Hard talkchess 2018 set - work in progress

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

Moderators: hgm, Dann Corbit, Harvey Williamson

Forum rules
This textbox is used to restore diagrams posted with the [d] tag before the upgrade.
Vinvin
Posts: 5009
Joined: Thu Mar 09, 2006 8:40 am
Full name: Vincent Lejeune

Re: Hard talkchess 2018 set - work in progress

Post by Vinvin » Tue Jul 27, 2021 9:39 am

Vinvin wrote:
Sun Apr 15, 2018 8:46 am
Vinvin wrote:...
9)[d]2bqrr1k/p5b1/1p1p2pp/nPpBp3/P1P1N2P/3PN1P1/R4P1K/3Q1R2 w - - 0 1
After a bit of learning. This one should require may be an hour to get an accurate evaluation.
Stockfish_180407_x64_modern:
31/51 01:08 652 451k 9 553k +1,76 1.Kg1 Qe7 2.h5 g5 3.g4 Be6 4.Ng3 Rf4 5.Be4 Ref8 6.f3 Qd7 7.Raf2 Qd8 8.Re1 Qd7 9.Rd2 Rd8 10.Nd5 Rff8 11.Rf2 Rf7 12.Ne3 Rff8 13.Ree2 Rf7 14.Bg6 Rff8 15.Ngf5 d5 16.Nxg7 Qxg7 17.cxd5 Bxd5
31/35 01:08 652 451k 9 553k +1,76 1.h5 g5 2.g4 Qe7 3.Ng3 Rf4 4.f3 Be6 5.Be4 Qd8 6.Nd5 Rff8 7.Raf2 Qd7 8.Ne3 Rd8 9.Re2 Rf7 10.Kg1 Rff8 11.Rff2 Rf7 12.Bg6 Rff8 13.Ngf5 d5 14.Nxg7 Qxg7 15.cxd5 Bxd5 16.Nf5 Bb3
31/49 01:08 652 451k 9 553k +1,93 1.Nc3 Be6 2.Bxe6 Rxe6 3.Ned5 Kh7 4.Qg4 Ree8 5.h5 g5 6.Ne4 Kh8 7.Rb2 Rf7 8.Kg2 Nb7 9.Re2 Qc8 10.f3 Kg8 11.Ra2 Kh7 12.Qxc8 Rxc8 13.g4 Rd8 14.Ng3 Rdd7 15.Nf5 Na5 16.Rb1 Nb7 17.Kg3 Nd8 18.a5 bxa5 19.Rxa5 Ne6 20.Nde3 Nd4 21.Kf2 Ne6 22.Rd1 Nd4
---------------------------------------------------------------------------
32/46 01:38 951 091k 9 651k +1,75 1.Kg1 Qe7 2.h5 g5 3.g4 Be6 4.Ng3 Rf4 5.Re1 Qd7 6.Bxe6 Qxe6 7.Ngf5 Qd7 8.f3 Rf8 9.Rf1 Rf7 10.Rff2 Nb7 11.Kh1 Na5 12.Kg2 Rf8 13.Rae2 Rf7 14.Rf1 Rf8 15.Ref2 Rf7 16.Re1 Nb7 17.Ra2 Na5 18.Rf1
32/37 01:38 951 091k 9 651k +1,83 1.h5 g5 2.g4 Qe7 3.Ng3 Rf4 4.f3 Be6 5.Be4 Qd8 6.Nd5 Rff8 7.Raf2 Qd7 8.Kg2 Rb8 9.Re2 Rb7 10.Ne3 Rbb8 11.Rff2 Rbe8 12.Nd5 Rd8 13.Qc2 Rc8 14.Rf1 Rcd8 15.Bg6 Bf7 16.Bf5 Be6
32/50 01:38 951 091k 9 651k +2,06 1.Nc3 h5 2.Be4 Re6 3.f3 Bh6 4.Ned5 Kg7 5.Raf2 Bb7 6.Qc2 Kh8 7.Kg2 Rf7 8.Qd1 Kg7 9.Kg1 Re8 10.g4 hxg4 11.fxg4 Rxf2 12.Rxf2 Bxd5 13.Nxd5 Qxh4 14.Qf3 Kh8 15.Qf6+ Qxf6 16.Rxf6 Bg5 17.Rxd6 Rg8 18.Re6 Nb3 19.Rxe5 Bd2 20.Re7 Rg7
---------------------------------------------------------------------------
33/35 02:12 1 282 845k 9 656k +1,75 1.Kg1 Qe7 2.h5 g5 3.g4 Be6 4.Ng3 Rf4 5.Re1 Qd7 6.Bxe6 Qxe6 7.Ngf5 Kh7 8.f3 Rf8 9.Rf2 Kh8 10.Ref1 Qd7 11.Kg2 Rf7 12.Re2 Rf8 13.Rff2 Rf7 14.Ra2 Rf8 15.Kf1 Nb7 16.Qb1 Rf7 17.Kg1 Rf8 18.Rf1
33/50 02:12 1 282 845k 9 656k +1,76 1.h5 g5 2.g4 Rf4 3.Ng3 Qe7 4.Be4 Qd8 5.f3 Be6 6.Nd5 Rff8 7.Re2 Qd7 8.Kg2 Rd8 9.Qc2 Rb8 10.Kg1 Rbc8 11.Rff2 Rcd8 12.Kg2 Rde8 13.Bg6 Bf7 14.Bf5 Be6 15.Bxe6 Rxe6 16.Nf5 Ree8 17.Qd1 Nb7
33/61 02:12 1 282 845k 9 656k +1,93 1.Nc3 Qd7 2.Be4 Qe6 3.Ncd5 Rd8 4.Kg1 Bb7 5.h5 g5 6.g4 Rb8 7.Bg6 Qd7 8.Re1 Qd8 9.Bf5 Qe8 10.Rd2 Qd8 11.Rde2 Qe8 12.Ra2 Qd8 13.Rf1 Rf7 14.Bg6 Rf8 15.Re2 Bc8 16.Nf5 Rb7 17.Re4 Be6 18.Kh2 Bxd5 19.cxd5 Rf6
Depth 55 with 3 best moves (after 8 hours), eval for Nc3 continues to climb (2.10, 2.29, 2.46, 2.55) :

Code: Select all

 51/62	1:10:28	 44 182 440k	10 448k	+1,88	1.Kg1 Qd7 2.g4 Qe7 3.h5 g5 4.Ng3 Rf4 5.Kg2 Qd8 6.Rd2 Qc7 7.f3 Rff8 8.Ngf5 Rf6 9.Ra2 Ref8 10.Ng3 Rf4 11.Ne2 R4f6 12.Ra1 Qd8 13.Ng3 Re8 14.Be4 Ref8 15.Ra3 Be6 16.Bf5 Bxf5 17.Ngxf5 Qd7 18.Ng3 R6f7 19.Nd5 Nb7 20.Qe2 Nd8 21.a5 Ne6 22.axb6 axb6 23.Qe4 Nd4 24.Nxb6 Qd8 25.Nd5 Nxf3 26.Rxf3 Rxf3 27.Qxf3
 51/52	1:10:28	 44 182 440k	10 448k	+1,88	1.h5 g5 2.g4 Rf4 3.Ng3 Rf6 4.Ra1 Ref8 5.Kg2 Rf4 6.Rb1 Re8 7.f3 Rff8 8.Nef5 Rf6 9.Qe1 Rff8 10.Qe2 Qc7 11.Qe4 Rd8 12.Qe1 Rde8 13.Qe2 Rd8 14.Ne3 Qe7 15.Ngf5 Qe8 16.Kf2 Rf6 17.Ng3 Rf8 18.Kg1 Qe7 19.Ngf5 Qe8 20.Kg2 Rf6 21.Qc2 Qf8 22.Ng3 Qe7 23.Qf2 Rff8 24.Qe2 Qc7 25.Qd1 Rde8 26.Be4 Be6
 51/65	1:10:28	 44 182 440k	10 448k	+2,10	1.Nc3 Bf5 2.h5 g5 3.Nxf5 Rxf5 4.Be4 Rff8 5.Kg2 Re7 6.Nd5 Ref7 7.g4 Qd7 8.f3 Qe6 9.Bg6 Rd7 10.Kh1 Rb7 11.Rh2 Rbb8 12.Re2 Rbd8 13.Kg2 Qd7 14.Ra2 Nb7 15.Be4 Na5 16.Kg1 Qe6 17.Bg6 Qd7 18.Re2 Qe6 19.Be4 Qd7 20.Kg2 Qf7 21.Ne3 Rb8 22.Ra2 Qd7 23.Bg6 Nb7 24.Bf5 Qe8 25.Nd5 Na5 26.Re1 Qf7
---------------------------------------------------------------------------
 52/55	3:01:47	115 523 470k	10 591k	+1,88	1.Kg1 Qd7 2.g4 Qe7 3.h5 g5 4.Ng3 Rf4 5.Kg2 Qd8 6.Rd2 Qc7 7.f3 Rff8 8.Ngf5 Rf6 9.Ra2 Ref8 10.Ng3 Rf4 11.Ne2 R4f6 12.Ra1 Re8 13.Ng3 Rff8 14.Rb1 Nb7 15.Be4 Be6 16.Ngf5 Na5 17.Bd5 Bc8 18.Qe2 Qd8 19.Qc2 Qd7 20.Ng3 Qd8 21.Kg1 Rf6 22.Ne4 Rff8 23.Qe2 Qc7 24.Kg2 Rd8 25.Ng3 Qe7 26.Ngf5 Qe8 27.Kf2 Rf6 28.Ng3
 52/57	3:01:47	115 523 470k	10 591k	+1,88	1.h5 g5 2.g4 Rf4 3.Ng3 Rf6 4.Ra1 Ref8 5.Kg2 Rf4 6.Rb1 Re8 7.f3 Rff8 8.Nef5 Rf6 9.Qe1 Rff8 10.Qe2 Qc7 11.Qe4 Rd8 12.Qe1 Rde8 13.Qe2 Rd8 14.Ne3 Qe7 15.Ngf5 Qe8 16.Kf2 Rf6 17.Ng3 Rf8 18.Kg1 Qe7 19.Ngf5 Qe8 20.Kg2 Rf6 21.Be4 Be6 22.Qe1 Qf7 23.Ng3 Rf8 24.Qe2 Qd7 25.Bd5 Bf7 26.Bxf7 R6xf7 27.Nd5 Nb7 28.Qe4 Nd8 29.a5
 52/77	3:01:47	115 523 470k	10 591k	+2,29	1.Nc3 Kh7 2.Be4 Bf5 3.h5 Bxe4 4.dxe4 gxh5 5.Qxh5 Qd7 6.Ncd5 Qf7 7.Qd1 Qf3 8.Qxf3 Rxf3 9.Kg2 Rf7 10.Rb1 Bf6 11.Nf5 Re6 12.Rc2 Bg7 13.Rh1 Nb3 14.Rc3 Na5 15.Kf3 Rg6 16.Ke3 Bf8 17.Ke2 Re6 18.Nde3 Kg8 19.Rh5 Rh7 20.Nd5 Rf7 21.Kd2 Rg6 22.Ke3 Kh7 23.Ke2 Kg8 24.Rh1 Re6 25.Ke3 Rg6 26.Kd3 Rg5 27.Nde3 Rh7 28.Ke2
---------------------------------------------------------------------------
 53/55	3:16:00	124 623 041k	10 596k	+1,88	1.Kg1 Qd7 2.g4 Qe7 3.h5 g5 4.Ng3 Rf4 5.Kg2 Qd8 6.Rd2 Qc7 7.f3 Rff8 8.Ngf5 Rf6 9.Ra2 Ref8 10.Ng3 Rf4 11.Ne2 R4f6 12.Ra1 Re8 13.Ng3 Rff8 14.Rb1 Nb7 15.Be4 Be6 16.Ngf5 Na5 17.Bd5 Bc8 18.Qe2 Qd8 19.Qc2 Qd7 20.Ng3 Qd8 21.Kg1 Rf6 22.Qe2 Qe7 23.Kg2 Bd7 24.Qd1 Bc8 25.Be4 Be6 26.Bg6 Rb8 27.Ne4 Rff8 28.Nf5
 53/55	3:16:00	124 623 041k	10 596k	+1,88	1.h5 g5 2.g4 Rf4 3.Ng3 Rf6 4.Ra1 Ref8 5.Kg2 Rf4 6.Rb1 Re8 7.f3 Rff8 8.Nef5 Rf6 9.Qe1 Rff8 10.Qe2 Qc7 11.Qe4 Rd8 12.Qe1 Rde8 13.Qe2 Rd8 14.Ne3 Qe7 15.Ngf5 Qe8 16.Kf2 Rf6 17.Ng3 Rf8 18.Kg1 Qe7 19.Ngf5 Qe8 20.Kg2 Rf6 21.Be4 Be6 22.Qe1 Qf7 23.Ng3 Rf8 24.Qe2 Qd7 25.Bd5 Re8 26.Bxe6 Rfxe6 27.Nd5 Rf8 28.Nf5
 53/59	3:16:00	124 623 041k	10 596k	+2,29	1.Nc3 Kh7 2.Be4 Bf5 3.h5 Bxe4 4.dxe4 gxh5 5.Qxh5 Qd7 6.Ncd5 Qf7 7.Qd1 Qf3 8.Qxf3 Rxf3 9.Kg2 Rf7 10.Rb1 Bf6 11.Nf5 Re6 12.Rc2 Bg7 13.Rh1 Nb3 14.Rc3 Na5 15.Kf3 Kg8 16.Ke2 Bf8 17.Rh4 Rg6 18.Rh5 Re6 19.Kd2 Rg6 20.Ke3 Kh7 21.f3 Re6 22.Ke2 Kg8 23.Kf2 Rh7 24.Rc2 Rf7 25.Kg2 Rh7 26.f4 exf4 27.gxf4 Rg6+ 28.Kf2 Rf7 29.Rc3
---------------------------------------------------------------------------
 54/57	6:31:04	249 349 642k	10 627k	+1,88	1.Kg1 Qd7 2.g4 Qe7 3.h5 g5 4.Ng3 Rf4 5.Kg2 Qd8 6.Rd2 Qc7 7.f3 Rff8 8.Ngf5 Rf6 9.Ra2 Ref8 10.Ng3 Qb8 11.Ne4 Rf4 12.Nc3 Rd8 13.Ne2 Rff8 14.Ng3 Qc7 15.Ne4 Rde8 16.Ra1 Qd8 17.Rb1 Rf4 18.Ng3 Rff8 19.Nef5 Rf6 20.Qe1 Rff8 21.Qe2 Qc7 22.Qe4 Rd8 23.Qe1 Rde8 24.Qe2 Rd8 25.Kf2 Rde8 26.Ne3 Nb7 27.Bc6 Bd7 28.Be4 Be6 29.Ngf5
 54/69	6:31:04	249 349 642k	10 627k	+1,88	1.h5 g5 2.g4 Rf4 3.Ng3 Rf6 4.Ra1 Ref8 5.Kg2 Rf4 6.Rb1 Re8 7.f3 Rff8 8.Nef5 Rf6 9.Qe1 Rff8 10.Qe2 Qc7 11.Qe4 Rd8 12.Qe1 Rde8 13.Qe2 Rd8 14.Kf2 Rde8 15.Ne3 Nb7 16.Bc6 Bd7 17.Be4 Be6 18.Ngf5 Na5 19.Bd5 Bc8 20.Kg2 Qd8 21.Kg1 Rf6 22.Rf2 Ref8 23.Ng3 Rf4 24.Ne4 Bb7 25.Rff1 Qd7 26.Ng3 Bxd5 27.Nxd5 Rd4 28.Ne4 Rd8 29.Kg2 Rf8
 54/81	6:31:04	249 349 642k	10 627k	+2,46	1.Nc3 Re7 2.Be4 Bf5 3.h5 Bxe4 4.dxe4 g5 5.Kg2 Qe8 6.Ncd5 Rd7 7.Ra3 Rdf7 8.Qe2 Rf3 9.Rd3 Qe6 10.Nf5 Rxd3 11.Qxd3 Rd8 12.Rd1 Rd7 13.Kg1 Kh7 14.g4 Kh8 15.Qc2 Kh7 16.Kf1 Rf7 17.Qc3 Kh8 18.Kg2 Rd7 19.Qd3 Rf7 20.Kf1 Nb7 21.Kg1 Bf8 22.Kg2 Rd7 23.Qc3 Na5 24.Kf1 Bg7 25.Ke2 Nb7 26.Qd3 Qf7 27.Kf1 Qe6 28.Rd2 Na5
---------------------------------------------------------------------------
 55/59	8:00:09	307 103 330k	10 660k	+1,88	1.Kg1 Qd7 2.g4 Qe7 3.h5 g5 4.Ng3 Rf4 5.Kg2 Qd8 6.Rd2 Qc7 7.f3 Rff8 8.Ngf5 Rf6 9.Ra2 Ref8 10.Ng3 Qb8 11.Ne4 Rf4 12.Nc3 Rd8 13.Ne2 Rff8 14.Ng3 Qc7 15.Ne4 Rde8 16.Ra1 Qd8 17.Rb1 Rf4 18.Ng3 Rff8 19.Nef5 Rf6 20.Qe1 Rff8 21.Qe2 Qc7 22.Qe4 Rd8 23.Qe1 Rde8 24.Qe2 Rd8 25.Ne3 Bd7 26.Ngf5 Rde8 27.Qe1 Bc8 28.Ng3 Qd8 29.Ne4 Qc7 30.Kg1
 55/60	8:00:09	307 103 330k	10 660k	+1,88	1.h5 g5 2.g4 Rf4 3.Ng3 Rf6 4.Ra1 Ref8 5.Kg2 Rf4 6.Rb1 Re8 7.f3 Rff8 8.Nef5 Rf6 9.Qe1 Rff8 10.Qe2 Qc7 11.Qe4 Rd8 12.Qe1 Rde8 13.Qe2 Rd8 14.Ne3 Bd7 15.Ngf5 Rde8 16.Qc2 Bc8 17.Qd1 Qd8 18.Qe1 Qc7 19.Ng3 Rd8 20.Qf2 Rde8 21.Ngf5 Qd7 22.Be4 Qd8 23.Qe2 Be6 24.Ng3 Re7 25.Qe1 Bg8 26.Ngf5 Rd7 27.Bc6 Nxc6 28.bxc6 Rdf7
 55/72	8:00:09	307 103 330k	10 660k	+2,55	1.Nc3 Qd7 2.Be4 Qe6 3.Ncd5 Rf7 4.Kg1 Ref8 5.g4 Bb7 6.f3 Kg8 7.Rg2 Kh8 8.Rgf2 Rf4 9.Rh2 R4f7 10.Rg2 Kg8 11.Rgf2 Ba8 12.h5 g5 13.Bf5 Qe8 14.Bg6 Qe6 15.Bxf7+ Rxf7 16.Kg2 Kf8 17.Qe2 Kg8 18.Qa2 Bb7 19.Rb1 Kh8 20.Qc2 Bc8 21.Qe2 Qe8 22.Nc3 Be6 23.Rff1 Qd7 24.Ne4 Kg8 25.Qd1 Rf8 26.Nd5 Kh8 27.Kg1 Qd8 28.Kg2 Rf7 29.Kg1
Depth 58 with a recent Stockfish, probably 250 Elo higher as the previous test : Stockfish_21071317_x64_avx2
Second move is so high, I will probably delete this position because it's not good enough.

Code: Select all

 57/10	7:11:16	704.951.344k	27.243k	+3,09	1.h5 g5
 58/74	7:11:16	704.951.344k	27.243k	+3,25	1.Kg2 Qd7 2.Rh1 Rd8 3.Re2 Rde8 4.Ree1 Re7 5.Rh2 Bb7 6.Bxb7 Nxb7 7.h5 g5 8.Rhh1 Kg8 9.Qe2 Qe6 10.Rhf1 Rd7 11.Rb1 Rdf7 12.Qd1 Na5 13.f3 Kh8 14.Rf2 Nb7 15.Ra1 Kg8 16.Rf1 Na5 17.Ra2 Kh8 18.Nd5 Qd7 19.Rff2 Nb7 20.Qh1 Na5 21.Rab2 Qe6 22.Qf1 Kg8 23.Qd1 Kh8 24.Rb1 Nb7 25.Qh1 Na5 26.Rbf1 Bf6 27.Nexf6 Rxf6 28.Nxf6 Qxf6 29.Qh3 Nb3 30.f4 gxf4
 58/78	7:11:16	704.951.344k	27.243k	+4,58	1.Nc3 Be6 2.Be4 Bf5 3.h5 Bxe4 4.dxe4 Qg5 5.hxg6 Qxg6 6.Kg2 Rf7 7.Rd2 Re6 8.Rd3 Rf8 9.Qc2 Qf7 10.Nf5 Nxc4 11.Nd5 Na5 12.Rf3 Qd7 13.Rh1 c4 14.Qd1 Nb3 15.Rc3 Nc5 16.Rxc4 Rg6 17.Rh5 Kh7 18.Qe1 Qe6 19.Qe3 Rgf6 20.Qf3 R6f7 21.Qg4 Qg6 22.Qxg6+ Kxg6 23.g4 Rd7 24.f3 Rfd8 25.Kg3 Rh8 26.Nh4+ Kf7 27.Rc1 Rc8 28.Kg2 Bf6 29.Rxh6 Bxh4 30.Rxh4 Kg6

Spliffjiffer
Posts: 336
Joined: Thu Aug 02, 2012 5:48 pm
Location: Germany

Re: Hard talkchess 2018 set - work in progress

Post by Spliffjiffer » Tue Jul 27, 2021 6:15 pm

i had some decent analysis sessions already on this position after u asked for help in another thread about 3-4 years ago (i did not answer because i could not work out anything decisive) and i was aiming to show that 1.h5 eg is winning as well but all my efforts failed to manage so...to remove this pos might be premature imo !?....if not premature than u have to delete others too....HTC.069,4R3/P7/2bB2p1/r3pk2/2p2p1P/2P2P1K/6P1/8 w - - bm Rxe5+ for example

greetings,Peer
Wahrheiten sind Illusionen von denen wir aber vergessen haben dass sie welche sind.

Vinvin
Posts: 5009
Joined: Thu Mar 09, 2006 8:40 am
Full name: Vincent Lejeune

Re: Hard talkchess 2018 set - work in progress

Post by Vinvin » Tue Jul 27, 2021 10:26 pm

Spliffjiffer wrote:
Tue Jul 27, 2021 6:15 pm
i had some decent analysis sessions already on this position after u asked for help in another thread about 3-4 years ago (i did not answer because i could not work out anything decisive) and i was aiming to show that 1.h5 eg is winning as well but all my efforts failed to manage so...to remove this pos might be premature imo !?....if not premature than u have to delete others too....HTC.069,4R3/P7/2bB2p1/r3pk2/2p2p1P/2P2P1K/6P1/8 w - - bm Rxe5+ for example

greetings,Peer
Thanks for your message !
I already also spent a lot of time on 1.h5 with no definitive eval. But it seems winning and I don't want to spend more time on it.

About HTC.069,4R3/P7/2bB2p1/r3pk2/2p2p1P/2P2P1K/6P1/8 w - - bm Rxe5+ I plan to remove it also. The eval for the second move is above 8.

I feel these 2 test positions are interesting, but may be not for everybody.

Andrew
Posts: 187
Joined: Wed Mar 08, 2006 11:51 pm
Location: Australia

Re: Hard talkchess 2018 set - work in progress

Post by Andrew » Wed Jul 28, 2021 9:02 am

Vinvin wrote:
Tue Jul 27, 2021 10:26 pm
Spliffjiffer wrote:
Tue Jul 27, 2021 6:15 pm
i had some decent analysis sessions already on this position after u asked for help in another thread about 3-4 years ago (i did not answer because i could not work out anything decisive) and i was aiming to show that 1.h5 eg is winning as well but all my efforts failed to manage so...to remove this pos might be premature imo !?....if not premature than u have to delete others too....HTC.069,4R3/P7/2bB2p1/r3pk2/2p2p1P/2P2P1K/6P1/8 w - - bm Rxe5+ for example

greetings,Peer
Thanks for your message !
I already also spent a lot of time on 1.h5 with no definitive eval. But it seems winning and I don't want to spend more time on it.

About HTC.069,4R3/P7/2bB2p1/r3pk2/2p2p1P/2P2P1K/6P1/8 w - - bm Rxe5+ I plan to remove it also. The eval for the second move is above 8.

I feel these 2 test positions are interesting, but may be not for everybody.
Using multi-pv analysis on this last position with SF 14, it probably has 6 winning moves, including a8Q and a8R !

Andrew

Vinvin
Posts: 5009
Joined: Thu Mar 09, 2006 8:40 am
Full name: Vincent Lejeune

Re: Hard talkchess 2018 set - work in progress

Post by Vinvin » Mon Aug 02, 2021 11:24 pm

Vinvin wrote:
Tue Jul 27, 2021 10:26 pm
Spliffjiffer wrote:
Tue Jul 27, 2021 6:15 pm
i had some decent analysis sessions already on this position after u asked for help in another thread about 3-4 years ago (i did not answer because i could not work out anything decisive) and i was aiming to show that 1.h5 eg is winning as well but all my efforts failed to manage so...to remove this pos might be premature imo !?....if not premature than u have to delete others too....HTC.069,4R3/P7/2bB2p1/r3pk2/2p2p1P/2P2P1K/6P1/8 w - - bm Rxe5+ for example

greetings,Peer
Thanks for your message !
I already also spent a lot of time on 1.h5 with no definitive eval. But it seems winning and I don't want to spend more time on it.
I got my new computer 2 weeks ago : an AMD Ryzen 9 5950X (16-Core) running at 4 GHz with 128 GB of memory.
I let it run on this position to know if h5 is winning or not :

1) 3 best moves at depth 64 with Honey-v14-x86-64-avx2

Code: Select all

 64/61	9:49:38	890.273.862k	25.165k	+1,99	1.Kg2 Qd7 2.g4 Qd8 3.h5 g5 4.Nc3 Rf4 5.Be4 Be6 6.f3 Ref8 7.Ncd5 R4f7 8.Bg6 Rb7 9.Rb2 Rb8 10.Bf5 Bg8 11.Rb1 Rb7 12.Be4 Rbf7 13.Nc3 Rf4 14.Ne2 Be6 15.Ng3 Qd7 16.Qe1 Bg8 17.Ne2 R4f6 18.Nc3 Rb8 19.Rb2 Rf4 20.Qd1 Rbf8 21.Ncd5 Rd8 22.Re2 Rff8 23.Ra2 Be6 24.Rff2 Bg8 25.Qc2 Rde8 26.Kg3 Be6 27.Rf1 Rb8 28.Qd1 Nb7 29.Kg2 Nd8 30.a5 Bxd5 31.Bxd5
 64/66	9:49:38	890.273.862k	25.165k	+1,99	1.h5 g5 2.g4 Rf4 3.f3 Qd7 4.Ng3 Ref8 5.Kg2 Qd8 6.Raf2 Re8 7.Be4 Be6 8.Nd5 Rff8 9.Re1 Qb8 10.Nf1 Rd8 11.Ne7 Rf7 12.Nf5 Qc7 13.N1e3 Qd7 14.Ng3 Rdf8 15.Ref1 Kg8 16.Nd5 Nb7 17.Ra2 Na5 18.Ra3 Kh8 19.Qe2 Qd8 20.Ne3 Rf4 21.Bd5 Bc8 22.Qc2 Qe7 23.Be4 Be6 24.Ra2 Qd8 25.Qb1 Qe7 26.Nd5 Qd8 27.Qd1 R4f7 28.Qc2 Qd7 29.Ne3 Rf4 30.Bd5 Bf6 31.Bxe6 Qxe6 32.Ra3 Rd4
 64/77	9:49:38	890.273.862k	25.165k	+3,33	1.Nc3 Be6 2.Be4 Bf5 3.h5 Bxe4 4.dxe4 gxh5 5.Qxh5 Qd7 6.Ncd5 Qf7 7.Qd1 Qf3 8.Qxf3 Rxf3 9.Kg2 Rf7 10.Nf5 Re6 11.Rc2 Bf8 12.Rc3 Kh7 13.Rh1 Kg8 14.Rh5 Rh7 15.Kf3 Rf7 16.Ke2 Rg6 17.Rh1 Kh7 18.Rh3 Kg8 19.f3 Rh7 20.g4 Kf7 21.Rc1 Rg5 22.Kf2 Rg6 23.Rh2 Re6 24.Ke3 Rg6 25.Ke2 Rg5 26.Nde3 Kf6 27.Rd1 Nb7 28.Kf2 Kg6 29.Kg3 Kf7 30.Rdh1 Kg6 31.Rg1 h5 32.Kf2 hxg4 33.Rxh7 Kxh7 34.Rh1+ Kg6
2) 3 best moves after 1.h5?! with Stockfish_21071317_x64_avx2 on depth 42 (1...g5 is clearly the best):

Code: Select all

FEN: 2bqrr1k/p5b1/1p1p2pp/nPpBp2P/P1P1N3/3PN1P1/R4P1K/3Q1R2 b - - 0 1

Stockfish_21071317_x64_avx2:
 42/60	09:46	 17.746.610k	30.241k	+6,74	1. ... gxh5 2.Qxh5 Re7 3.Kg2 Be6 4.Rh1 Qe8 5.Qg5 Bxd5 6.Nxd5 Re6 7.Nc7 Rg6 8.Nxe8 Rxg5 9.Nxg5 Rxe8 10.Ne4 d5 11.Nd6 Re7 12.cxd5 Rd7 13.Nc4 Rxd5 14.Nxa5 bxa5 15.Rc2 Kh7 16.Rc4 Kg6 17.Rd1 Bf6 18.Kf1 Kf5 19.Ke2 Bg5 20.Rh1 Kg6 21.Rb1 Bd8 22.Re4 Kf5 23.Rf1 Bg5 24.Rh1 Kg6 25.Rb1 Bd8 26.f3 Kf5 27.Rh1 Bg5 28.Rc4 Kg6 29.Rb1 Bd8 30.Rbc1 Be7 31.Rg4+ Kf7
 43/62	09:46	 17.746.610k	30.241k	+6,58	1. ... Bf5 2.hxg6 Bxg6 3.Qg4 Bh7 4.Kg2 Qc8 5.Qh4 Qd8 6.Qh5 Re7 7.Rh1 Qe8 8.Qh4 Rd7 9.Qh3 Rd8 10.Rb2 Qg6 11.Rbb1 Rf4 12.f3 Rdf8 13.Rbf1 Bg8 14.Qh5 Qxh5 15.Rxh5 Bf7 16.gxf4 Bxh5 17.f5 Be8 18.Rh1 Bd7 19.Nxd6 Rf6 20.Nf7+ Kh7 21.Ng5+ Kh8 22.Ne6 Kh7 23.Be4 Bxe6 24.fxe6+ Kg8 25.Rh5 Rxe6 26.Bd5 Kf7 27.Nf5 Kg6 28.Nxg7 Rf6
 43/45	09:46	 17.746.610k	30.241k	+3,16	1. ... g5 2.Ra3 Qd7 3.Rh1 Qh3+ 4.Kg1 Qd7 5.Kg2 Re7 6.Ra2 Nb7 7.Kg1 Na5 8.Re2 Ree8 9.Kg2 Qe7 10.Rb2 Bd7 11.Rf1 Bc8 12.Rb1 Qd7 13.Rh1 Re7 14.Qe2 Qd8 15.Qd2 Nb7 16.Ra1 Na5 17.Qc2 Qd7 18.Ra2 Qd8 19.Nc3 Ree8 20.Qd1 Qc7 21.Re2 Re7 22.Be4 Ref7 23.Ra2 Be6
3) After 1.h5 g5 with Stockfish_21071317_x64_avx2 on depth 74 !!
>+3 eval is probably enouhg to win.

Code: Select all

FEN: 2bqrr1k/p5b1/1p1p3p/nPpBp1pP/P1P1N3/3PN1P1/R4P1K/3Q1R2 w - - 0 2

Stockfish_21071317_x64_avx2:
NNUE evaluation using nn-9e3c6298299a.nnue enabled
...
 73/79	1:53:33	176.053.912k	25.840k	+3,09	2.Re1 Qd7 3.Rh1 Qe7 4.Re2 Qd7 5.Kg2 Qd8 6.Ree1 Re7 7.Ref1 Qd7 8.Qc2 Nb7 9.Bc6 Qe6 10.Rb1 Na5 11.Bd5 Qd7 12.Qe2 Qd8 13.Rh3 Qd7 14.Rh2 Nb7 15.Ra1 Na5 16.Ra3 Qc7 17.Rh1 Ree8 18.Ra2 Re7 19.Qd1 Qd7 20.Ra3 Ree8 21.Nc3 Qe7 22.Ra2 Rd8 23.Re2 Rde8 24.Be4 Be6 25.Rb2 Rb8 26.Ncd5 Qd7 27.Rb1 Rbe8 28.Bg6 Rb8 29.Qe2 Nb7 30.Be4 Rf7 31.Qd1 Rbf8 32.f3 Nd8 33.Rb2 Bxd5 34.Nxd5 Ne6 35.Ra2 Nd4 36.Rf1 Bf6 37.a5 Bd8 38.axb6 Bxb6 39.Ne3 Qe6 40.Bd5 Qd7 41.Bxf7
 74/86+	5:22:44	500.766.346k	25.861k	+3,17	2.Re1

Spliffjiffer
Posts: 336
Joined: Thu Aug 02, 2012 5:48 pm
Location: Germany

Re: Hard talkchess 2018 set - work in progress

Post by Spliffjiffer » Tue Aug 03, 2021 10:24 am

Vinvin wrote:
Mon Aug 02, 2021 11:24 pm


3) After 1.h5 g5 with Stockfish_21071317_x64_avx2 on depth 74 !!
>+3 eval is probably enouhg to win.

Code: Select all

FEN: 2bqrr1k/p5b1/1p1p3p/nPpBp1pP/P1P1N3/3PN1P1/R4P1K/3Q1R2 w - - 0 2

Stockfish_21071317_x64_avx2:
NNUE evaluation using nn-9e3c6298299a.nnue enabled
...
 73/79	1:53:33	176.053.912k	25.840k	+3,09	2.Re1 Qd7 3.Rh1 Qe7 4.Re2 Qd7 5.Kg2 Qd8 6.Ree1 Re7 7.Ref1 Qd7 8.Qc2 Nb7 9.Bc6 Qe6 10.Rb1 Na5 11.Bd5 Qd7 12.Qe2 Qd8 13.Rh3 Qd7 14.Rh2 Nb7 15.Ra1 Na5 16.Ra3 Qc7 17.Rh1 Ree8 18.Ra2 Re7 19.Qd1 Qd7 20.Ra3 Ree8 21.Nc3 Qe7 22.Ra2 Rd8 23.Re2 Rde8 24.Be4 Be6 25.Rb2 Rb8 26.Ncd5 Qd7 27.Rb1 Rbe8 28.Bg6 Rb8 29.Qe2 Nb7 30.Be4 Rf7 31.Qd1 Rbf8 32.f3 Nd8 33.Rb2 Bxd5 34.Nxd5 Ne6 35.Ra2 Nd4 36.Rf1 Bf6 37.a5 Bd8 38.axb6 Bxb6 39.Ne3 Qe6 40.Bd5 Qd7 41.Bxf7
 74/86+	5:22:44	500.766.346k	25.861k	+3,17	2.Re1
as i sayed: im sceptical...my mainline went with 1.h5 g5 2.f3 and my efforts ended in draws...my suggestion would be (if u dont like to investigate into the lines manually):
take the pos after 1.h5 g5 2.f3 Be6 (as your depth 74 !! search is not able to produce anything better) and have another try with your new monster (congrats, awesome, i had the same idea but decided for a cheap r5 3600)...if it still thinks its 3.17 then u had at least a deeper look ;-) and your decision is a bit better backuped.
if ure interested in, my main efforts went into a line like this (unfortunately my SSD died so im not sure if it was 2...Be6 or 2...qe7): 1.h5 g5 2.f3 Be6 3.g4 Qe7
and i thought white cant win !??
Wahrheiten sind Illusionen von denen wir aber vergessen haben dass sie welche sind.

Spliffjiffer
Posts: 336
Joined: Thu Aug 02, 2012 5:48 pm
Location: Germany

Re: Hard talkchess 2018 set - work in progress

Post by Spliffjiffer » Tue Aug 03, 2021 5:39 pm

as i did never analyse this with a NNUE-engine i tried corchess now (nn-56a5f1c4173a.nnue) and offered some ideas that i still had in mind and came to my "new" mainline:


[d]3r3k/p4rb1/1p1p3p/nPpqpNpP/P1P1N1P1/3P1P2/4RK2/1R6 w - - 0 21
cxd5 +-

as i couldnt find any serious mistakes after 4...Qxe6 for black by my own(corechess), this might be a good line to backtrack from if somebody is interested in (1.Nc3!! stays oc the best move)

this took me around 3-4 hours of computeranalysis and now i believe that 1.h5 might be indeed winning as well lol, hf
Wahrheiten sind Illusionen von denen wir aber vergessen haben dass sie welche sind.

Post Reply