about Sting-sf-3

[lech]

Why the topic was written?
Sting (bred on Stockfish 2.1.1) is a deep idea, and not a try to exist here.
As you know, Sting SF 2 can find such difficult sacrifices like this:
[d]8/5pKN/5Pp1/5kP1/8/Bb6/8/7q b - - bm Qxh7+;
without any tricks (one brain in one head) and can play on a high level (very important for me too! ).
However, a real challange for me are zugzwangs, especially the delayed by pawn moves.
Now, I found some intersting new idea and Sting is able to solve the two very difficult positions.
It is a motivation to work and to make Sting-sf-3 published.

Sting-sf-3 beta, 2 threads, 1056M hash.
[d]n1QBq1k1/5p1p/5KP1/p7/8/8/8/8 w - - bm Bc7; (modified Matous study)

Code: Select all

info depth 32 seldepth 28 multipv 1 score cp 0 nodes 1255152615 nps 1048759 time
 1196797 pv g6h7 g8h8 c8a8 a4a3 f6f5 e8e6 f5f4 e6d6 f4f3 d6d3 f3f2 d3d2 f2f3 d2d
3
info depth 33
info currmove g6h7 currmovenumber 1
info nodes 1741482087 nps 1052384 time 1654797
info currmove c8a8 currmovenumber 2
info nodes 1743199448 nps 1052259 time 1656625
info currmove g6f7 currmovenumber 3
info nodes 1873604326 nps 1049702 time 1784891
info currmove f6g5 currmovenumber 4
info currmove f6f5 currmovenumber 5
info nodes 1877050561 nps 1049602 time 1788344
info currmove g6g7 currmovenumber 6
info currmove c8f5 currmovenumber 7
info currmove d8a5 currmovenumber 8
info currmove c8b7 currmovenumber 9
info currmove c8c2 currmovenumber 10
info currmove c8a6 currmovenumber 11
info currmove d8e7 currmovenumber 12
info nodes 1878197481 nps 1049501 time 1789609
info currmove c8c4 currmovenumber 13
info currmove c8g4 currmovenumber 14
info currmove c8c6 currmovenumber 15
info currmove c8b8 currmovenumber 16
info nodes 1879353303 nps 1049414 time 1790859
info currmove c8d7 currmovenumber 17
info currmove c8c7 currmovenumber 18
info currmove d8c7 currmovenumber 19
info depth 33 seldepth 0 multipv 1 score cp 8 lowerbound nodes 1879638514 nps 10
49372 time 1791203 pv d8c7 e8c8 g6f7 g8h8 c7e5 c8c5 e5a1 a8c7 a1b2 a4a3 b2a1 a3a
2 a1b2 a2a1n b2a1
info nodes 1879638514 nps 1049372 time 1791203
info currmove d8c7 currmovenumber 1
info depth 33 seldepth 0 multipv 1 score cp 16 lowerbound nodes 1879744928 nps 1
049358 time 1791328 pv d8c7 e8c8 g6f7 g8h8 c7e5 c8c5 e5a1 a8c7 a1b2 a4a3 b2a1 a3
a2 a1b2 a2a1n b2a1

[d]4k1rr/1R1R3p/8/8/8/2K5/7P/8 w k - bm Rd2; (Benko study)

Code: Select all

info depth 40 seldepth 53 multipv 1 score cp 210 nodes 1554806967 nps 1037932 ti
me 1497984 pv d7e7 e8d8 e7f7 d8c8 b7a7 c8b8 f7b7 b8c8 b7c7 c8b8 a7b7 b8a8 c7h7 h
8h7 b7h7 a8b8 h2h4 b8c8 h4h5 c8d8 h5h6 g8g3 c3d4 g3h3 d4e5 h3h2 e5f6 h2f2 f6e6 f
2e2 e6d5 e2h2 d5d6 h2d2 d6e6 d2e2
info depth 41
info currmove d7e7 currmovenumber 1
info nodes 1763254288 nps 1034735 time 1704062
info currmove d7h7 currmovenumber 2
info nodes 1763754724 nps 1034735 time 1704546
info currmove c3d4 currmovenumber 3
info nodes 1830445318 nps 1032846 time 1772234
info currmove b7b8 currmovenumber 4
info nodes 1830730141 nps 1032833 time 1772531
info currmove b7a7 currmovenumber 5
info nodes 1879337730 nps 1031646 time 1821687
info currmove d7d1 currmovenumber 6
info nodes 1980701278 nps 1031187 time 1920796
info currmove d7d4 currmovenumber 7
info nodes 2028736599 nps 1031148 time 1967453
info currmove b7b5 currmovenumber 8
info currmove d7d3 currmovenumber 9
info nodes 2178672310 nps 1028813 time 2117656
info currmove d7d5 currmovenumber 10
info nodes 2213463958 nps 1028979 time 2151125
info currmove c3d2 currmovenumber 11
info currmove b7b1 currmovenumber 12
info nodes 2214202748 nps 1028942 time 2151921
info currmove d7d2 currmovenumber 13
info depth 41 seldepth 0 multipv 1 score cp 218 lowerbound nodes 2266939963 nps
1028032 time 2205125 pv d7d2 g8f8 c3c2 h8g8 c2b2 h7h5 h2h4 g8h8 b2a2 h8g8 b7a7 f
8f3 d2e2 e8f8 e2c2
info nodes 2266939963 nps 1028032 time 2205125
info currmove d7d2 currmovenumber 1
info depth 41 seldepth 0 multipv 1 score cp 226 lowerbound nodes 2279974230 nps
1027717 time 2218484 pv d7d2 g8f8 c3c2 h8g8 c2b2 h7h5 h2h4 g8h8 b2a2 h8g8 b7a7 f
8f3 d2e2 e8f8 e2c2 f3e3 c2f2 f8e8 a7a8
info nodes 2279974230 nps 1027717 time 2218484
info currmove d7d2 currmovenumber 1
info depth 41 seldepth 0 multipv 1 score cp 238 lowerbound nodes 2284168067 nps
1027638 time 2222734 pv d7d2 g8f8 c3c2 h8g8 c2b2 h7h5 h2h4 g8h8 b2a2 h8h6 d2e2 e
8d8 b7b8

[Paloma]

Great Work, Marek

[Graham Banks]

Why the topic was written?
Sting (bred on Stockfish 2.1.1) is a deep idea, and not a try to exist here.
As you know, Sting SF 2 can find such difficult sacrifices like this:
[d]8/5pKN/5Pp1/5kP1/8/Bb6/8/7q b - - bm Qxh7+;
without any tricks (one brain in one head) and can play on a high level (very important for me too! ).
However, a real challange for me are zugzwangs, especially the delayed by pawn moves.
Now, I found some intersting new idea and Sting is able to solve the two very difficult positions.
It is a motivation to work and to make Sting-sf-3 published.

[MikeGL]

Marek, where can we download this Sting3 for testing

[lech]

Graham it was a wonderful Polish day in Auckland (Speedway GP)

[jpqy]

Can't wait to test Sting SF3

JP.

[Graham Banks]

Graham it was a wonderful Polish day in Auckland (Speedway GP)

Yes - I saw some TV coverage of the Western Springs event.

[lech]

Can't wait to test Sting SF3

JP.

Jean-Paul believe me, I am working at Sting all the time. Sting is not able to get advantage against Critter or Houdini thanks to a worse static evaluation. Many people try to work for Stockfish but they can mix water only if they are not able to improve the static evoluation.
I try to get an intersting changes in endgames or tactical positions. The second goal is: to play without ELO loss.