Sting Sf 13
Moderators: bob, hgm, Harvey Williamson
Forum rules
This textbox is used to restore diagrams posted with the [d] tag before the upgrade.
This textbox is used to restore diagrams posted with the [d] tag before the upgrade.
Sting Sf 13
Some next very important modifications and corrections: trapped pieces, zero beta, lazy king, fortress.
download:
http://www.mediafire.com/file/jq0gp5b7w ... 3.zip/file
download:
http://www.mediafire.com/file/jq0gp5b7w ... 3.zip/file
Maybe, I can't be friendly, but let me be useful.
Re: Sting Sf 13
The "IQ" test is a very important part of my project (Sting) and doesn't contain middlegame tactical positions and technical endgames.
Here are positions dedicated to test socalled by me quasiintelligent abilities of engines.
They show full spectrum of possible problems for a clessic model of alphabeta engines.
Set parameters: "threads" to 1 and "hash" to 1024 and clear hash before each of items.
Don't use such external tricks like: rule50 changing, nullmove disabling, "backwardforward" moves, special parameters and others.
MultiPV, tablebases and "Monte Carlo" are not welcome too.
Time to solve each of items should be 30 minutes for 1 000 000 nodes/second speed. It means 3 minutes for 10 000 000 nodes/second only. Of course, the going on last depth should not be broken.
The DRAW result means ZERO score or a missing gain.
The WIN result means the correct solution mainly and any gain.
[Click square "d" to get the full pgn]
The authors of studies selected to the "IQ" test:
by YACPDB:
Blathy Otto  2, 11, 12
Benko Pal  15, 74
Hasek Josef  21, 27
Gurvich Abram  22
Kasparyan Genrikh  23
Neghina Mihai  26
Grasemann Herbert  36
Popov Georgy & Hanyan Aleksei  38 (part of study)
Rudolph William  45
Troitzky Aleksei  56
Reti Richard  57
Matous Mario  59
Behting Carl  61
Klyatskin Mikhail  65
Kubbel Karl  68
Novomesky Daniel  72
Lazard Frederic  73
Marwitz Jan  78
Lamford Paul  81
Simkhovich Froim  85
by other sources (Talkchess.com):
Solovyov Y.  5
Babic M.  8
Neghina Mihai  10, 30, 83, 86, 88, 91
Simkhovich Froim  19
Plaskett  35
Hoch Yehuda & Aloni Hillel  54
Troitzky Aleksei  58
Smyslov Vasily  66
Bondarenko F. & Kuznetsov A.  67
Surya Sekhar Ganguly  80
Aleksandr Stavrietsky  87
A. Kazantsev  92
Here are positions dedicated to test socalled by me quasiintelligent abilities of engines.
They show full spectrum of possible problems for a clessic model of alphabeta engines.
Set parameters: "threads" to 1 and "hash" to 1024 and clear hash before each of items.
Don't use such external tricks like: rule50 changing, nullmove disabling, "backwardforward" moves, special parameters and others.
MultiPV, tablebases and "Monte Carlo" are not welcome too.
Time to solve each of items should be 30 minutes for 1 000 000 nodes/second speed. It means 3 minutes for 10 000 000 nodes/second only. Of course, the going on last depth should not be broken.
The DRAW result means ZERO score or a missing gain.
The WIN result means the correct solution mainly and any gain.
[Click square "d" to get the full pgn]
The authors of studies selected to the "IQ" test:
by YACPDB:
Blathy Otto  2, 11, 12
Benko Pal  15, 74
Hasek Josef  21, 27
Gurvich Abram  22
Kasparyan Genrikh  23
Neghina Mihai  26
Grasemann Herbert  36
Popov Georgy & Hanyan Aleksei  38 (part of study)
Rudolph William  45
Troitzky Aleksei  56
Reti Richard  57
Matous Mario  59
Behting Carl  61
Klyatskin Mikhail  65
Kubbel Karl  68
Novomesky Daniel  72
Lazard Frederic  73
Marwitz Jan  78
Lamford Paul  81
Simkhovich Froim  85
by other sources (Talkchess.com):
Solovyov Y.  5
Babic M.  8
Neghina Mihai  10, 30, 83, 86, 88, 91
Simkhovich Froim  19
Plaskett  35
Hoch Yehuda & Aloni Hillel  54
Troitzky Aleksei  58
Smyslov Vasily  66
Bondarenko F. & Kuznetsov A.  67
Surya Sekhar Ganguly  80
Aleksandr Stavrietsky  87
A. Kazantsev  92
Maybe, I can't be friendly, but let me be useful.
Re: Sting Sf 13
Some words more about the new version.
Looking back at the weak versions 10.x and 11.x Sting 13 is a big jump in quality. It is less selective and is based on unified ideas. I had rejected all doubt ideas and added a few new ones.
Sting solves more than 90 (of 94) items of the "IQ" test and it is a very good result.
The "IQ" test is not a random work (as others) and is done by (I hope) a proper person.
Looking back at the weak versions 10.x and 11.x Sting 13 is a big jump in quality. It is less selective and is based on unified ideas. I had rejected all doubt ideas and added a few new ones.
Sting solves more than 90 (of 94) items of the "IQ" test and it is a very good result.
The "IQ" test is not a random work (as others) and is done by (I hope) a proper person.
Maybe, I can't be friendly, but let me be useful.
Re: Sting Sf 13
Thanks Marek, i will test this version on a few testsets next week.
Re: Sting Sf 13
Maybe some words about the future of my project Sting.
There are using some tricks to solve difficult positions. Two (maybe more) of them are very interesting and can be implemented by engines directly, of course, in a specific way without human orders (tricks).
Now Sting (from version 13) is ready ("IQ" test) to take such an intelligent support and to breake the "horizon effect" for alphabeta engines.
It is not an easy work (time) but a posible effect seems to be worth to try it.
There are using some tricks to solve difficult positions. Two (maybe more) of them are very interesting and can be implemented by engines directly, of course, in a specific way without human orders (tricks).
Now Sting (from version 13) is ready ("IQ" test) to take such an intelligent support and to breake the "horizon effect" for alphabeta engines.
It is not an easy work (time) but a posible effect seems to be worth to try it.
Maybe, I can't be friendly, but let me be useful.

 Posts: 195
 Joined: Fri Mar 10, 2006 9:26 am
 Location: France
Re: Sting Sf 13
Too many games lost on time! Pity!
Re: Sting Sf 13
The real point of this project is to create a smart analysis tool, especially for tough endgames and fortresstype positions. Using it for engine matches is a waste of resources, imho.

 Posts: 195
 Joined: Fri Mar 10, 2006 9:26 am
 Location: France
Re: Sting Sf 13
Thanks for the precision.
Re: Sting Sf 13
Many thanks for this information. I will take a look at this question too.
It is important for me to prove that Sting with my changes is not much weaker than Stockfish 2.1.1 .
BTW. I can't understand why in the ELO tests only the time menagment is used.
"Time per move" can return the real strength of engines.
Maybe, I can't be friendly, but let me be useful.
Re: Sting Sf 13
Hi Marek, it seems like with Sting 13 this very difficult problem is solved.
This is great !
I have just a crash engine after 10 hours but I used 8Go Hash.
Analysis by Sting 13 based on Stockfish 2.1.1 64bit:
.....
1.b4xa5 Rd3e3+ 2.Ke5d6 Re3d3+ 3.Kd6e5
= (0.00) Depth: 35 00:06:22 3309MN
1.b4xa5 Rd3e3+ 2.Ke5d6 Re3d3+ 3.Kd6e5
= (0.00) Depth: 36 00:08:24 4368MN
1.b4b5
= (0.08 ++) Depth: 37 00:10:13 5317MN
1.b4b5
= (0.16 ++) Depth: 37 00:12:05 6305MN
1.b4b5 Rd3e3+ 2.Ke5d6 Re3d3+ 3.Kd6c7 Rd3f3 4.e6e7 Rf3e3 5.Kc7d7 Re3d3+ 6.Kd7e6 Rd3e3+ 7.Ke6f7 Re3f3+ 8.Kf7e8 Rf3e3 9.Rg2g1 g5g4 10.Rg1g2 Kc4b4 11.Rg2c2 Re3e4 12.Rc2d2 Kb4c3 13.Rd2g2 Kc3b4 14.Rg2xg3 Kb4xb5 15.Ke8f7 Re4f4+ 16.Kf7e6 Rf4e4+ 17.Ke6d6 Re4d4+ 18.Kd6e5 Rd4e4+ 19.Ke5xe4 Kb5b4 20.e7e8Q Kb4c4 21.Qe8xa4+ Kc4c5 22.Rg3xa3 Kc5d6 23.Qa4xa5
= (0.28 ++) Depth: 37 00:12:52 6723MN
1.b4b5 Rd3e3+ 2.Ke5d6 Re3d3+ 3.Kd6c7 Rd3f3 4.e6e7 Rf3e3 5.Kc7d7 Re3d3+ 6.Kd7e6 Rd3e3+ 7.Ke6f7 Re3f3+ 8.Kf7e8 Rf3e3 9.Rg2g1 g5g4 10.Rg1g2 Kc4b4 11.Rg2c2 Kb4xb5 12.Rc2d2 Kb5c4 13.Ke8d8 Kc4c3 14.Rd2g2 Kc3d3 15.Rg2xg3 Re3xg3 16.e7e8Q Kd3d2 17.Qe8xa4 Rg3d3+ 18.Kd8e7 g4g3 19.Qa4xa5+ Kd2e2 20.Qa5e5+ Rd3e3
+/= (0.32) Depth: 37 00:14:23 7544MN
1.b4b5 Rd3e3+
= (0.20 ) Depth: 38 00:25:10 13255MN
1.b4b5 Rd3e3+ 2.Ke5d6 Re3d3+ 3.Kd6c7 Rd3e3 4.Kc7d7 Re3d3+ 5.Kd7e8 Rd3e3 6.e6e7 Re3f3 7.Ke8d7 Rf3d3+
+/= (0.32) Depth: 38 00:29:30 15536MN
1.b4b5 Rd3e3+ 2.Ke5d6 Re3d3+ 3.Kd6c7 Rd3e3 4.Kc7d7 Re3d3+ 5.Kd7e8 Rd3e3 6.e6e7 Re3f3 7.Ke8d7 Rf3d3+
+/= (0.32) Depth: 39 00:35:18 18662MN
1.b4b5 Rd3e3+ 2.Ke5d6 Re3d3+ 3.Kd6c7 Rd3e3 4.Kc7d7 Re3d3+ 5.Kd7e8 Rd3e3 6.e6e7 Re3f3 7.Ke8d7 Rf3d3+
+/= (0.32) Depth: 40 00:44:38 23622MN
1.b4b5 Rd3e3+ 2.Ke5d6 Re3d3+ 3.Kd6c7 Rd3e3 4.Kc7d7 Re3d3+ 5.Kd7e8 Rd3e3 6.e6e7 Re3f3 7.Ke8d7 Rf3d3+
+/= (0.32) Depth: 41 01:33:22 49130MN
1.b4b5 Rd3e3+ 2.Ke5d6 Re3d3+ 3.Kd6c7 Rd3e3 4.Kc7d7 Re3d3+ 5.Kd7e8 Rd3e3 6.e6e7 Re3f3 7.Ke8d7
+/= (0.32) Depth: 42 02:14:07 69683MN
1.b4b5
+/= (0.40 ++) Depth: 43 02:17:22 71411MN
1.b4b5
+/= (0.48 ++) Depth: 43 02:23:02 74417MN
1.b4b5 Rd3e3+ 2.Ke5d6 Re3d3+ 3.Kd6c7 Rd3e3 4.Kc7d7 Re3d3+ 5.Kd7e8 g5g4 6.e6e7 Rd3e3 7.Rg2g1 Re3e2 8.Rg1xg3 Re2xa2 9.Rg3xg4+ Kc4xb5 10.Rg4e4 Kb5b6 11.Ke8f7 Ra2f2+ 12.Kf7g8 Rf2g2+ 13.Kg8xh7 a3a2 14.e7e8Q a2a1Q 15.Re4e6+ Kb6b7 16.Qe8c6+ Kb7b8 17.Qc6xg2 Qa1a2 18.Qg2xa2 Kb8b7 19.Re6xh6 Kb7b8 20.Qa2xa4 Kb8b7 21.Qa4xa5 Kb7b8 22.Kh7g8 Kb8b7 23.Rh6e6 Kb7b8 24.Kg8f7 Kb8b7 25.Re6e1 Kb7b8 26.Kf7g8 Kb8b7 27.Kg8g7
+/= (0.60 ++) Depth: 43 02:31:23 78831MN
1.b4b5 Rd3e3+ 2.Ke5d6 Re3d3+ 3.Kd6c7 Rd3e3 4.Kc7d7 Re3d3+ 5.Kd7e8 g5g4 6.e6e7 Rd3e3 7.Rg2g1 Re3e2 8.Rg1xg3 Re2xa2 9.Rg3xg4+ Kc4xb5 10.Rg4e4 Kb5b6 11.Ke8f7 Ra2f2+ 12.Kf7g8 Rf2g2+ 13.Kg8xh7 a3a2 14.e7e8Q a2a1Q 15.Re4e6+ Kb6b7 16.Qe8c6+ Kb7b8 17.Qc6xg2 Qa1a2 18.Qg2xa2 Kb8b7 19.Re6xh6 Kb7b8 20.Qa2xa4 Kb8b7 21.Qa4xa5 Kb7b8 22.Kh7g8 Kb8b7 23.Rh6e6 Kb7b8 24.Kg8f7 Kb8b7 25.Re6e1 Kb7b8 26.Kf7g8
+/ (0.78 ++) Depth: 43 02:35:10 80813MN
1.b4b5 Rd3e3+ 2.Ke5d6 Re3d3+ 3.Kd6c7 Rd3e3 4.Kc7d7 Re3d3+ 5.Kd7e8 g5g4 6.e6e7 Rd3e3 7.Rg2g1 Re3e2 8.Rg1xg3 Re2xa2 9.Rg3xg4+ Kc4xb5 10.Rg4e4 Kb5b6 11.Ke8f7 Ra2f2+ 12.Kf7g8 Rf2g2+ 13.Kg8xh7 a3a2 14.e7e8Q a2a1Q 15.Re4e6+ Kb6b7 16.Qe8c6+ Kb7b8 17.Qc6xg2 Qa1a2 18.Qg2xa2 Kb8b7 19.Re6xh6 Kb7b8 20.Qa2xa4 Kb8b7 21.Qa4xa5 Kb7b8 22.Kh7g8 Kb8b7 23.Rh6e6 Kb7b8 24.Kg8f7 Kb8b7 25.Re6e1 Kb7b8 26.Kf7g8
+/ (1.06 ++) Depth: 43 03:43:39 116707MN
1.b4b5 Rd3e3+ 2.Ke5d6 Re3d3+ 3.Kd6c7 Rd3e3 4.Kc7d7 Re3d3+ 5.Kd7e8 g5g4 6.e6e7 Rd3e3 7.Rg2g1 Re3e2 8.Rg1xg3 Re2xa2 9.Rg3xg4+ Kc4xb5 10.Rg4e4 Kb5b6 11.Ke8f7 Ra2f2+ 12.Kf7g8 Rf2g2+ 13.Kg8xh7 a3a2 14.e7e8Q a2a1Q 15.Re4e6+ Kb6b7 16.Qe8c6+ Kb7b8 17.Qc6xg2 Qa1a2 18.Qg2xa2 Kb8b7 19.Re6xh6 Kb7b8 20.Qa2xa4 Kb8b7 21.Qa4xa5 Kb7b8 22.Kh7g8 Kb8b7 23.Rh6e6 Kb7b8 24.Kg8f7 Kb8b7 25.Re6e1 Kb7b8 26.Kf7g8
+/ (1.46 ++) Depth: 43 04:08:08 129372MN
1.b4b5 Rd3e3+ 2.Ke5d6 Re3d3+ 3.Kd6c7 Rd3e3 4.Kc7d7 Re3d3+ 5.Kd7e8 g5g4 6.e6e7 Rd3e3 7.Rg2g1 g3g2 8.Rg1xg2 g4g3 9.Rg2g1 Re3d3 10.Ke8f7 Rd3f3+ 11.Kf7g7
+/ (1.29) Depth: 43 06:25:20 201624MN
1.b4b5
+/ (1.41 ++) Depth: 44 06:55:47 217861MN
(, 05.02.2019)