Stockfish

Thomas Miller wrote:

Vladimir wrote:

Stockfish 4 x64 Gauntlet
https://www.chess2u.com/t7501-stockfish-4-x64-gauntlet#45734

Code:

     Program                          Elo    +   -   Games   Score   Av.Op.  Draws

     Stockfish 4 x64                : 3229   13  13  1500    51.0 %   3222   48.7 %
     Stockfish 3 JA 64bit           : 3174    9   9  2650    44.1 %   3215   53.5 %

https://www.chess2u.com/t6886p15-engineswar-rating-list-updated#45735

Amazing results! 55 Elo over its predecessor.

Thanks for sharing your efforts Vladimir!

Same here:

Marco Costalba wrote:

According to our internal tests we have increased about 55 ELO in just the last 4 months

http://www.chessdom.com/stockfish-4-to-play-in-the-new-season-of-tcec/

Wow...it happened exactly as he predicted..

Congratulations to the Stockfish team.

Me too..
+56 ELO !

Stockfish (26-08-2013)

Windows x64 for modern computers + sse4.2
Windows x64 for modern computers
Windows x64
Windows 32
Linux x64 for modern computers + sse4.2
Linux x64 for modern computers
Linux x64

changelog:
https://github.com/mcostalba/Stockfish/commits/master


Time management: move faster if PV is stable

Move faster but compensate by allocating more
time when the best move changes.

Passed short TC at 15+0.05
LLR: 2.93 (-2.94,2.94)
Total: 13895 W: 3030 L: 2882 D: 798

Long TC at 60+0.05
LLR: 2.96 (-2.94,2.94)
Total: 9266 W: 1777 L: 1624 D: 5865

At time increment 30+0.5
LLR: 2.96 (-2.94,2.94)
Total: 6703 W: 1238 L: 1134 D: 4331

And at fixed game number, longer TC 120+0.05
ELO: 5.17 +-2.8 (95%) LOS: 100.0%
Total: 19306 W: 3378 L: 3091 D: 12837

Wait...Time Management is still being tuned! A big step was made...but the end will be even better.

I´m predicting around ~10 ELO over SF4 (maybe more) on self-testing and much more (like ~20 ELO) against foreign opponents like Houdini and Critter as they use "time management tactics" hardcore to win games. This will be the end of false magicians.

Stockfish number 1 soon!

@Kallah wrote:

Wait...Time Management is still being tuned! A big step was made...but the end will be even better.

I´m predicting around ~10 ELO over SF4 (maybe more) on self-testing and much more (like ~20 ELO) against foreign opponents like Houdini and Critter as they use "time management tactics" hardcore to win games. This will be the end of false magicians.

Stockfish number 1 soon!

Nice...well said.

can anyone do engine matches (at least 100 matches) for the latest

Stockfish (29-08-2013) engine?

I'm doing the tourney too

and my result says Stockfish (29-08-2013) is super strong

and stronger than all the other engines in that tourney I'm running now

we should compare the engine match results!!

@thorhammer wrote:

can anyone do engine matches (at least 100 matches) for the latest

Stockfish (29-08-2013) engine?

I'm doing the tourney too

and my result says Stockfish (29-08-2013) is super strong

and stronger than all the other engines in that tourney I'm running now

we should compare the engine match results!!

What Settings, Thorhammer ?

the stockfish 2013 8 30

is up there!!!

can any of you do the test for stockfish 2013 9 1 ? 

Stockfish 020913 (2013 9 2)  -  Houdini 3 Pro w32


***


started a new tourney with the aforementioned engines

after hundreds of matches
I'll post the whole pgn file and the results    

now I know how to use
"use the same opening line twice"

so each engine will use the exact same opening in reversed colors!!!


***


http://abrok.eu/stockfish/


***


Stockfish 020913 (2013 9 2)

Rewrite KBBKN endgame

This was thought to be a draw but the bishops generally win. However,
it takes up to 66 moves. The position in the diagram was thought to be
a draw for over one hundred years, but tablebases show that White wins
in 45 moves. All of the long wins go through this type of semi-fortress
position. It takes several moves to force Black out of the temporary
fortress in the corner; then precise play with the bishops prevents Black
from forming the temporary fortress in another corner (Nunn 1995:265ff).

Before computer analysis, Speelman listed this position as unresolved,
but "probably a draw" (Speelman 1981:109).

100 games played / Tournament is finished
Name of the tournament: SF4Dev10
Site/ Country: RICARDO-PC, United States
Level: Tournament Game in 5 Minutes
Hardware: Intel(R) Core(TM)2 Duo CPU E4600 @ 2.40GHz with 3,072 MB Memory
Operating system: Windows 7 Ultimate Professional (Build 7600)
PGN-File: C:\Program Files\Arena\Tournaments\SF4Dev10.pgn

2 Cores

Engine Score
1: Stockfish_13090118_32bit 53.0/100
2: Houdini_20_w32 47.0/100

Rank Name Elo + - games score oppo. draws
1 Stockfish_13090118_32bit 10 29 29 100 53% -10 30%
2 Houdini_20_w32 -10 29 29 100 47% 10 30%

5 min. No more losses on time!! SF4dev already ahead of Houdini 2 at 5 min!!! SF was ahead all time.

PGN file all 100 games . (small sample due to time control - 5 min)

sendspace.com vbzgbz

@thorhammer wrote:

Stockfish 020913 (2013 9 2)  -  Houdini 3 Pro w32


***


started a new tourney with the aforementioned engines

after hundreds of matches
I'll post the whole pgn file and the results    

now I know how to use
"use the same opening line twice"

so each engine will use the exact same opening in reversed colors!!!


***


http://abrok.eu/stockfish/


***


Stockfish 020913 (2013 9 2)

Rewrite KBBKN endgame

This was thought to be a draw but the bishops generally win. However,
it takes up to 66 moves. The position in the diagram was thought to be
a draw for over one hundred years, but tablebases show that White wins
in 45 moves. All of the long wins go through this type of semi-fortress
position. It takes several moves to force Black out of the temporary
fortress in the corner; then precise play with the bishops prevents Black
from forming the temporary fortress in another corner (Nunn 1995:265ff).

Before computer analysis, Speelman listed this position as unresolved,
but "probably a draw" (Speelman 1981:109).

the tourney is almost finished and I'll post the whole pgn and the result soon!! 


***


I only ran like 100 something matches or so for the engines

thus the statistical noise and the size of error (in Elo ratings fluctuation) are big

so the result here is not very accurate
(because we have to run the matches for like 1000 times)

so can any of us do hundreds or matches (at least 500) for the engines to get the accurate tourney results for all the chess enthusiasts?  


***



Stockfish 020913 (2013 9 2)  -  Houdini 3 Pro w32


140 matches were done




Stockfish 020913          : 140 (+ 46,= 51,- 43)

Houdini 3 Pro w32            




the whole pgn match file:
mediafire.com ?eb4xdxtl3i8417r

@thorhammer wrote:

@thorhammer wrote:

Stockfish 020913 (2013 9 2)  -  Houdini 3 Pro w32


***


http://abrok.eu/stockfish/



Stockfish 020913 (2013 9 2)

Rewrite KBBKN endgame

This was thought to be a draw but the bishops generally win. However,
it takes up to 66 moves. The position in the diagram was thought to be
a draw for over one hundred years, but tablebases show that White wins
in 45 moves. All of the long wins go through this type of semi-fortress
position. It takes several moves to force Black out of the temporary
fortress in the corner; then precise play with the bishops prevents Black
from forming the temporary fortress in another corner (Nunn 1995:265ff).

Before computer analysis, Speelman listed this position as unresolved,
but "probably a draw" (Speelman 1981:109).

the tourney is almost finished and I'll post the whole pgn and the result soon!! 


***


I only ran like 100 something matches or so for the engines

thus the statistical noise and the size of error (in Elo ratings fluctuation) are big

so the result here is not very accurate
(because we have to run the matches for like 1000 times)

so can any of us do hundreds or matches (at least 500) for the engines to get the accurate tourney results for all the chess enthusiasts?  


***



Stockfish 020913 (2013 9 2)  -  Houdini 3 Pro w32


140 matches were done


Stockfish 020913          : 140 (+ 46,= 51,- 43)

Houdini 3 Pro w32            




the whole pgn match file:
mediafire.com ?eb4xdxtl3i8417r

more matches are being done

after 170 matches
they made a tie



Stockfish 020913 : 170 (+ 56,= 58,- 56), 50.0 %

Houdini 3 Pro w32 : 170 (+ 56,= 58,- 56), 50.0 %

after hundreds of more matches
the full pgn file will be posted

about 300 matches are done in total
between Stockfish 020913 (2013 9 2)  -  Houdini 3 Pro w32


and the elostat says
the + - error gap, elo fluctuation is


only
+33 and -33
so elo ratings become more accurate as time goes on  

Stockfish 020913 (2013 9 2)  -  Houdini 3 Pro w32

140 matches were done


Stockfish 020913          : 140 (+ 46,= 51,- 43)

Houdini 3 Pro w32            



the whole pgn match file:
mediafire.com ?eb4xdxtl3i8417r


***


more matches are being done

after 170 matches
they made a tie


Stockfish 020913 : 170 (+ 56,= 58,- 56), 50.0 %

Houdini 3 Pro w32 : 170 (+ 56,= 58,- 56), 50.0 %

after hundreds of more matches
the full pgn file will be posted


***


about 300 matches are done in total
between Stockfish 020913 (2013 9 2)  -  Houdini 3 Pro w32


and the elostat says
the + - error gap, elo fluctuation is


only
+33 and -33
so elo ratings become more accurate as time goes on   


***


1 Houdini 3 Pro w32              : 3241  
3 Stockfish 020913               : 3234


Houdini 3 Pro w32         : 294 (+ 98,=105,- 91),
Stockfish 020913              :  


Houdini 3 Pro w32 is the king of the kings!!



the whole pgn matches download:  

mediafire.com view/asu724s252d7b64/houdini_king.pgn

Margin seems narrow, don't forget the hardware used changes the standings of the engines... stronger computer = longer time controls. So when CCRL tests with 4 cpus, well then...

...I'm assuming they test with 4 cpus not 4 cores, or I'll rage.
CCRL's only entry of Stockfish 4 has it trailing by Houdini by only 11 elo.

I see big improvements and important fixes in the development versions , like the fix of the endgame of KBP vs K where the bishop is of wrong colour, therefore I think it worth releasing an official version like Stockfish 4.1

Stockfish (5-09-2013)

Windows x64 for modern computers + sse4.2
Windows x64 for modern computers
Windows x64
Windows 32
Linux x64 for modern computers + sse4.2
Linux x64 for modern computers
Linux x64

changelog:
https://github.com/mcostalba/Stockfish/commits/master


Revert "Fix check for bishop pair in material imbalance"

Idea is sound but implementation is partial. Ryan and Joona noticed that
we leave an hole in material table. Also we got another report by an user
of an odd behaviour. Namely, if you start stockfish and from the prompt
give 'bench' you get 3453941, then if you run again bench you get 3453940.

The reason is that two different positions with the same number of pieces,
but one with a bishop pair and another without have the same material key.
But after Eelco patch also different material imbalance and this yields
to this issue.

Restesting at long TC shows the patch does not really contribute at
ELO improvement. Actually patch failed at long TC.

LLR: -2.97 (-2.94,2.94)
Total: 23109 W: 4104 L: 4092 D: 14913

So revert.

can any 1 help for configuration setting of stockfish 4 x64 , and any tips to make engine strong ?

@DARE_DEVIL wrote:

can any 1 help for configuration setting of stockfish 4 x64 ,  and any tips  to  make engine strong ?

https://www.chess2u.com/t7513-stockfish-tuning

Thanks nicolas

Stockfish (7-09-2013)

Windows x64 for modern computers + sse4.2
Windows x64 for modern computers
Windows x64
Windows 32
Linux x64 for modern computers + sse4.2
Linux x64 for modern computers
Linux x64

changelog:
https://github.com/mcostalba/Stockfish/commits/master


Singular extension at 8 plies also for PV nodes

Passed both short TC:
LLR: 2.96 (-2.94,2.94)
Total: 11451 W: 2455 L: 2282 D: 6714

And long TC
LLR: 2.96 (-2.94,2.94)
Total: 15813 W: 2907 L: 2723 D: 10183


Union of 2 changes

Union of
- LMR >= 3 plies from Gary tests.stockfishchess.org/tests/view/522522960ebc595d328fcafd

- allows() tweak from Reuven tests.stockfishchess.org/tests/view/5225fa1c0ebc595d328fcb53

Both passed Step I and failed Step II.

Instead this union passed both short TC:
LLR: 2.95 (-2.94,2.94)
Total: 14525 W: 3063 L: 2874 D: 8588

And long TC
LLR: 2.94 (-2.94,2.94)
Total: 31075 W: 5566 L: 5308 D: 20201

master vs sf_4_bas

ELO: 24.54 +-3.1 (95%) LOS: 100.0%
Total: 17430 W: 3786 L: 2557 D: 11087

http://tests.stockfishchess.org/tests

Vladimir wrote:

master vs sf_4_bas

ELO: 24.54 +-3.1 (95%) LOS: 100.0%
Total: 17430 W: 3786 L: 2557 D: 11087

http://tests.stockfishchess.org/tests

Big Vladimir! Where are your tests again after day 26? I´m missing them!

Best regards!

@Kallah wrote:

Big Vladimir! Where are your tests again after day 26? I´m missing them!

Best regards!

Stockfish 07-09-2013 64bit Gauntlet
https://www.chess2u.com/t7142p30-stockfish-under-microscope#46175

Let's Get This Party Started!

Dark_wizzie wrote:

I believe Houdini 3 was heavily optimized for blitz games though and as a result is much stronger than Houdini 2 in blitz. Correct me if I'm wrong.

In longer time controls however Stockfish creeps up very closely to Houdini 3!

Passed it actually...checkout latest versions !