Stockfish 1.3

[jpqy]

Here against Grapefruit...after checking the games,score is 6 - 4 win for Stockfish!

-----------------Grapefruit 1.0 alpha 3-----------------
Grapefruit 1.0 alpha 3 - Stockfish_13_win32_ja : 2,5/10 2-7-1 (0100=01000) 25% -191
-----------------Stockfish_13_win32_ja-----------------
Stockfish_13_win32_ja - Grapefruit 1.0 alpha 3 : 7,5/10 7-2-1 (1011=10111) 75% +191

JP.

[Tord Romstad]

I have merged a patch from Tord to allow Stockfish to use the same book of Glaurung 1.xx series.

I evidently need to comment my code better: That patch was meant for the iPhone version of Glaurung, which uses entirely different opening book code, which is not even written in the same programming language as the book code in Stockfish and Glaurung 2.x.

I see that there is already a 1.3.1 version which fixes this bug now.

Tord

[ernest]

Jean-Paul,

There is little sense posting the result of 10-game matches...

You should know better!

[jpqy]

Hi Ernest,

if you do 10 games 100 times is also 1000 games!
So every time i test a new engine i see directly after a few 10 games and different test suites how good this program will be.
Rybka and Naum you see directly it comes at top..also with 10 games.
If it's a very strong engine it will not stay down in a list.
And Stockfish will gave his place i my list when he has around 200 games!

JP.

[Robert Weck]

Hi Ernest,

if you do 10 games 100 times is also 1000 games!

But the statistical aberration is much too high!

One example:
In 1986 a german chess computer magazine published a result of 9:1 of a chess computer match between Computer A and Computer B. You would say the winner should be clearly stronger?

But the solution was: They simply let a Mephisto Amsterdam play against itself! The side with the white pieces was alternating called "Computer A" and "Computer B".

The more games you play, the less will be the statistical aberration. Every statistician will confirm that...


best regards,
Robert

[ernest]

Thanks Jim !

Hi Marco and Jim,
Indeed 1.3 and 1.3.1 are "almost" the same!

Start_position, 32-bit, 1 thread (for reproducibility), 512 MB hash

Analysis by Stockfish 1.3 JA:
1.d4 c5 2.dxc5 Na6 3.Nc3 Nxc5 4.b4 Na6 5.a3 Nf6 6.e4 h6
= (0.17) Depth: 12 00:00:01 877kN
1.Nf3 Nf6 2.d4 e6 3.e3 Nc6 4.Bd3 Be7 5.0-0 0-0 6.Nc3 d5
= (0.25) Depth: 12 00:00:01 1305kN
1.Nf3 Nf6 2.d4 e6 3.e3 d5 4.Bd2 Bd6 5.Nc3 Nc6 6.Bd3 0-0 7.0-0
= (0.21) Depth: 13 00:00:02 1636kN
1.e4
+/= (0.45) Depth: 13 00:00:02 2360kN
1.e4 e5 2.Nf3 Nf6 3.Nxe5 d6 4.Nc4 Nxe4 5.Qe2 d5 6.Nc3 Bf5 7.d3 dxc4 8.Nxe4
+/= (0.47) Depth: 14 00:00:06 5267kN
1.e4 e5 2.Nf3 Nf6 3.Nxe5 d6 4.Nc4 Nxe4 5.Qe2 Qe7 6.d3 Nf6 7.Nc3 d5 8.Ne5
+/= (0.39) Depth: 15 00:00:08 7486kN
1.e4 e5 2.Nf3 Nc6 3.Bb5 Nf6 4.0-0 Bc5 5.d3 Qe7 6.Nc3 0-0 7.Bxc6 dxc6 8.Be3 Bg4 9.Bxc5 Qxc5
+/= (0.41) Depth: 16 00:00:14 13149kN
1.e4 e5 2.Nf3 Nf6 3.Nxe5 d6 4.Nf3 Nxe4 5.Bd3 Nf6 6.Nc3 Be7 7.0-0 0-0 8.Re1 Nc6 9.Ng5
+/= (0.35) Depth: 17 00:00:27 25412kN
1.e4 e5 2.Nf3 Nf6 3.Nxe5 d6 4.Nf3 Nxe4 5.d3 Nf6 6.d4 Nc6 7.Nc3 Be7 8.Bd3 0-0 9.0-0 Be6
= (0.25) Depth: 18 00:00:53 49488kN

Analysis by Stockfish 1.3.1 JA:
1.d4 c5 2.dxc5 Na6 3.Nc3 Nxc5 4.b4 Na6 5.a3 Nf6 6.e4 h6
= (0.17) Depth: 12 00:00:01 877kN
1.Nf3 Nf6 2.d4 e6 3.e3 Nc6 4.Bd3 Be7 5.0-0 0-0 6.Nc3 d5
= (0.25) Depth: 12 00:00:01 1305kN
1.Nf3 Nf6 2.d4 e6 3.e3 d5 4.Bd2 Bd6 5.Nc3 Nc6 6.Bd3 0-0 7.0-0
= (0.21) Depth: 13 00:00:02 1636kN
1.e4
+/= (0.45) Depth: 13 00:00:02 2360kN
1.e4 e5 2.Nf3 Nf6 3.Nxe5 d6 4.Nc4 Nxe4 5.Qe2 d5 6.Nc3 Bf5 7.d3 dxc4 8.Nxe4
+/= (0.47) Depth: 14 00:00:06 5267kN
1.e4 e5 2.Nf3 Nf6 3.Nxe5 d6 4.Nc4 Nxe4 5.Qe2 Qe7 6.d3 Nf6 7.Nc3 d5 8.Ne5
+/= (0.39) Depth: 15 00:00:08 7486kN
1.e4 e5 2.Nf3 Nc6 3.Bb5 Nf6 4.0-0 Bc5 5.d3 Qe7 6.Nc3 0-0 7.Bxc6 dxc6 8.Be3 Bg4 9.Bxc5 Qxc5
+/= (0.41) Depth: 16 00:00:14 13149kN
1.e4 e5 2.Nf3 Nf6 3.Nxe5 d6 4.Nf3 Nxe4 5.Bd3 Nf6 6.Nc3 Be7 7.0-0 0-0 8.Re1 Nc6 9.Ng5
+/= (0.35) Depth: 17 00:00:27 25799kN
1.e4 e5 2.Nf3 Nf6 3.Nxe5 d6 4.Nf3 Nxe4 5.d3 Nf6 6.Nc3 Nc6 7.d4 Be7 8.Bd3 0-0 9.0-0 Be6
= (0.25) Depth: 18 00:00:53 50092kN

[mcostalba]

Thanks Jim !

Hi Marco and Jim,
Indeed 1.3 and 1.3.1 are "almost" the same!

The only difference is a change in the zobrist key initialization, in particular in the zobCastle keys used to update castle information.

Could be that this change caused a bit more/less hash collisions when looking up a position in the transposition table. And this could lead to a small difference in the searched nodes.

The difference is:

Code: Select all

-  for(int i = 15; i >= 0; i--)
-      zobCastle&#91;&#40;i&8&#41; | &#40;i&1&#41; | (&#40;i&2&#41; << 1&#41; | (&#40;i&4&#41; >> 1&#41;&#93; = genrand_int64&#40;);
+  for &#40;int i = 0; i < 16; i++)
+      zobCastle&#91;i&#93; = genrand_int64&#40;);

[Jouni]

Hi Marco,

Testing 1.3 JA I must say it's strong. But one thing: sometimes it takes
ridiculous time to start engine in Fritz 10. I just checked from clock 45s!
And before somebody asks: it's no hard disk swapping. I don't remember
this kind of delay previously with other engines.

Jouni

[mcostalba]

Hi Marco,

Testing 1.3 JA I must say it's strong. But one thing: sometimes it takes
ridiculous time to start engine in Fritz 10. I just checked from clock 45s!
And before somebody asks: it's no hard disk swapping. I don't remember
this kind of delay previously with other engines.

Jouni

If this behaviour is reproducible and you are interested I can send you a debug version that writes in a log file each initialization step togheter with timestamp, so that we could understund from where it comes from...

Let me know. BTW I didn't experienced this delay.

[Jouni]

Same 45s delay with Arena. Here's log, don't know if it helps:
(I send my email address in PM to You)

Arena 2.0
1.351**----------New game---2009-05-11 09:18:05,041 ma -------------
7.000*1*------------------------Starting engine 1 Stockfish-------------------------
7.000*1*Configured Engine 1 Type: UCI
7.121*1*Engine 1 dir: C:\Program Files\ChessBase\Engines\Stockfish
7.160*1*Engine 1 commandline: "C:\Program Files\ChessBase\Engines\Stockfish\Stockfish.exe"
49.751*1*Engine 1 ProcessID: 1416
49.891-->1:uci
50.041<--1:Stockfish 1.3 JA. Copyright (C) 2004-2008 Tord Romstad, Marco Costalba.
50.041<--1:id name Stockfish 1.3 JA
50.041<--1:id author Tord Romstad, Marco Costalba
50.041<--1:option name Use Search Log type check default false
50.041<--1:option name Search Log Filename type string default SearchLog.txt
50.041<--1:option name Book File type string default book.bin
50.041<--1:option name Mobility (Middle Game) type spin default 100 min 0 max 200
50.041<--1:option name Mobility (Endgame) type spin default 100 min 0 max 200
50.041<--1:option name Pawn Structure (Middle Game) type spin default 100 min 0 max 200
50.041<--1:option name Pawn Structure (Endgame) type spin default 100 min 0 max 200
50.041<--1:option name Passed Pawns (Middle Game) type spin default 100 min 0 max 200
50.041<--1:option name Passed Pawns (Endgame) type spin default 100 min 0 max 200
50.041<--1:option name Space type spin default 100 min 0 max 200
50.041<--1:option name Aggressiveness type spin default 100 min 0 max 200
50.041<--1:option name Cowardice type spin default 100 min 0 max 200
50.041<--1:option name King Safety Curve type combo default Quadratic var Quadratic var Linear
50.041<--1:option name King Safety Coefficient type spin default 40 min 1 max 100
50.041<--1:option name King Safety X Intercept type spin default 0 min 0 max 20
50.041<--1:option name King Safety Max Slope type spin default 30 min 10 max 100
50.041<--1:option name King Safety Max Value type spin default 500 min 100 max 1000
50.041<--1:option name Queen Contact Check Bonus type spin default 3 min 0 max 8
50.041<--1:option name Queen Check Bonus type spin default 2 min 0 max 4
50.041<--1:option name Rook Check Bonus type spin default 1 min 0 max 4
50.052<--1:option name Bishop Check Bonus type spin default 1 min 0 max 4
50.052<--1:option name Knight Check Bonus type spin default 1 min 0 max 4
50.052<--1:option name Discovered Check Bonus type spin default 3 min 0 max 8
50.052<--1:option name Mate Threat Bonus type spin default 3 min 0 max 8
50.052<--1:option name Check Extension (PV nodes) type spin default 2 min 0 max 2
50.052<--1:option name Check Extension (non-PV nodes) type spin default 1 min 0 max 2
50.052<--1:option name Single Reply Extension (PV nodes) type spin default 2 min 0 max 2
50.052<--1:option name Single Reply Extension (non-PV nodes) type spin default 2 min 0 max 2
50.052<--1:option name Mate Threat Extension (PV nodes) type spin default 0 min 0 max 2
50.052<--1:option name Mate Threat Extension (non-PV nodes) type spin default 0 min 0 max 2
50.052<--1:option name Pawn Push to 7th Extension (PV nodes) type spin default 1 min 0 max 2
50.052<--1:option name Pawn Push to 7th Extension (non-PV nodes) type spin default 1 min 0 max 2
50.052<--1:option name Passed Pawn Extension (PV nodes) type spin default 1 min 0 max 2
50.052<--1:option name Passed Pawn Extension (non-PV nodes) type spin default 0 min 0 max 2
50.052<--1:option name Pawn Endgame Extension (PV nodes) type spin default 2 min 0 max 2
50.052<--1:option name Pawn Endgame Extension (non-PV nodes) type spin default 2 min 0 max 2
50.052<--1:option name Full Depth Moves (PV nodes) type spin default 14 min 1 max 100
50.052<--1:option name Full Depth Moves (non-PV nodes) type spin default 3 min 1 max 100
50.052<--1:option name Threat Depth type spin default 5 min 0 max 100
50.052<--1:option name Futility Pruning (Main Search) type check default true
50.052<--1:option name Futility Pruning (Quiescence Search) type check default true
50.052<--1:option name LSN filtering type check default false
50.052<--1:option name LSN Time Margin (sec) type spin default 4 min 1 max 10
50.052<--1:option name LSN Value Margin type spin default 200 min 100 max 600
50.063<--1:option name Randomness type spin default 0 min 0 max 10
50.112<--1:option name Minimum Split Depth type spin default 4 min 4 max 7
50.112<--1:option name Maximum Number of Threads per Split Point type spin default 5 min 4 max 8
50.112<--1:option name Threads type spin default 1 min 1 max 8
50.112<--1:option name Hash type spin default 32 min 4 max 4096
50.112<--1:option name Clear Hash type button
50.112<--1:option name Ponder type check default true
50.112<--1:option name OwnBook type check default true
50.112<--1:option name MultiPV type spin default 1 min 1 max 500
50.112<--1:option name UCI_ShowCurrLine type check default false
50.112<--1:option name UCI_Chess960 type check default false
50.112<--1:uciok
50.112-->1:setoption name Hash value 128
50.112-->1:isready
50.663<--1:readyok
51.103-->1:isready
51.113<--1:readyok
51.113-->1:ucinewgame
51.113-->1:isready
51.113<--1:readyok
51.124-->1:position startpos
51.124-->1:go infinite
51.344<--1:info depth 2
51.353<--1:info depth 2 score cp 11 time 220 nodes 49 nps 222 pv g1f3 g8f6
51.353<--1:info depth 3
51.353<--1:info depth 3 score cp 60 time 220 nodes 171 nps 777 pv g1f3 g8f6 b1c3
51.353<--1:info depth 4
51.353<--1:info depth 4 score cp 11 time 220 nodes 404 nps 1836 pv g1f3 g8f6 b1c3 b8c6
51.364<--1:info depth 5
51.374<--1:info depth 5 score cp 39 time 240 nodes 1291 nps 5379 pv g1f3 g8f6 b1c3 b8c6 d2d4
51.374<--1:info depth 5 score cp 43 time 240 nodes 1958 nps 8158 pv d2d4 b8c6 b1c3 g8f6 e2e4
51.393<--1:info depth 6
51.393<--1:info depth 6 score cp 11 time 250 nodes 4257 nps 17028 pv d2d4 b8c6 b1c3 d7d5 g1f3 g8f6
51.393<--1:info depth 6 score cp 13 time 250 nodes 5938 nps 23752 pv e2e4 b8c6 b1c3 g8f6 d2d4 d7d5
51.393<--1:info depth 7
51.414<--1:info depth 7 score cp 15 time 270 nodes 9159 nps 33922 pv e2e4 b8c6 b1c3 g8f6 g1f3 d7d5 f1d3
51.423<--1:info depth 7 score cp 31 time 280 nodes 11657 nps 41632 pv d2d4 b8c6 b1c3 d7d5 g1f3 g8f6 c1f4
51.423<--1:info depth 7 score cp 37 time 290 nodes 15274 nps 52668 pv e2e3 b8c6 g1f3 g8f6 f1d3 c6b4 e1g1 b4d3 c2d3
51.434<--1:info depth 8
51.455<--1:info depth 8 score cp 29 time 320 nodes 22832 nps 71350 pv e2e3 c7c5 g1f3 g8f6 f1d3 b8c6 e1g1 d7d5
51.485<--1:info depth 9
51.515<--1:info depth 9 score cp 43 time 381 nodes 43180 nps 113333 pv e2e3 b8c6 g1f3 g8f6 b1c3 d7d5 f1d3 d5d4 e3d4 c6d4 f3d4 d8d4
51.534<--1:info depth 10
51.614<--1:info depth 10 score cp 17 time 481 nodes 76067 nps 158143 pv e2e3 b8c6 g1f3 e7e5 b1c3 f8d6 f1d3 g8f6 e1g1 e8g8
51.765<--1:info depth 10 score cp 35 time 631 nodes 125625 nps 199088 pv e2e4 b8c6 g1f3 g8f6 f1d3 d7d5 e4d5 f6d5 e1g1 e7e5
51.814<--1:info depth 11
51.954<--1:info depth 11 score cp 17 time 831 nodes 193935 nps 233375 pv e2e4 b8c6 g1f3 g8f6 b1c3 d7d5 e4d5 f6d5 f1c4 c8e6 c3d5 e6d5
52.196<--1:info nodes 270000 nps 251865 time 1072 hashfull 0
52.565<--1:info depth 11 score cp 29 time 1432 nodes 392643 nps 274192 pv d2d4 d7d5 e2e3 c7c5 d4c5 e7e6 b2b4 b8c6 c2c3 g8f6 g1f3
52.565<--1:info currmove c2c4 currmovenumber 4
52.565<--1:info currmove b1a3 currmovenumber 5
52.576<--1:info currmove g1h3 currmovenumber 6
52.586<--1:info currmove d2d3 currmovenumber 7
52.595<--1:info currmove b2b3 currmovenumber 8
52.606<--1:info currmove h2h3 currmovenumber 9
52.606<--1:info currmove g2g3 currmovenumber 10
52.606<--1:info currmove h2h4 currmovenumber 11
52.606<--1:info currmove a2a4 currmovenumber 12
52.616<--1:info currmove a2a3 currmovenumber 13
52.616<--1:info currmove b1c3 currmovenumber 14
52.625<--1:info currmove g2g4 currmovenumber 15
52.635<--1:info currmove b2b4 currmovenumber 16
52.635<--1:info currmove f2f4 currmovenumber 17
52.635<--1:info currmove c2c3 currmovenumber 18
52.635<--1:info currmove g1f3 currmovenumber 19
52.646<--1:info currmove f2f3 currmovenumber 20
52.646<--1:info depth 12
52.646<--1:info currmove d2d4 currmovenumber 1
53.247<--1:info nodes 600000 nps 285306 time 2103 hashfull 0
54.068<--1:info depth 12 score cp 17 time 2934 nodes 877389 nps 299041 pv d2d4 c7c5 d4c5 b8a6 b1c3 a6c5 b2b4 c5a6 a2a3 g8f6 e2e4 h7h6
54.068<--1:info currmove e2e4 currmovenumber 2
54.087<--1:info currmove e2e3 currmovenumber 3
54.187-->1:stop
54.227<--1:info nodes 930000 nps 299613 time 3104 hashfull 0
54.227<--1:info nodes 930095 nps 299644 time 3104 hashfull 0
54.227<--1:bestmove d2d4 ponder c7c5

Jouni