sting-sf-67

[lech]

download: http://www.mediafire.com/download/54pqm ... -sf-67.zip

[Paloma]

Thanks Marek

[Graham Banks]

Thanks Marek.

[Damir]

Thanks Marek.

[lech]

A few words about the new 6.7 version of Sting.
Sting is based on Stockfish 2.1.1 as earlier.
I found a serious bug in Hash management (Stockfish 2.1.1).
Now you may use more of Hash if you can and Sting should return a correct size.
It may be very important for users of 64-bit versions to try solve many very difficult positions.
Sting is quasi-intelligent, "one brain in one head" engine. It means that unlike some other ELO engines, Sting can try to solve tough positions ("horizon effect") without an artifical or the human intelligence based on some tricks like: Monte Carlo, nullmove disabling, start parameters (or versions), "rule 50", tablebases or other.
Of course, it is very important to do it without a serious ELO loss too.
I added a few tools (code) making Sting more selective. It means I diagnosed many branch of problems leading to the horizon effect.
Maybe one branche "lazy King" needs still a code, but first I will check the code of Stockfish 2.1.1 (too many bugs). The code of search and static evaluation (passed pawns) need my look too.
I decided to leave some other branch: endgame question (like KQ/KP or other) for next versons. I have an interesting idea to solve the problem of all such endgames too.

Sting needs still some solid work. But I believe that it is the only correct way.

[velmarin]

A few words about the new 6.7 version of Sting.
Sting is based on Stockfish 2.1.1 as earlier.
I found a serious bug in Hash management (Stockfish 2.1.1).

Maybe one branche "lazy King" needs still a code, but first I will check the code of Stockfish 2.1.1 (too many bugs). The code of search and static evaluation (passed pawns) need my look too.

Another option is to go to a later version...
We must think that if Stockfish is number ONE now, is that improved everything.
pass code what is updated. It is as simple as that.

[carldaman]

It is a little self-defeating to use an inferior code-base, when better versions are readily available. Even Stockfish 2.2.2 was both stronger and had a much more attractive playing style than v2.1.1.

However, I'm sure that Marek's goal is not to compete for a top position with a derivative, so that's why he does not use the latest SF sources.

My recommendation is to use something at least as strong as Stockfish DD, which was a distinct improvement over earlier versions while sporting a pleasant attacking style.

Regards,
CL

A few words about the new 6.7 version of Sting.
Sting is based on Stockfish 2.1.1 as earlier.
I found a serious bug in Hash management (Stockfish 2.1.1).

Maybe one branche "lazy King" needs still a code, but first I will check the code of Stockfish 2.1.1 (too many bugs). The code of search and static evaluation (passed pawns) need my look too.

Another option is to go to a later version...
We must think that if Stockfish is number ONE now, is that improved everything.
pass code what is updated. It is as simple as that.

[lech]

I think that it would be easier to move my ideas to the higher version of Stockfish.
But first I have to end my work at my tools.
Today I found here a very interesying example::
http://talkchess.com/forum/viewtopic.php?t=59264

Thanks to this I found how to improve one of my tools (much !).
It is worth to wait for the next version!

Sting SF 7 beta (2CPU, 1024 HASH, ~40 minutes)
[d]6bn/p4p1k/rp3Pp1/rp2BBP1/1p6/pK6/P2P4/8 w - - (#13);

Code: Select all

info depth 42 multipv 1 score cp 0 nodes 2056327836 nps 968481 time 2123250 pv f5c8 a5a4 d2d4 a6a5 d4d5 a7a6 c8f5 g6f5 d5d6 h8g6 d6d7 g6e5 d7d8q f5f4 d8d4 h7g6 d4e4 g6g5 e4e5 g5g4 e5e4 g4g3 e4f5 g3h4 f5f4 h4h5 f4f5 h5h4 f5f4
info depth 43
info currmove f5c8 currmovenumber 1
info depth 43 multipv 1 score cp 8 lowerbound nodes 2386547085 nps 963063 time 2478078 pv f5c8
info currmove f5c8 currmovenumber 1
info depth 43 multipv 1 score cp 16 lowerbound nodes 2386547228 nps 963057 time 2478094 pv f5c8
info currmove f5c8 currmovenumber 1
info depth 43 multipv 1 score cp 28 lowerbound nodes 2386547462 nps 963057 time 2478094 pv f5c8
info currmove f5c8 currmovenumber 1
info depth 43 multipv 1 score cp 46 lowerbound nodes 2386547817 nps 963057 time 2478094 pv f5c8
info currmove f5c8 currmovenumber 1
info depth 43 multipv 1 score cp 73 lowerbound nodes 2386591025 nps 963044 time 2478172 pv f5c8
info currmove f5c8 currmovenumber 1
info depth 43 multipv 1 score cp 114 lowerbound nodes 2386610089 nps 963040 time  2478203 pv f5c8
info currmove f5c8 currmovenumber 1
info depth 43 multipv 1 score cp 175 lowerbound nodes 2386740335 nps 963020 time 2478391 pv f5c8
info currmove f5c8 currmovenumber 1
info depth 43 multipv 1 score cp 267 lowerbound nodes 2387022369 nps 962976 time 2478797 pv f5c8
info currmove f5c8 currmovenumber 1
info depth 43 multipv 1 score cp 404 lowerbound nodes 2467954022 nps 959488 time 2572156 pv f5c8
info currmove f5c8 currmovenumber 1
info depth 43 multipv 1 score cp 609 lowerbound nodes 2499489149 nps 959209 time 2605781 pv f5c8
info currmove f5c8 currmovenumber 1
info depth 43 multipv 1 score cp 916 lowerbound nodes 2527370869 nps 957688 time 2639031 pv f5c8
info currmove f5c8 currmovenumber 1
info depth 43 multipv 1 score mate 16 lowerbound nodes 2527372457 nps 957683 time 2639047 pv f5c8
info depth 43 multipv 1 score mate 16 lowerbound nodes 2527372457 nps 957683 time 2639047 pv f5c8

[Vinvin]

Sting-SF is a bit deceiving on this position
( from http://www.talkchess.com/forum/viewtopi ... 117#672117 )
[pgn][Event "18.p WCCT7 Ty (tt)"]
[Site "?"]
[Date "2004.??.??"]
[Round "?"]
[White "Predrag=N"]
[Black "(=1031.57d3a1)"]
[Result "1/2-1/2"]
[SetUp "1"]
[FEN "7Q/4p3/4p3/p1p1P3/Pp2P3/3Kp3/p1PbP3/kN6 w - - 0 1"]
[PlyCount "22"]
[EventDate "2004.??.??"]

{EG#14182} 1. Na3 (1. Nxd2 $2 exd2 2. Kxd2 Kb2 3. Qh3 c4 $1) 1... bxa3 (1...
Kb2 2. Nc4+ Kb1 3. Qh1+ Bc1 4. Nxa5 $1 a1=Q 5. Qxc1+ Kxc1 6. Nb3+) 2. Qh1+ $1
Kb2 3. Qa1+ $1 Kxa1 4. c4 $1 Bc1 (4... Kb1) (4... Be1 5. Kc2 Bd2 6. Kd1 Kb1) 5.
Kc2 Bb2 6. Kb3 $1 Bxe5 (6... Kb1) 7. Kc2 Bb2 8. e5 $1 (8. Kb3 $2 Kb1 $1 9. e5
a1=N#) 8... Bxe5 9. Kc1 Bc3 (9... Bb2+ 10. Kc2) 10. Kc2 Bb2 (10... Bd2 11. Kd1
$1 (11. Kd3 $2 Kb2) 11... Kb1) 11. Kb3 Kb1 1/2-1/2

[/pgn]