf3+ 5.Kd2 Qd5 6.Kc3 Ke4 7.Nd2+ Ke3 8.Nc4+ Ke4 9.Nd2+ Ke3 10.Nc4+ Kf4
 2/20	00:00,859	 31k	36k	+0,31	1.Nc4+ Kf4 2.cxd4 Qf3+ 3.Kd2 Qf2+ 4.Kd3 Qf3+ 5.Kd2 Qd5 6.Kc3 Ke4 7.Nd2+ Ke3 8.Nc4+ Ke4 9.Nd2+ Ke3 10.Nc4+ Kf4
 2/21	00:01,094	 41k	37k	+0,35	1.Nc4+ Kf4 2.cxd4 Qf3+ 3.Kd2 Qf2+ 4.Kd3 Qf3+ 5.Kd2 Qd5 6.Kc3 Ke4 7.Nd2+ Ke3 8.Nc4+ Ke4 9.Nd2+ Ke3 10.Nc4+ Kf4
 2/22	00:02,602	 103k	39k	+0,43	1.Nc4+ Kf4 2.cxd4 Qf3+ 3.Kd2 Qf2+ 4.Kc3 Ke4 5.Nd2+ Kf4 6.Nb3 Qe1+ 7.Kc4 Qe2+ 8.Kc3 Qe1+ 9.Kb2 Qb4 10.Ra8 Ke4
 2/23	00:04,326	 185k	43k	+0,49	1.Nc4+ Kf4 2.cxd4 Qf3+ 3.Kd2 Qf2+ 4.Kc3 Qe1+ 5.Kb3 Qd1 6.Kc3 Qe1+ 7.Kb3 Qb1+ 8.Kc3 Qe1+
 2/29	00:05,777	 274k	47k	+0,49	1.Nc4+ Kf4 2.cxd4 Qf3+ 3.Kd2 Qf2+ 4.Kc3 Qe1+ 5.Kb3 Qd1 6.Kc3 Qe1+ 7.Kb3 Qb1+ 8.Kc3 Qe1+
 2/29	00:06,320	 310k	49k	+20,27	1.Ra5 Qxa5 2.Nc4+ Kd5 3.Nxa5 dxc3 4.Kxc3 Kc5 5.Nb3+ Kb5 6.Kd4 Kb4 7.c4 Kxb3 8.c5 Kb4 9.c6 Kb5 10.c7 Kc6 11.c8Q+ Kd6 12.Qb8+ Kc6 13.Kc4 Kd7
 2/31	00:07,178	 367k	51k	+20,74	1.Ra5 Qxa5 2.Nc4+ Kd5 3.Nxa5 dxc3 4.Kxc3 Kc5 5.Nb3+ Kb5 6.Kd4 Kb4 7.c4 Kxb3 8.c5 Kb4 9.c6 Kb5 10.c7 Kc6 11.c8Q+ Kd6 12.Qb8+ Kc6 13.Kc4 Kd7
 2/32	00:07,680	 400k	52k	+20,99	1.Ra5 Qxa5 2.Nc4+ Kd5 3.Nxa5 dxc3 4.Kxc3 Kc5 5.Nb3+ Kb5 6.Kd4 Kb4 7.c4 Kxb3 8.c5 Kb4 9.c6 Kb5 10.c7 Kc6 11.c8Q+ Kd6 12.Qb8+ Kc6 13.Kc4 Kd7 14.Kd5
 2/32	00:12,684	 735k	58k	+22,38	1.Ra5 Qxa5 2.Nc4+ Kd5 3.Nxa5 dxc3 4.Kxc3 Kc5 5.Nb3+ Kb5 6.Kd4 Kb4 7.Nd2 Kb5 8.c3 Kc6 9.c4 Kd6 10.c5+ Kc6 11.Nc4 Kc7 12.Kd5 Kd7 13.c6+ Kc7
 2/32	00:17,688	 1.073k	61k	+23,32	1.Ra5 Qxa5 2.Nc4+ Kd5 3.Nxa5 dxc3 4.Kxc3 Kc5 5.Nb3+ Kb5 6.Kd4 Kb4 7.Na1 Kb5 8.c3 Kc6 9.c4 Kd6 10.Nb3 Kc6 11.c5 Kc7 12.Kd5 Kd7 13.Nd4 Kc7 14.c6 Kc8
 2/33	00:19,591	 1.206k	62k	+23,72	1.Ra5 Qxa5 2.Nc4+ Kd5 3.Nxa5 dxc3 4.Kxc3 Kc5 5.Nb3+ Kb5 6.Kd4 Kb4 7.Na1 Kb5 8.c3 Kc6 9.c4 Kd6 10.Nb3 Kc6 11.c5 Kc7 12.Kd5 Kd7 13.Nd4 Kc7 14.c6 Kc8 15.Nb5
 2/34	00:19,828	 1.223k	62k	+23,77	1.Ra5 Qxa5 2.Nc4+ Kd5 3.Nxa5 dxc3 4.Kxc3 Kc5 5.Nb3+ Kb5 6.Kd4 Kb4 7.Na1 Kb5 8.c3 Kc6 9.c4 Kd6 10.Nb3 Kc6 11.c5 Kc7 12.Kd5 Kd7 13.Nd4 Kc7 14.c6 Kc8 15.Nb5
 2/34	00:22,167	 1.382k	62k	+24,16	1.Ra5 Qxa5 2.Nc4+ Kd5 3.Nxa5 dxc3 4.Kxc3 Kc5 5.Nb3+ Kb5 6.Kd4 Kb4 7.Na1 Kb5 8.Kd5 Kb4 9.c4 Kc3 10.c5 Kb4 11.c6 Kb5 12.c7 Kb6 13.c8Q Ka7 14.Nc2 Kb6 15.Qb8+ Ka6 16.Kc5 Ka5 17.Qa7+
Wow! That is fast! Most a-b engines will find it within a millisecond, it is obvious that tactics is not one of Leela's strong points.
"Most" a-b engines will still lose to Leela (with the exception of the very top, for now...), so who cares about the time it takes :mrgreen:
It must be nice living in la-la land :)
Judge without bias, or don't judge at all...
Top
User avatar
CMCanavessi
Posts: 1142
Joined: Thu Dec 28, 2017 4:06 pm
Location: Argentina

Re: will Leela ever solve this ?

Post by CMCanavessi »

glennsamuel32 wrote: ↑Mon Jun 11, 2018 2:47 pm
CMCanavessi wrote: ↑Mon Jun 11, 2018 2:29 pm
Joost Buijs wrote: ↑Mon Jun 11, 2018 1:32 pm

Wow! That is fast! Most a-b engines will find it within a millisecond, it is obvious that tactics is not one of Leela's strong points.
"Most" a-b engines will still lose to Leela (with the exception of the very top, for now...), so who cares about the time it takes :mrgreen:
It must be nice living in la-la land :)
Click my signature, if you dare... and get a cold shower of something called "reality".
Follow my tournament and some Leela gauntlets live at http://twitch.tv/ccls
Top
glennsamuel32
Posts: 136
Joined: Sat Dec 04, 2010 5:31 pm
Location: 223

Re: will Leela ever solve this ?

Post by glennsamuel32 »

CMCanavessi wrote: ↑Mon Jun 11, 2018 2:48 pm
glennsamuel32 wrote: ↑Mon Jun 11, 2018 2:47 pm
CMCanavessi wrote: ↑Mon Jun 11, 2018 2:29 pm

"Most" a-b engines will still lose to Leela (with the exception of the very top, for now...), so who cares about the time it takes :mrgreen:
It must be nice living in la-la land :)
Click my signature, if you dare... and get a cold shower of something called "reality".
Mind elaborating on the game conditions ?
Instead of "reality", I would call this a "Blast-From-The-Past" by playing against stockfish 5 :D
Judge without bias, or don't judge at all...
Top
User avatar
CMCanavessi
Posts: 1142
Joined: Thu Dec 28, 2017 4:06 pm
Location: Argentina

Re: will Leela ever solve this ?

Post by CMCanavessi »

glennsamuel32 wrote: ↑Mon Jun 11, 2018 2:58 pm
CMCanavessi wrote: ↑Mon Jun 11, 2018 2:48 pm
glennsamuel32 wrote: ↑Mon Jun 11, 2018 2:47 pm

It must be nice living in la-la land :)
Click my signature, if you dare... and get a cold shower of something called "reality".
Mind elaborating on the game conditions ?
More than "reality", I would call this a "Blast-From-The-Past" by playing against stockfish 5 :D
True, because "most" a-b engines can easily beat Stockfish 5 (3244 ccrl)...
Follow my tournament and some Leela gauntlets live at http://twitch.tv/ccls
Top
glennsamuel32
Posts: 136
Joined: Sat Dec 04, 2010 5:31 pm
Location: 223

Re: will Leela ever solve this ?

Post by glennsamuel32 »

CMCanavessi wrote: ↑Mon Jun 11, 2018 3:01 pm
glennsamuel32 wrote: ↑Mon Jun 11, 2018 2:58 pm
CMCanavessi wrote: ↑Mon Jun 11, 2018 2:48 pm

Click my signature, if you dare... and get a cold shower of something called "reality".
Mind elaborating on the game conditions ?
More than "reality", I would call this a "Blast-From-The-Past" by playing against stockfish 5 :D
True, because "most" a-b engines can easily beat Stockfish 5 (3244 ccrl)...
Sorry but I can't comprehend your evasive answers.
What are the game conditions and why aren't you using SF 9 ?
Also, I see you had Scorpio using 1 thread in another test -- any reason for that ?
Judge without bias, or don't judge at all...
Top
Milos
Posts: 4190
Joined: Wed Nov 25, 2009 1:47 am

Re: will Leela ever solve this ?

Post by Milos »

Joost Buijs wrote: ↑Mon Jun 11, 2018 1:32 pm
CMCanavessi wrote: ↑Sun Jun 10, 2018 9:39 pm
Spliffjiffer wrote: ↑Sun Jun 10, 2018 3:30 pm [d]8/8/8/3qk3/3p4/R1PK4/2PN4/8 w - - 0 1

Ra5, game over :-)
Finds it in 6 seconds:

Code: Select all

FEN: 8/8/8/3qk3/3p4/R1PK4/2PN4/8 w - - 0 1

Lc0 ID 25 (TEST):
 1/2	00:00,010	 2	200	-1,57	1.cxd4+ Qxd4+
 1/3	00:00,011	 3	272	-0,72	1.cxd4+ Qxd4+ 2.Ke2
 1/4	00:00,012	 6	500	 0,00	1.cxd4+ Qxd4+ 2.Ke2 Qg4+
 1/5	00:00,013	 15	1k	-0,07	1.cxd4+ Qxd4+ 2.Ke2 Qc5 3.Re3+
 1/6	00:00,015	 33	2k	-0,10	1.cxd4+ Qxd4+ 2.Ke2 Qc5 3.Re3+ Kf5
 1/7	00:00,017	 65	4k	-0,01	1.cxd4+ Qxd4+ 2.Ke2 Qc5 3.Re3+ Kd5 4.c4+
 1/8	00:00,018	 97	5k	+0,13	1.cxd4+ Qxd4+ 2.Ke2 Qg4+ 3.Nf3+ Kd6 4.Rd3+ Kc5
 1/9	00:00,020	 148	7k	+0,03	1.cxd4+ Qxd4+ 2.Ke2 Qg4+ 3.Nf3+ Kd6 4.Rd3+ Kc5 5.Ke3
 1/9	00:00,022	 187	9k	+0,68	1.c4 Qc5 2.Ra8 Qb4 3.Nf3+ Kf4 4.Nxd4
 1/10	00:00,026	 282	11k	+0,73	1.c4 Qc5 2.Ra8 Qb4 3.Nf3+ Kf4 4.Nxd4 Qb1
 1/11	00:00,035	 504	14k	+0,38	1.c4 Qg2 2.Ra5+ Kd6 3.Ne4+ Kc7 4.Rf5 Qh3+
 1/12	00:00,076	 2k	21k	+0,16	1.c4 Qg2 2.Ra5+ Kf4 3.c5 Qd5 4.Nb3 Ke5 5.c6 Kd6
 1/13	00:00,099	 2k	23k	+0,01	1.c4 Qg2 2.Ra5+ Kf4 3.c5 Qg6+ 4.Kxd4 Qxc2 5.Nc4 Qe4+ 6.Kc3 Qe1+
 1/13	00:00,111	 3k	24k	+0,20	1.Nc4+ Kf4 2.cxd4 Qf3+ 3.Kd2 Qd5 4.Kc3 Ke4 5.Nd2+ Ke3 6.Nc4+ Ke4 7.Nd2+
 1/14	00:00,123	 3k	24k	+0,25	1.Nc4+ Kf4 2.cxd4 Qf3+ 3.Kd2 Qd5 4.Kc3 Ke4 5.Nd2+ Ke3 6.Nc4+ Ke4 7.Nd2+
 1/15	00:00,143	 4k	25k	+0,23	1.Nc4+ Kf4 2.cxd4 Qf3+ 3.Kd2 Qd5 4.Kc3 Ke4 5.Nd2+ Ke3 6.Nc4+ Ke4 7.Nd2+ Ke3
 1/16	00:00,178	 5k	27k	+0,22	1.Nc4+ Kf4 2.cxd4 Qf3+ 3.Kd2 Qd5 4.Kc3 Ke4 5.Nd2+ Ke3 6.Nc4+ Ke4 7.Nd2+ Ke3 8.Nc4+
 1/17	00:00,223	 6k	29k	+0,25	1.Nc4+ Kf4 2.cxd4 Qf3+ 3.Kd2 Qd5 4.Kc3 Ke4 5.Nd2+ Ke3 6.Nc4+ Ke4 7.Nd2+ Ke3 8.Nc4+ Ke4
 1/19	00:00,396	 13k	32k	+0,25	1.Nc4+ Kf4 2.cxd4 Qf3+ 3.Kd2 Qd5 4.Kc3 Ke4 5.Nd2+ Ke3 6.Nc4+ Ke4 7.Nd2+ Ke3 8.Nc4+ Ke4
 1/20	00:00,776	 28k	36k	+0,30	1.Nc4+ Kf4 2.cxd4 Qf3+ 3.Kd2 Qf2+ 4.Kd3 Qf3+ 5.Kd2 Qd5 6.Kc3 Ke4 7.Nd2+ Ke3 8.Nc4+ Ke4 9.Nd2+ Ke3 10.Nc4+ Kf4
 2/20	00:00,859	 31k	36k	+0,31	1.Nc4+ Kf4 2.cxd4 Qf3+ 3.Kd2 Qf2+ 4.Kd3 Qf3+ 5.Kd2 Qd5 6.Kc3 Ke4 7.Nd2+ Ke3 8.Nc4+ Ke4 9.Nd2+ Ke3 10.Nc4+ Kf4
 2/21	00:01,094	 41k	37k	+0,35	1.Nc4+ Kf4 2.cxd4 Qf3+ 3.Kd2 Qf2+ 4.Kd3 Qf3+ 5.Kd2 Qd5 6.Kc3 Ke4 7.Nd2+ Ke3 8.Nc4+ Ke4 9.Nd2+ Ke3 10.Nc4+ Kf4
 2/22	00:02,602	 103k	39k	+0,43	1.Nc4+ Kf4 2.cxd4 Qf3+ 3.Kd2 Qf2+ 4.Kc3 Ke4 5.Nd2+ Kf4 6.Nb3 Qe1+ 7.Kc4 Qe2+ 8.Kc3 Qe1+ 9.Kb2 Qb4 10.Ra8 Ke4
 2/23	00:04,326	 185k	43k	+0,49	1.Nc4+ Kf4 2.cxd4 Qf3+ 3.Kd2 Qf2+ 4.Kc3 Qe1+ 5.Kb3 Qd1 6.Kc3 Qe1+ 7.Kb3 Qb1+ 8.Kc3 Qe1+
 2/29	00:05,777	 274k	47k	+0,49	1.Nc4+ Kf4 2.cxd4 Qf3+ 3.Kd2 Qf2+ 4.Kc3 Qe1+ 5.Kb3 Qd1 6.Kc3 Qe1+ 7.Kb3 Qb1+ 8.Kc3 Qe1+
 2/29	00:06,320	 310k	49k	+20,27	1.Ra5 Qxa5 2.Nc4+ Kd5 3.Nxa5 dxc3 4.Kxc3 Kc5 5.Nb3+ Kb5 6.Kd4 Kb4 7.c4 Kxb3 8.c5 Kb4 9.c6 Kb5 10.c7 Kc6 11.c8Q+ Kd6 12.Qb8+ Kc6 13.Kc4 Kd7
 2/31	00:07,178	 367k	51k	+20,74	1.Ra5 Qxa5 2.Nc4+ Kd5 3.Nxa5 dxc3 4.Kxc3 Kc5 5.Nb3+ Kb5 6.Kd4 Kb4 7.c4 Kxb3 8.c5 Kb4 9.c6 Kb5 10.c7 Kc6 11.c8Q+ Kd6 12.Qb8+ Kc6 13.Kc4 Kd7
 2/32	00:07,680	 400k	52k	+20,99	1.Ra5 Qxa5 2.Nc4+ Kd5 3.Nxa5 dxc3 4.Kxc3 Kc5 5.Nb3+ Kb5 6.Kd4 Kb4 7.c4 Kxb3 8.c5 Kb4 9.c6 Kb5 10.c7 Kc6 11.c8Q+ Kd6 12.Qb8+ Kc6 13.Kc4 Kd7 14.Kd5
 2/32	00:12,684	 735k	58k	+22,38	1.Ra5 Qxa5 2.Nc4+ Kd5 3.Nxa5 dxc3 4.Kxc3 Kc5 5.Nb3+ Kb5 6.Kd4 Kb4 7.Nd2 Kb5 8.c3 Kc6 9.c4 Kd6 10.c5+ Kc6 11.Nc4 Kc7 12.Kd5 Kd7 13.c6+ Kc7
 2/32	00:17,688	 1.073k	61k	+23,32	1.Ra5 Qxa5 2.Nc4+ Kd5 3.Nxa5 dxc3 4.Kxc3 Kc5 5.Nb3+ Kb5 6.Kd4 Kb4 7.Na1 Kb5 8.c3 Kc6 9.c4 Kd6 10.Nb3 Kc6 11.c5 Kc7 12.Kd5 Kd7 13.Nd4 Kc7 14.c6 Kc8
 2/33	00:19,591	 1.206k	62k	+23,72	1.Ra5 Qxa5 2.Nc4+ Kd5 3.Nxa5 dxc3 4.Kxc3 Kc5 5.Nb3+ Kb5 6.Kd4 Kb4 7.Na1 Kb5 8.c3 Kc6 9.c4 Kd6 10.Nb3 Kc6 11.c5 Kc7 12.Kd5 Kd7 13.Nd4 Kc7 14.c6 Kc8 15.Nb5
 2/34	00:19,828	 1.223k	62k	+23,77	1.Ra5 Qxa5 2.Nc4+ Kd5 3.Nxa5 dxc3 4.Kxc3 Kc5 5.Nb3+ Kb5 6.Kd4 Kb4 7.Na1 Kb5 8.c3 Kc6 9.c4 Kd6 10.Nb3 Kc6 11.c5 Kc7 12.Kd5 Kd7 13.Nd4 Kc7 14.c6 Kc8 15.Nb5
 2/34	00:22,167	 1.382k	62k	+24,16	1.Ra5 Qxa5 2.Nc4+ Kd5 3.Nxa5 dxc3 4.Kxc3 Kc5 5.Nb3+ Kb5 6.Kd4 Kb4 7.Na1 Kb5 8.Kd5 Kb4 9.c4 Kc3 10.c5 Kb4 11.c6 Kb5 12.c7 Kb6 13.c8Q Ka7 14.Nc2 Kb6 15.Qb8+ Ka6 16.Kc5 Ka5 17.Qa7+
Wow! That is fast! Most a-b engines will find it within a millisecond, it is obvious that tactics is not one of Leela's strong points.
6 seconds is irrelevant, it is actually 310 000 playouts which is a lot. Most AB engines find it in QS within 1000 nodes evaluated.
Top
Joost Buijs
Posts: 1563
Joined: Thu Jul 16, 2009 10:47 am
Location: Almere, The Netherlands

Re: will Leela ever solve this ?

Post by Joost Buijs »

glennsamuel32 wrote: ↑Mon Jun 11, 2018 3:06 pm
CMCanavessi wrote: ↑Mon Jun 11, 2018 3:01 pm
glennsamuel32 wrote: ↑Mon Jun 11, 2018 2:58 pm

Mind elaborating on the game conditions ?
More than "reality", I would call this a "Blast-From-The-Past" by playing against stockfish 5 :D
True, because "most" a-b engines can easily beat Stockfish 5 (3244 ccrl)...
Sorry but I can't comprehend your evasive answers.
What are the game conditions and why aren't you using SF 9 ?
Also, I see you had Scorpio using 1 thread in another test -- any reason for that ?
Indeed, Stockfish 5 is doing like 1 to 1.5 million n/s which is very slow, a couple of days ago I even saw ~100 kn/s If you really want to compare strength you have to let Leela face Stockfish 9 at all 8 cores, or even better.
Last edited by Joost Buijs on Mon Jun 11, 2018 3:19 pm, edited 1 time in total.
Top
User avatar
CMCanavessi
Posts: 1142
Joined: Thu Dec 28, 2017 4:06 pm
Location: Argentina

Re: will Leela ever solve this ?

Post by CMCanavessi »

glennsamuel32 wrote: ↑Mon Jun 11, 2018 3:06 pm
CMCanavessi wrote: ↑Mon Jun 11, 2018 3:01 pm
glennsamuel32 wrote: ↑Mon Jun 11, 2018 2:58 pm

Mind elaborating on the game conditions ?
More than "reality", I would call this a "Blast-From-The-Past" by playing against stockfish 5 :D
True, because "most" a-b engines can easily beat Stockfish 5 (3244 ccrl)...
Sorry but I can't comprehend your evasive answers.
What are the game conditions and why aren't you using SF 9 ?
Also, I see you had Scorpio using 1 thread in another test -- any reason for that ?
No evasive nothing. TC is 30 min + 10 sec, 2 move book (with reverse games), 1 thread for SF 5, 1 thread for Lc0, 4gb hash for SF, ponder off, no TBs.
I'm using SF 5 because that's the purpose of this series in the stream, Leela has been beating the "yearly CCRL champions", Rybka 4, Houdini 1.5a, etc... right now at year 2014. If she can beat SF 5 in the 10-game match, we'll proceed to 2015+
Follow my tournament and some Leela gauntlets live at http://twitch.tv/ccls
Top

Return to “Computer Chess Club: General Topics”