Sting SF 6.2

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

Moderators: hgm, Harvey Williamson, bob

Forum rules
This textbox is used to restore diagrams posted with the [d] tag before the upgrade.
Post Reply
lech
Posts: 921
Joined: Sun Feb 14, 2010 9:02 pm

Sting SF 6.2

Post by lech » Tue Feb 23, 2016 10:17 am

I found a next, I hope now exactly last, problem leading to the horizon effect too.
It happens in such a position and many others.
[d]6k1/6p1/4Rb1p/p7/rp2P1N1/2P3P1/1P5P/6K1 b - - bm Bxc3;
I solved this problem and, additionally, I added to the code of Sting an important endgame knowladge KQ/KP.

download: http://www.mediafire.com/download/k41dl ... -sf-62.zip

Now Sting should solve all difficult positions. What it mens? Sting contains many good tools to help to solve all the tough for engines positions. But for many, It is still a question of CPU(s), hash size and good or better static evaluation and search, probably the main problem.

The list of tough positions: Some time ago it was impossible to solve all the positions without tricks by only one strong engine. Only ELO racing is able to.get it never.

7k/6rp/5BpN/1r4P1/8/8/8/K7 b - - bm Rf5 (next Bd4[Bc3] Rxg5);
8/5pKN/5Pp1/5kP1/8/Bb6/8/7q b - - bm Qxh7+;
3N2r1/2K1p3/4Pk2/8/Bp5b/8/2P5/8 b - - bm Bg3+ (next Kd7 Rxd8+)
6k1/6p1/4Rb1p/p7/rp2P1N1/2P3P1/1P5P/6K1 b - - bm Bxc3;
8/8/p2k1p2/1p1p3p/1P1P3p/P3NPP1/5K2/1b6 w - - bm Ng2;
1n6/Pp1p1p1p/1P1P1P1P/4K3/2p5/p7/rpp1P3/qkb1R3 w - - bm axb8N;
1N4K1/1pp1p3/7k/2r3pP/p2pp1P1/P6N/8/2B5 w - - bm Nc6;
5K2/k4p1p/5p1p/1p3p2/bP6/1p6/1P5p/7B w - - bm Ke8(e7);
k1b5/1p1p1p1p/1PpPpP2/2B5/8/2p1p2P/2PbP3/3K4 w - - Bd4;
2b1r3/r2ppN2/8/1p1p1k2/pP1P4/2P3R1/PP3PP1/2K5 w - - bm Nd6+;
7r/p3k3/2p5/1pPp4/3P4/PP4P1/3P1PB1/2K5 w - - bm Kd1;
1K3kB1/8/6p1/7R/8/4b3/8/6n1 w - - bm Rh8 (next Kg7 Bh7);
8/8/R7/1b4k1/5p2/1B3r2/7P/7K w - - bm h4+;
1k1n4/1p1b1p2/1Bp1pPp1/1PP1P3/8/6P1/8/1K6 w - - bm Bc7+ (win);
4q1kr/p6p/1prQPppB/4n3/4P3/2P5/PP2B2P/R5K1 w - - bm Qxe5;
q1b4n/1p1pKp2/1p1p1Pp1/1P1P2Pp/2P4P/8/7R/1N1k4 w - - bm Ra2;
3k4/2pq3p/pp5R/4P3/P6P/2PN4/1PN3K1/8 w - - bm Nd4 (win);
8/7p/6pP/p4pP1/3BpP2/p1KpP3/pn1N4/k7 w - - bm Bh8;
8/6p1/p7/rp1K2p1/kb3pP1/2p2p1b/P1Np1P2/3N2R1 w - - bm Ke6(e4);
1qN5/b1pB4/1pP2k2/pP3P2/P1P5/2K5/8/8 w - - (Kc3->d8);
8/5K2/3p3p/3p3P/pp1P4/rkb1R3/p1p3P1/n1B2B2 w - - bm Rd3;
r7/1b1r4/k1p1p1p1/1p1pPpPp/p1PP1P1P/PP1K4/8/4Q3 w - - bm Qa5+;
3B4/1r2p3/r2p1p2/bkp1P1p1/1p1P1PPp/p1P4P/PPB1K3/8 w - - bm Ba4+;
Maybe, I can't be friendly, but let me be useful.
Top
Vinvin
Posts: 4024
Joined: Thu Mar 09, 2006 8:40 am

Re: Sting SF 6.2

Post by Vinvin » Tue Feb 23, 2016 10:44 am

Thanks, Marek !
I'll test it soon !
Top
zullil
Posts: 4878
Joined: Mon Jan 08, 2007 11:31 pm
Location: PA USA

Re: Sting SF 6.2

Post by zullil » Tue Feb 23, 2016 11:06 am

lech wrote:I found a next, I hope now exactly last, problem leading to the horizon effect too.
It happens in such a position and many others.
[d]6k1/6p1/4Rb1p/p7/rp2P1N1/2P3P1/1P5P/6K1 b - - bm Bxc3;
The latest Stockfish solves this quickly---albeit with 20 threads and endgame tables. :wink:

Code: Select all

info depth 29 seldepth 45 multipv 1 score cp 38 lowerbound nodes 873917808 nps 37131110 hashfull 843 tbhits 248311 time 23536 pv f6c3
info depth 29 currmove f6c3 currmovenumber 1
info depth 29 seldepth 45 multipv 1 score cp 45 lowerbound nodes 902920436 nps 37243047 hashfull 858 tbhits 261578 time 24244 pv f6c3
info depth 29 currmove f6c3 currmovenumber 1
info depth 29 seldepth 45 multipv 1 score cp 56 lowerbound nodes 931298007 nps 37395519 hashfull 868 tbhits 270923 time 24904 pv f6c3
info depth 29 currmove f6c3 currmovenumber 1
info depth 29 seldepth 45 multipv 1 score cp 70 lowerbound nodes 970648778 nps 37592903 hashfull 880 tbhits 280461 time 25820 pv f6c3
info depth 29 currmove f6c3 currmovenumber 1
info depth 29 seldepth 45 multipv 1 score cp 91 lowerbound nodes 1006390942 nps 37852745 hashfull 884 tbhits 286595 time 26587 pv f6c3
info depth 29 currmove f6c3 currmovenumber 1
info depth 29 seldepth 45 multipv 1 score cp 118 lowerbound nodes 1053699656 nps 38134691 hashfull 896 tbhits 292727 time 27631 pv f6c3
info depth 29 currmove f6c3 currmovenumber 1
info depth 29 seldepth 45 multipv 1 score cp 153 lowerbound nodes 1142158136 nps 38663489 hashfull 915 tbhits 306343 time 29541 pv f6c3
info depth 29 currmove f6c3 currmovenumber 1
info depth 29 seldepth 45 multipv 1 score cp 200 lowerbound nodes 1285965407 nps 39454053 hashfull 939 tbhits 332006 time 32594 pv f6c3
info depth 29 currmove f6c3 currmovenumber 1
info depth 29 seldepth 45 multipv 1 score cp 260 lowerbound nodes 1459691897 nps 40375401 hashfull 956 tbhits 373003 time 36153 pv f6c3
info depth 29 currmove f6c3 currmovenumber 1
info depth 29 seldepth 53 multipv 1 score cp 337 lowerbound nodes 1868787418 nps 41321085 hashfull 986 tbhits 457894 time 45226 pv f6c3
info depth 29 currmove f6c3 currmovenumber 1
info depth 29 currmove b4c3 currmovenumber 2
info depth 29 currmove g8f7 currmovenumber 3
info depth 29 currmove g8h7 currmovenumber 4
info depth 29 currmove b4b3 currmovenumber 5
info depth 29 currmove a4a1 currmovenumber 6
info depth 29 currmove a4a2 currmovenumber 7
info depth 29 currmove f6e7 currmovenumber 8
info depth 29 currmove a4a3 currmovenumber 9
info depth 29 currmove g8f8 currmovenumber 10
info depth 29 currmove f6e5 currmovenumber 11
info depth 29 currmove f6d4 currmovenumber 12
info depth 29 currmove h6h5 currmovenumber 13
info depth 29 currmove g8h8 currmovenumber 14
info depth 29 currmove g7g5 currmovenumber 15
info depth 29 currmove g7g6 currmovenumber 16
info depth 29 currmove f6d8 currmovenumber 17
info depth 29 currmove f6g5 currmovenumber 18
info depth 29 currmove f6h4 currmovenumber 19
info depth 29 seldepth 55 multipv 1 score cp 396 nodes 2585465071 nps 42701786 hashfull 996 tbhits 652295 time 60547 pv f6c3 g1f1 c3b2 e6b6 b2d4 b6b8 g8f7 f1e2 a4a3 b8d8 d4c3 d8b8 a3a2 e2f3 a5a4 e4e5 h6h5 g4e3 a4a3 b8b6 a2d2 e5e6 f7f6 e3f5 d2d5 e6e7 f6f7 f5e3 d5d3 b6a6 f7e7 f3e4 d3d4 e4f3 d4d2 e3c4 d2h2 c4a3 b4a3 a6a3 c3e5 a3a7 e7d6 f3e4 h2e2 e4f5
Actually, much faster without tables and with a smaller hash:

Code: Select all

info depth 27 seldepth 45 multipv 1 score cp 51 lowerbound nodes 633105489 nps 46517670 hashfull 997 tbhits 0 time 13610 pv f6c3
info depth 27 currmove f6c3 currmovenumber 1
info depth 27 seldepth 45 multipv 1 score cp 58 lowerbound nodes 649491923 nps 46622060 hashfull 997 tbhits 0 time 13931 pv f6c3
info depth 27 currmove f6c3 currmovenumber 1
info depth 27 seldepth 45 multipv 1 score cp 69 lowerbound nodes 668880049 nps 46804285 hashfull 997 tbhits 0 time 14291 pv f6c3
info depth 27 currmove f6c3 currmovenumber 1
info depth 27 seldepth 45 multipv 1 score cp 84 lowerbound nodes 696685186 nps 46946441 hashfull 998 tbhits 0 time 14840 pv f6c3
info depth 27 currmove f6c3 currmovenumber 1
info depth 27 seldepth 45 multipv 1 score cp 104 lowerbound nodes 726591893 nps 47059060 hashfull 998 tbhits 0 time 15440 pv f6c3
info depth 27 currmove f6c3 currmovenumber 1
info depth 27 seldepth 45 multipv 1 score cp 131 lowerbound nodes 767175331 nps 47222413 hashfull 998 tbhits 0 time 16246 pv f6c3
info depth 27 currmove f6c3 currmovenumber 1
info depth 27 seldepth 45 multipv 1 score cp 167 lowerbound nodes 856728058 nps 47477309 hashfull 998 tbhits 0 time 18045 pv f6c3
info depth 27 currmove f6c3 currmovenumber 1
info depth 27 seldepth 45 multipv 1 score cp 213 lowerbound nodes 961225933 nps 47972547 hashfull 999 tbhits 0 time 20037 pv f6c3
info depth 27 currmove f6c3 currmovenumber 1
info depth 27 seldepth 45 multipv 1 score cp 273 lowerbound nodes 1159459675 nps 48651379 hashfull 999 tbhits 0 time 23832 pv f6c3
info depth 27 currmove f6c3 currmovenumber 1
info depth 27 currmove b4c3 currmovenumber 2
info depth 27 currmove g8f7 currmovenumber 3
info depth 27 currmove a4a1 currmovenumber 4
info depth 27 currmove g8h7 currmovenumber 5
info depth 27 currmove b4b3 currmovenumber 6
info depth 27 currmove a4a3 currmovenumber 7
info depth 27 currmove f6d4 currmovenumber 8
info depth 27 currmove g8f8 currmovenumber 9
info depth 27 currmove a4a2 currmovenumber 10
info depth 27 currmove f6e5 currmovenumber 11
info depth 27 currmove g8h8 currmovenumber 12
info depth 27 currmove f6g5 currmovenumber 13
info depth 27 currmove f6e7 currmovenumber 14
info depth 27 currmove f6h4 currmovenumber 15
info depth 27 currmove g7g6 currmovenumber 16
info depth 27 currmove f6d8 currmovenumber 17
info depth 27 currmove h6h5 currmovenumber 18
info depth 27 currmove g7g5 currmovenumber 19
info depth 27 seldepth 45 multipv 1 score cp 332 nodes 1411737192 nps 49421921 hashfull 999 tbhits 0 time 28565 pv f6c3 g1f1 c3b2 e6b6 b2d4 b6b8 g8f7 f1e2 a4a3 b8d8 d4c3 d8d1 b4b3 e2d3 b3b2 d3c2 a3a1 d1b1 c3d4 b1b2 d4b2 c2b2 a1e1 g4e5 f7f6 e5d7 f6g5 d7c5 e1e2 b2c3 e2h2 c3d3 h2g2 e4e5 g2g3 d3d4 g3g4 d4d5
Top
lech
Posts: 921
Joined: Sun Feb 14, 2010 9:02 pm

Re: Sting SF 6.2

Post by lech » Tue Feb 23, 2016 2:15 pm

zullil wrote:
lech wrote:I found a next, I hope now exactly last, problem leading to the horizon effect too.
It happens in such a position and many others.
[d]6k1/6p1/4Rb1p/p7/rp2P1N1/2P3P1/1P5P/6K1 b - - bm Bxc3;
The latest Stockfish solves this quickly---albeit with 20 threads and endgame tables. :wink:
I know that much more engines is able to solve this position easly, especially if they use ext = ONE_PLY for check moves.
I removed in Sting such an extension to avoid many problems in other tough positions.
For this reason, for Sting it was a very difficult position.
My new tool should work good in other similar (by problem) positions too.
Maybe, I can't be friendly, but let me be useful.
Top
Damir
Posts: 1796
Joined: Mon Feb 11, 2008 2:53 pm

Re: Sting SF 6.2

Post by Damir » Tue Feb 23, 2016 2:24 pm

Hi Marek :)

Can you make Sting stronger than the current Stockfish ? :wink:
Top
lech
Posts: 921
Joined: Sun Feb 14, 2010 9:02 pm

Re: Sting SF 6.2

Post by lech » Tue Feb 23, 2016 2:47 pm

Damir wrote:Hi Marek :)

Can you make Sting stronger than the current Stockfish ? :wink:
Hi Damir

I think no one is able to do it. :lol:

Seriously, Sting is based on Stockfish 2.1.1 and it is not good idea.
Maybe, I can't be friendly, but let me be useful.
Top
User avatar
Graham Banks
Posts: 30733
Joined: Sun Feb 26, 2006 9:52 am
Location: Auckland, NZ

Re: Sting SF 6.2

Post by Graham Banks » Tue Feb 23, 2016 5:43 pm

Thanks Marek. :)
My email addresses:
gbanksnz at gmail.com
gbanksnz at yahoo.co.nz
Top
lech
Posts: 921
Joined: Sun Feb 14, 2010 9:02 pm

Re: Sting SF 6.2

Post by lech » Wed Feb 24, 2016 10:27 am

Sorry, but I had to close link to the version of Sting SF 6.2 .
In my code I found two serious bug (both in added knowladge KO/KP) .
I will correct it soon (today) or never. :oops: :oops: :oops:
Maybe, I can't be friendly, but let me be useful.
Top
Paloma
Posts: 645
Joined: Thu Dec 25, 2008 8:07 pm

Re: Sting SF 6.2

Post by Paloma » Wed Feb 24, 2016 6:51 pm

Hi Marek,

i prefer soon today :wink:

btw, thanks for your work and good luck.

Have a nice evening.
Top
Post Reply

Return to “Computer Chess Club: General Topics”