sting-sf-7 release

[lech]

download: http://www.mediafire.com/download/223t9 ... g-sf-7.zip

[lech]

What is new?
I modified one my idea and now Sting is able to solve such positions too:
[d]1b6/2p5/1kB5/1PN5/8/3K4/8/q7 w - - bm Nb3;
Instead of 1.Nd7+ ?? Ka5 !! 2.Nxb8
I added a new interesting idea and now Sting can reject the move 1.Nxb6?? (no static eval trick!)
[d]2k5/2p5/1q1p4/pPpPp1pp/N1P1Pp2/P4PbP/KQ4P1/8 w - - am Nxb6+;
and it helps to find a better move in famous position
[d]1r6/4k3/r2p2p1/2pR1p1p/2P1pP1P/pPK1P1P1/P7/1B6 b - - bm Rxb3+;
I added a code for stalemate positions and endgame positions like QK/PK too.

It is probably the last version based on old Stockfish 2.1.1. I will try to move my ideas to the newer 3.0 version.

The project Sting can to prove that chess engines (only?) don't need an artifical intelligence. It is possible without some tricks tablebases, null move disabling, rule_50, Monte Carlo and others.
Some there Elo questions are important but not only one.

[carldaman]

It is probably the last version based on old Stockfish 2.1.1. I will try to move my ideas to the newer 3.0 version.

The project Sting can to prove that chess engines (only?) don't need an artifical intelligence. It is possible without some tricks tablebases, null move disabling, rule_50, Monte Carlo and others.
Some there Elo questions are important but not only one.

Thanks for releasing Sting to us, Marek.

If I may make a recommendation, I would suggest using Stockfish DD (named after the late Don Dailey) as a coding base, rather than the weaker and stylistically unimpressive SF 3.

Regards,
CL

[tpoppins]

Reporting a bug under Chessbase GUIs (specifically tested under DF 14.38).

The GUI seems to be confused by the "Maximum Number of Threads per Split Point" param. Apparently it interpets the param's value as the max number of cores the engine can use and ignores the "Threads" value. As the former param maxes out at 8, the GUI won't allow the engine to use more than 8 cores.

The same issue affects Sting SF 6.

There's a workaround: insert the following line in the [OPTIONS] section of your Sting SF .uci file*:

Code: Select all

Maximum Number of Threads per Split Point=xx

where xx is the max number of cores on your system. Sting SF then appears to use the number of cores specified. I don't know what value it uses for the "Maximum Number of Threads per Split Point" param in that case, though; probably the default (5).


* this file is typically stored in C:\Users\<your user name>\AppData\Roaming\ChessBase\Engines.UCI

[Vinvin]

download: http://www.mediafire.com/download/223t9 ... g-sf-7.zip

Thanks, Marek !

[leslies]

thank you very much.

[fishpov]

Hi Marek,

For this position
[D]2k5/2p5/1q1p4/pPpPp1pp/N1P1Pp2/P4PbP/KQ4P1/8 w - - 0 1

Sting SF7.2 is actually able to find another move instead of Nxb6 ?? but with no particular plan.
After several hours Nxb6 ? is considered again.

Analysis by Sting SF 7.2 64bit:

.....
1.Nxb6+ cxb6 2.h4 gxh4 3.Kb3 Kd7 4.Kc3 Ke7 5.Kd3 Kf6 6.Qd2 Ke7 7.Qc2 Kf6 8.Qd1 h3 9.gxh3 h4 10.Qd2 Ke7 11.Qe2 Kf6 12.Qc2 Kf7 13.Qg2 Kf6 14.Qe2 Ke7 15.Qc2 Kf7
+- (8.84) Depth: 37 00:00:05 17400kN
1.Nxb6+ cxb6 2.h4 gxh4 3.Kb3 Kd7 4.Kc3 Ke7 5.Kd3 Kf6 6.Qd2 Ke7 7.Qc2 Kf6 8.Qd1 h3 9.gxh3 h4 10.Qd2 Ke7 11.Qe2 Kf6 12.Qc2 Kf7 13.Qg2 Kf6 14.Qe2 Ke7 15.Qb2 Kf6 16.Qc1 Kf7 17.Qc2 Ke7 18.Qd2 Kd7 19.Qc2 Ke7
+- (8.84) Depth: 38 00:00:07 25965kN
1.Nxb6+ cxb6 2.h4 gxh4 3.Kb3 Kd7 4.Kc3 Ke7 5.Kd3 Kf6 6.Qd2 Ke7 7.Qc2 Kf6 8.Qd1 h3 9.gxh3 h4 10.Qd2 Ke7 11.Qe2 Kf6 12.Qc2 Kf7 13.Qg2 Kf6 14.Qe2 Ke7 15.Qb2 Kd7 16.Qg2 Ke7 17.Qd2 Kd7 18.Qc2 Kd8 19.Qe2 Ke7
+- (8.84) Depth: 39 00:00:10 44481kN
1.Nxb6+ cxb6 2.h4 gxh4 3.Kb3 Kd7 4.Kc3 Ke7 5.Kd3 Kf6 6.Qd2 Ke7 7.Qc2 Kf6 8.Qd1 h3 9.gxh3 h4 10.Qd2 Ke7 11.Kc3 Kf6 12.Kd3
+- (8.84) Depth: 40 00:00:11 49862kN
1.Kb3 Qb8 2.Qd2 Qa7 3.Nb2 Kb7 4.Ka4 Kb6 5.Nd3 Qa8 6.Nc1
+- (8.85) Depth: 41 00:00:16 75351kN
1.Kb3 Qb8 2.Qd2 Qa7 3.Nb2 Kb7 4.Ka4 Kb6 5.Nd3 Qa8 6.Nc1 Qa7 7.Nb3 Kb7 8.Nxa5+ Kb6 9.Qc3 Qa8
+- (8.85) Depth: 42 00:00:23 109MN
1.Kb3 Qb8 2.Qd2 Qa7 3.Nc3 Qb7 4.Ne2 Bf2 5.Nxf4 gxf4 6.Qxf2 Kd7 7.Qd2 Qb6 8.Kc2
+- (8.85) Depth: 43 00:00:36 172MN
1.Kb3 Qb8 2.Qd2 Qa7 3.Nc3 Kb7 4.Ne2 Bh4 5.Kc2 Kb6
+- (8.85) Depth: 44 00:01:00 289MN
1.Kb3 Qb8 2.Qd2 Qa8 3.Nb2 Kb7 4.Ka4 Kb6 5.Nd3 Bh2 6.Nc1 Qa7 7.Ne2 h4 8.Qe1 Bg3 9.Nxg3 hxg3 10.Qc3
+- (8.85) Depth: 45 00:56:07 17133MN
1.Kb3 Qb8 2.Qd2
+- (8.85) Depth: 46 05:02:02 91872MN
1.Nxb6+
+- (8.92 ++) Depth: 47 05:12:59 95180MN
1.Nxb6+ cxb6 2.Kb3 Kd7 3.Kc2 Ke7 4.Qc1 h4 5.Qd2 Kd7 6.Kd3 Ke8 7.Qe2 Ke7 8.Qf1 Kd7 9.Kc3 Ke7 10.Qe2 Kf6 11.Kb3 Ke7 12.Qf1 Kd7 13.Kc3
+- (8.88) Depth: 47 07:29:13 137239MN
1.Nxb6+
+- (8.96 ++) Depth: 48 07:29:16 137255MN

Thanks for great Sting improvement !
Nicolas.