Testing Stockfish 11-03-13. 480 Games.

Discussion of computer chess matches and engine tournaments.

Moderators: hgm, Rebel, chrisw

Tomcass
Posts: 786
Joined: Sun Apr 16, 2006 9:09 pm

Re: Testing Stockfish 15-03-13. 480 Games.

Post by Tomcass »

STOCKFISH 3 IPMAN : 240 GAMES

i7 975 3.33 Ghz.
4 real cores
GUI: Fritz 12
Book: Fritz 12
Time control: 4 min + 0 sec/ game
Ponder: Off
No tablebases

20130430_SF3_IP_4+0 2013

Stockfish 3 64bit SSE4.2 - Critter 1.6a 64-bitnob_4 23.5 - 16.5 +12/=23/-5 58.75%
Stockfish 3 64bit SSE4.2 - Deep Rybka 4 x64 (x4) 22.5 - 17.5 +11/=23/-6 56.25%
Stockfish 3 64bit SSE4.2 - Houdini 3 Pro x64_4 18.0 - 22.0 +9/=18/-13 45.00%

i7 975 3.33 Ghz.
4 real cores
GUI: Fritz 12
Book: Fritz 12
Time control: 2 min + 2 sec/ game
Ponder: Off
No tablebases

20130430_SF3_2+2 2013

Stockfish 3 64bit SSE4.2 - Critter 1.6a 64-bitnob_4 22.0 - 18.0 +9/=26/-5 55.00%
Stockfish 3 64bit SSE4.2 - Deep Rybka 4 x64 (x4) 24.0 - 16.0 +12/=24/-4 60.00%
Stockfish 3 64bit SSE4.2 - Houdini 3 Pro x64_4 20.5 - 19.5 +9/=23/-8 51.25%

240 games : http://www.mediafire.com/?19el5s20wn5r1he

Overall average with 4 cores ( 130.5– 109.5) : 54.37%

Against: Critter 1.6: 56.87% / Deep Rybka 4: 58.12%/ Houdini 3.0: 48.12%/


This time my test is only 240 games instead of 480. In my I7 980 3.33 Ghz. 6 real cores I have not been able to run properly SF3. I tried it with 2 different compiles and the result is: Although in the Fritz layout it appears 6 cores, in fact it works only with one.

Anyway the result of my test with 4 cores is excellent: 54.37% is a great result, including an impressive 48.12% against the king Houdini 3.0 Pro.

Regards from Barcelona.

Tom.
Tomcass
Posts: 786
Joined: Sun Apr 16, 2006 9:09 pm

Re: Testing Stockfish 15-03-13. 480 Games.

Post by Tomcass »

STOCKFISH 02-05-13: 480 GAMES

I7 980 3.33 Ghz.
6 real cores
Ponder: Off.
GUI: Fritz 12
Book: Perfect 2012
Time control: 4 min +0 sec/ game
No tablebases

SF020513 64p_6 - Critter 1.6 64-bitx6_nob 10-24-6 22.0-18.0 55.00%
SF020513 64p_6 - Deep Rybka 4.1 SSE42 x64 (x6) 12-22-6 23.0-17.0 57.50%
SF020513 64p_6 - Houdini 3 Pro x64_6 6-20-14 16.0-24.0 40.00%

I7 980 3.33 Ghz.
6 real cores
Ponder: Off.
GUI: Fritz 12
Book: Perfect 2012
Time control: 2 min +2 sec/ game
No tablebases

SF020513 64p_6 - Critter 1.6 64-bitx6_nob 9-22-9 20.0-20.0 50.00%
SF020513 64p_6 - Deep Rybka 4.1 SSE42 x64 (x6) 11-21-8 21.5-18.5 53.75%
SF020513 64p_6 - Houdini 3 Pro x64_6 4-25-11 16.5-23.5 40.63%

Overall average with 6 cores ( 120.0– 120.0) : 50.00%

X6 240 games: http://www.mediafire.com/?8y4caislsfd99ms

i7 975 3.33 Ghz.
4 real cores
GUI: Fritz 12
Book: Fritz 12
Time control: 4 min + 0 sec/ game
Ponder: Off
No tablebases

SF020513 64p- Critter 1.6a 64-bitnob_4 7-30-3 22.0-18.0 55.00%
SF020513 64p- Deep Rybka 4 x64 (x4) 8-25-7 20.5-19.5 51.25%
SF020513 64p- Houdini 3 Pro x64_4 9-15-16 16.5-23.5 40.63%

i7 975 3.33 Ghz.
4 real cores
GUI: Fritz 12
Book: Fritz 12
Time control: 2 min + 2 sec/ game
Ponder: Off
No tablebases

SF020513 64p - Critter 1.6a 64-bitnob_4 9-25-6 21.5-18.5 53.75%
SF020513 64p - Deep Rybka 4 x64 (x4) 15-21-5 24.5-15.5 61.25%
SF020513 64p - Houdini 3 Pro x64_4 5-23-12 16.5-23.5 40.63%

Overall average with 4 cores ( 120.5– 119.5) : 50.21%

X4 240 games: http://www.mediafire.com/?1whi2l2q6enfg2y

Global Score: 50.10%
Against Critter 1.6: 53.44% Deep Rybka 4: 55.94% Houdini 3.0: 40.93%



Not a new record this time. :-)

Regards,

Tom.
Tomcass
Posts: 786
Joined: Sun Apr 16, 2006 9:09 pm

Re: Testing Stockfish 15-03-13. 480 Games.

Post by Tomcass »

STOCKFISH 03-05-13: 480 GAMES

I7 980 3.33 Ghz.
6 real cores
Ponder: Off.
GUI: Fritz 12
Book: Perfect 2012
Time control: 4 min +0 sec/ game
No tablebases

SF030513_6 - Critter 1.6 64-bitx6_nob 10-25-5 22.5-17.5 56.25%
SF030513_6 - Deep Rybka 4.1 SSE42 x64 (x6) 16-21-3 26.5-13.5 66.25%
SF030513_6 - Houdini 3 Pro x64_6 5-20-15 15.0-25.0 37.50%

I7 980 3.33 Ghz.
6 real cores
Ponder: Off.
GUI: Fritz 12
Book: Perfect 2012
Time control: 2 min +2 sec/ game
No tablebases

SF030513_6 - Critter 1.6 64-bitx6_nob 7-25-8 19.5-20.5 48.75%
SF030513_6 - Deep Rybka 4.1 SSE42 x64 (x6) 8-25-7 20.5-19.5 51.25%
SF030513_6 - Houdini 3 Pro x64_6 6-19-15 15.5-24.5 38.75%

X6 SF030513 120 games: http://www.mediafire.com/?vpeuu1ys596v565
X6 SF030513 games 121 to 240: http://www.mediafire.com/?1pfvza115ne8l98

Overall average with 6 cores: (119.5 – 120.5) 49.79%

i7 975 3.33 Ghz.
4 real cores
GUI: Fritz 12
Book: Fritz 12
Time control: 4 min + 0 sec/ game
Ponder: Off
No tablebases

SF030513- Critter 1.6a 64-bitnob_4 8-23-9 19.5 – 20.5 48.75%
SF030513- Deep Rybka 4 x64 (x4) 13-21-6 23.5-16.5 58.75%
SF030513- Houdini 3 Pro x64_4 4-21-15 14.5-25.5 36.25%

i7 975 3.33 Ghz.
4 real cores
GUI: Fritz 12
Book: Fritz 12
Time control: 2 min + 2 sec/ game
Ponder: Off
No tablebases

SF030513 - Critter 1.6a 64-bitnob_4 5-31-4 20.5-19.5 51.25%
SF030513 - Deep Rybka 4 x64 (x4) 5-31-4 20.5-19.5 51.25%
SF030513 - Houdini 3 Pro x64_4 6-22-12 17.0-23.0 42.50%

Overall average with 4 cores ( 115.5– 124.5) : 48.12%

X4 240 games http://www.mediafire.com/?vsdylcyuc6imczg

Global Score: 48.96%
Against Critter 1.6: 51.25% Deep Rybka 4: 56.87% Houdini 3.0: 38.75%


Not so brilliant this time against Houdini 3.0 Pro..

Regards,

Tom.
mcostalba
Posts: 2684
Joined: Sat Jun 14, 2008 9:17 pm

Re: Testing Stockfish 15-03-13. 480 Games.

Post by mcostalba »

Thanks Tom for running these tests vs strong engines. We don't perform such tests in our framework, so your info is very well complementary to ours and gives us interesting clues about how we are going with development.
Tomcass
Posts: 786
Joined: Sun Apr 16, 2006 9:09 pm

Re: Testing Stockfish 15-03-13. 480 Games.

Post by Tomcass »

Thanks to you, Marco, for the great work you and the team are doing to improve 'our' Stockfish.

Tom. :)
Tomcass
Posts: 786
Joined: Sun Apr 16, 2006 9:09 pm

Re: Testing Stockfish 15-03-13. 480 Games.

Post by Tomcass »

STOCKFISH 05-05-13: 480 GAMES

I7 980 3.33 Ghz.
6 real cores
Ponder: Off.
GUI: Fritz 12
Book: Perfect 2012
Time control: 4 min +0 sec/ game
No tablebases

201305_SF05052013_IPMAN-1 2013

SF050513 64px6 - Critter 1.6 64-bitx6_nob 20.5 - 19.5 +8/=25/-7 51.25%
SF050513 64px6 - Deep Rybka 4.1 SSE42 x64 (x6) 22.0 - 18.0 +13/=18/-9 55.00%
SF050513 64px6 - Houdini 3 Pro x64_6 17.5 - 22.5 +9/=17/-14 43.75%

http://www.mediafire.com/?mefi4dqmv5gmxxf

I7 980 3.33 Ghz.
6 real cores
Ponder: Off.
GUI: Fritz 12
Book: Perfect 2012
Time control: 2 min +2 sec/ game
No tablebases

201305_SF05052013_IPMAN_2+2 2013

SF050513 64px6 - Critter 1.6 64-bitx6_nob 23.5 - 16.5 +10/=27/-3 58.75%
SF050513 64px6 - Deep Rybka 4.1 SSE42 x64 (x6) 24.5 - 15.5 +14/=21/-5 61.25%
SF050513 64px6 - Houdini 3 Pro x64_6 13.5 - 26.5 +6/=15/-19 33.75%

Overall average with 6 cores: (121.5 – 118.5) = 50.62%

i7 975 3.33 Ghz.
4 real cores
GUI: Fritz 12
Book: Fritz 12
Time control: 4 min + 0 sec/ game
Ponder: Off
No tablebases

20130505_SF3_IPMAN_4+0-1 2013

SF050513 64p - Critter 1.6a 64-bitnob_4 21.5 - 18.5 +8/=27/-5 53.75%
SF050513 64p - Deep Rybka 4 x64 (x4) 24.5 - 15.5 +13/=23/-4 61.25%
SF050513 64p - Houdini 3 Pro x64_4 15.5 - 24.5 +3/=25/-12 38.75%

http://www.mediafire.com/?4n8jbckjq63adf6

i7 975 3.33 Ghz.
4 real cores
GUI: Fritz 12
Book: Fritz 12
Time control: 2 min + 2 sec/ game
Ponder: Off
No tablebases

20130505_SF3_2+2_IPMAN 2013

SF050513 64p - Critter 1.6a 64-bitnob_4 19.5 - 20.5 +6/=27/-7 48.75%
SF050513 64p - Deep Rybka 4 x64 (x4) 24.0 - 16.0 +10/=28/-2 60.00%
SF050513 64p - Houdini 3 Pro x64_4 16.5 - 23.5 +4/=25/-11 41.25%

http://www.mediafire.com/?76jjowfv7a753gc

Overall average with 4 cores: (121.5 – 118.5) = 50.62%

Global Score: 50.62%

Against Critter 1.6: 53.12% Deep Rybka 4: 59.38% Houdini 3.0: 39.37%


Regards from Barcelona,

Tom.
Tomcass
Posts: 786
Joined: Sun Apr 16, 2006 9:09 pm

Re: Testing Stockfish 15-03-13. 480 Games.

Post by Tomcass »

STOCKFISH 09-05-13: 480 GAMES

I7 980 3.33 Ghz.
6 real cores
Ponder: Off.
GUI: Fritz 12
Book: Perfect 2012
Time control: 4 min +0 sec/ game
No tablebases

SF090513 64x6 - Critter 1.6 64-bitx6_nob 12-17-11 20.5-19.5 51.25%
SF090513 64x6 - Deep Rybka 4.1 SSE42 x64 (x6) 13-20-7 23.0-17.0 57.50%
SF090513 64x6 - Houdini 3 Pro x64_6 5-22-13 16.0-24.0 40.00%

I7 980 3.33 Ghz.
6 real cores
Ponder: Off.
GUI: Fritz 12
Book: Perfect 2012
Time control: 2 min +2 sec/ game
No tablebases

SF090513 64x6 - Critter 1.6 64-bitx6_nob 6-26-8 19.0-21.0 47.50%
SF090513 64x6 - Deep Rybka 4.1 SSE42 x64 (x6) 7-23-10 18.5-21.5 46.25%
SF090513 64x6 - Houdini 3 Pro x64_6 1-26-13 14.0-26.0 35.00%

Overall average with 6 cores: 46.25%

X6 240 games = http://www.mediafire.com/?mdg6lb5md5hpxve

i7 975 3.33 Ghz.
4 real cores
GUI: Fritz 12
Book: Fritz 12
Time control: 4 min + 0 sec/ game
Ponder: Off
No tablebases

SF090513 64 - Critter 1.6a 64-bitnob_4 9-25-6 21.5-18.5 53.75%
SF090513 64 - Deep Rybka 4 x64 (x4 9-25-6 21.5-18.5 53.75%
SF090513 64 - Houdini 3 Pro x64_4 10-19-11 19.5-20.5 48.75%

i7 975 3.33 Ghz.
4 real cores
GUI: Fritz 12
Book: Fritz 12
Time control: 2 min + 2 sec/ game
Ponder: Off
No tablebases

SF090513 64 - Critter 1.6a 64-bitnob_4 11-23-6 22.5-17.5 56.25%
SF090513 64 - Deep Rybka 4 x64 (x4) 5-29-6 19.5-20.5 48.75%
SF090513 64 - Houdini 3 Pro x64_4 3-26-11 16.0-24.0 40.00%

X4 240 games http://www.mediafire.com/?zmazuqbcn9tvo6z

Overall average with 4 cores: 50.21%

Global Score: 48.23%

Against Critter 1.6: 52.19% Deep Rybka 4: 51.56% Houdini 3.0: 40.94%


Poor performance at incremental TC. Only 45.63%. The worst in two months. Statistical effect or wrong changes in the time management?.

Regards,

Tom.
Tomcass
Posts: 786
Joined: Sun Apr 16, 2006 9:09 pm

Re: Testing Stockfish 15-03-13. 480 Games.

Post by Tomcass »

STOCKFISH 11-05-13: 480 GAMES

I7 980 3.33 Ghz.
6 real cores
Ponder: Off.
GUI: Fritz 12
Book: Perfect 2012
Time control: 4 min +0 sec/ game
No tablebases

SF110513 64x6 - Critter 1.6 64-bitx6_nob 6-29-5 20.5-19.5 51.25%
SF110513 64x6 - Deep Rybka 4.1 SSE42 x64 (x6) 13-21-6 23.5-16.5 58.75%
SF110513 64x6 - Houdini 3 Pro x64_6 6-17-17 14.5-26.5 36.35%

I7 980 3.33 Ghz.
6 real cores
Ponder: Off.
GUI: Fritz 12
Book: Perfect 2012
Time control: 2 min +2 sec/ game
No tablebases

SF110513 64x6 - Critter 1.6 64-bitx6_nob 9-24-7 21.0-19.0 52.50%
SF110513 64x6 - Deep Rybka 4.1 SSE42 x64 (x6) 11-28-1 25.0-15.0 62.50%
SF110513 64x6 - Houdini 3 Pro x64_6 6-20-14 16.0-24.0 40.00%

240 games X6 = http://www.mediafire.com/?ib3cg53vxdfp81f

Overall average with 6 cores: (120.5 – 119.5) = 50.21%

i7 975 3.33 Ghz.
4 real cores
GUI: Fritz 12
Book: Fritz 12
Time control: 4 min + 0 sec/ game
Ponder: Off
No tablebases

SF110513 64 - Critter 1.6a 64-bitnob_4 4-30-6 19.0-21.0 47.50%
SF110513 64 - Deep Rybka 4 x64 (x4 10-27-3 23.5-16.5 58.75%
SF110513 64 - Houdini 3 Pro x64_4 8-18-14 17.0-23.0 42.50%

i7 975 3.33 Ghz.
4 real cores
GUI: Fritz 12
Book: Fritz 12
Time control: 2 min + 2 sec/ game
Ponder: Off
No tablebases

SF110513 64 - Critter 1.6a 64-bitnob_4 13-22-5 24.0-16.0 60.00%
SF110513 64 - Deep Rybka 4 x64 (x4 9-26-5 22.0-18.0 55.00%
SF110513 64 - Houdini 3 Pro x64_4 6-24-10 18.0-22.0 45.00%

240 games X4= http://www.mediafire.com/?1j53my36mxqqxkf

Overall average with 4 cores: (123.5-116.5) = 51.46%

Global Score: 50.84%

Against Critter 1.6: 52.81% Deep Rybka 4: 59.06% Houdini 3.0: 40.94%


In my latest test I wondered if something wrong happened with the time management at incremental T.C. or if it was simply an statistical effect. This test seems to confirm that there is nothing wrong with incremental T.C., since there is no functional change between 090513 and 110513 versions.

Regards,

Tom.
gladius
Posts: 568
Joined: Tue Dec 12, 2006 10:10 am
Full name: Gary Linscott

Re: Testing Stockfish 15-03-13. 480 Games.

Post by gladius »

Tomcass wrote:STOCKFISH 11-05-13: 480 GAMES
Thanks Tom, it's really nice to see your matches, and watch the progress :). There were no changes in the TC management recently, so that must have been statistical noise.

I'm excited to see the next results, after Joona's new counter move patch.
Tomcass
Posts: 786
Joined: Sun Apr 16, 2006 9:09 pm

Re: Testing Stockfish 15-03-13. 480 Games.

Post by Tomcass »

Thanks Tom, it's really nice to see your matches, and watch the progress :). There were no changes in the TC management recently, so that must have been statistical noise.

I'm excited to see the next results, after Joona's new counter move patch.



Thanks for following my tests, Gary. I have started testing 160513 version, with Joona's improvements. Not a bad start. Let's see. :wink:



STOCKFISH 15-05-1 480 GAMES

I7 980 3.33 Ghz.
6 real cores
Ponder: Off.
GUI: Fritz 12
Book: Perfect 2012
Time control: 4 min +0 sec/ game
No tablebases

SF150513 64x6 - Critter 1.6 64-bitx6_nob 8-22-10 19.0-21.0 47.50%
SF150513 64x6 - Deep Rybka 4.1 SSE42 x64 (x6) 11-23-6 22.5-17.5 56.25%
SF150513 64x6 - Houdini 3 Pro x64_ 6-21-13 16.5-23.5 41.25%

I7 980 3.33 Ghz.
6 real cores
Ponder: Off.
GUI: Fritz 12
Book: Perfect 2012
Time control: 2 min +2 sec/ game
No tablebases

SF150513 64x6 - Critter 1.6 64-bitx6_nob 8-27-5 21.5-18.5 53.75%
SF150513 64x6 - Deep Rybka 4.1 SSE42 x64 (x6) 11-21-8 21.5-18.5 53.75%
SF150513 64x6 - Houdini 3 Pro x64_6 3-21-16 13.5-26.5 33.75%

240 games X6 = http://www.mediafire.com/?btb0ab3r2smpiii

Overall average with 6 cores: (114.5-125.5) = 47.71%

i7 975 3.33 Ghz.
4 real cores
GUI: Fritz 12
Book: Fritz 12
Time control: 4 min + 0 sec/ game
Ponder: Off
No tablebases

SF150513 64 - Critter 1.6a 64-bitnob_4 10-26-4 23.0-17.0 57.50%
SF150513 64 - Deep Rybka 4 x64 (x4 11-24-5 23.0-17.0 57.50%
SF150513 64 - Houdini 3 Pro x64_4 6-22-12 17.0-23.0 42.50%

i7 975 3.33 Ghz.

4 real cores
GUI: Fritz 12
Book: Fritz 12
Time control: 2 min + 2 sec/ game
Ponder: Off
No tablebases

SF150513 64 - Critter 1.6a 64-bitnob_4 6-28-6 20.0-20.0 50.00%
SF150513 64 - Deep Rybka 4 x64 (x4 11-23-6 22.5-17.5 56.25%
SF150513 64 - Houdini 3 Pro x64_4 6-27-7 19.5-20.5 48.75%

240 games X 4 = http://www.mediafire.com/?g8anqkjn6w83xc3

Overall average with 4 cores: (125.0-115.0) = 52.08%

Global Score: 49.89%

Against Critter 1.6: 52.19% Deep Rybka 4: 55.94% Houdini 3.0: 41.56%


Regards,

Tom.