CEGT - rating lists April 08th 2012

Discussion of computer chess matches and engine tournaments.

Moderators: hgm, Harvey Williamson, bob

User avatar
Werner
Posts: 2224
Joined: Wed Mar 08, 2006 9:09 pm

CEGT - rating lists April 08th 2012

Post by Werner » Sun Apr 08, 2012 4:05 pm

Hi all, :D

our actual rating lists are online and can be found under the attached links.

40 / 20:
New games: 1099 ; 27 different engines
Total: 583.226

NEW Engines
55 Stockfish 2.2.2 w32 1CPU: 2934 - 350 games (startrating is close to version 2.01)
816 DoubleCheck 3.1 x64: 2365 - 443 games (new here and about 10 points behind the blitz-result with version 3.0)

UPDATES
...not much differences:

386 Texel 1.01 x64: 2650 - 1174 games (+1)
646 RedQueen 1.1.1 x64 1CPU: 2472 - 656 games (some more games played with not so good results - name will be corrected later)

40 / 4:
New games: 5456
All games now: 1.001.760!!! - a jublilee!!!

New Engines
351 Texel 1.01 x64 1CPU: 2638 - 1000 games (much better than CuckooChess here too of course)
425 Texel 1.01 w32 1CPU: 2599 - 1300 games (39 points under 64bit-version)
883 Cheese 1.4 x64: 2317 - 1000 games (+90 to version 1.3b)
793 DoubleCheck 3.0 x64: 2377 - 1000 games (+160 to version 2.4)

Updates
51 Stockfish 2.2.2 w32 1CPU: 2923 - 1500 games (-3)
833 Zarkov 5 (MCS): 2352 - 1000 games (+5)
1048 LittleThought 1.04 w32: 2163 - 1000 games (-1)

40/120
See here our new single-list (with 7800 games):
http://www.husvankempen.de/nunn//40120n ... liste.html
New is Zappa Mexico II x64 with 2721/150 games.
Stockfish 2.2.2 x64 lost 8 points and Deep Junior 13 x64 lost 16:
Naum 4.2 x64: +4

40/20 pb=on
Our new list raises fast. We now have made 4000 games.
New is Chiron 1.1a x64 with 2822/520 games.

http://www.husvankempen.de/nunn/rating4020PBON.htm
A big „Thank you“ to all testers as usual!!

Links

40/20: http://www.husvankempen.de/nunn/rating.htm
Blitz: http://www.husvankempen.de/nunn/blitz.htm
40/120: http://www.husvankempen.de/nunn/rating120.htm
Tester: http://www.husvankempen.de/nunn/testers/testers.htm
40/20 pb=on: http://www.husvankempen.de/nunn/rating4020PBON.htm
Games of the week: http://www.husvankempen.de/nunn/40_40%2 ... on/gow.jpg

Werner Schuele
CEGT-Team

lkaufman
Posts: 3082
Joined: Sun Jan 10, 2010 5:15 am
Location: Maryland USA
Contact:

Re: CEGT - rating lists April 08th 2012

Post by lkaufman » Tue Apr 10, 2012 7:06 pm

One question regarding your two 40/20 lists (regular and pb on): I would like to know whether the komodo ratings would differ much if you separated games played with sse4 from those played without. We appear to be about 30 elo away from Houdini (talking 1 core) on both lists (averaging Houdini 1.5 and 2.0 since 1.5 is higher), but perhaps the gap is narrower if only sse4 games are considered. Since I estimate we have gained 25 elo since K4, I'm wondering if we might have finally caught or passed Houdini on sse4 machines.

User avatar
gleperlier
Posts: 1033
Joined: Sat Feb 04, 2012 9:03 pm

Re: CEGT - rating lists April 08th 2012

Post by gleperlier » Tue Apr 10, 2012 7:39 pm

Thanks !

Gab

User avatar
Houdini
Posts: 1471
Joined: Mon Mar 15, 2010 11:00 pm
Contact:

Re: CEGT - rating lists April 08th 2012

Post by Houdini » Tue Apr 10, 2012 7:53 pm

lkaufman wrote:One question regarding your two 40/20 lists (regular and pb on): I would like to know whether the komodo ratings would differ much if you separated games played with sse4 from those played without. We appear to be about 30 elo away from Houdini (talking 1 core) on both lists (averaging Houdini 1.5 and 2.0 since 1.5 is higher), but perhaps the gap is narrower if only sse4 games are considered. Since I estimate we have gained 25 elo since K4, I'm wondering if we might have finally caught or passed Houdini on sse4 machines.
Larry, you really should stop worrying about catching Houdini 1.5 or 2.
In case you don't follow the Houdini Facebook page, our current Houdini development version is about 60 Elo stronger than Houdini 1.5 - with an increasing trend at longer longer time controls.

Robert

beram
Posts: 1187
Joined: Wed Jan 06, 2010 2:11 pm

Re: CEGT - rating lists April 08th 2012

Post by beram » Tue Apr 10, 2012 8:02 pm

Houdini wrote:
lkaufman wrote:One question regarding your two 40/20 lists (regular and pb on): I would like to know whether the komodo ratings would differ much if you separated games played with sse4 from those played without. We appear to be about 30 elo away from Houdini (talking 1 core) on both lists (averaging Houdini 1.5 and 2.0 since 1.5 is higher), but perhaps the gap is narrower if only sse4 games are considered. Since I estimate we have gained 25 elo since K4, I'm wondering if we might have finally caught or passed Houdini on sse4 machines.

Larry, you really should stop worrying about catching Houdini 1.5 or 2.
In case you don't follow the Houdini Facebook page, our current Houdini development version is about 60 Elo stronger than Houdini 1.5 - with an increasing trend at longer longer time controls.

Robert
:lol: .......... :( .......... :lol:

User avatar
gleperlier
Posts: 1033
Joined: Sat Feb 04, 2012 9:03 pm

Re: CEGT - rating lists April 08th 2012

Post by gleperlier » Tue Apr 10, 2012 8:05 pm

Please Robert, do not launch an Houdini stronger than 2.0c version... not now ! :D

All engine chess rooms are full of Houdini everywhere :cry:

Please think about diversity :lol:

Kidding :roll:

lkaufman
Posts: 3082
Joined: Sun Jan 10, 2010 5:15 am
Location: Maryland USA
Contact:

Re: CEGT - rating lists April 08th 2012

Post by lkaufman » Tue Apr 10, 2012 8:13 pm

Well, you were clearly wrong in claiming that H2 was stronger than H1.5 (except in blitz or using multi-cores), so maybe you will again be proven wrong. Your claim of +60 over H1.5 and +40 over H2 is already suspicious, unless you are talking about blitz, since (40 + 0 = 60) is false.

In any case, I think it is desirable for any commercial program to be stronger than any free one, so passing h1.5 (and thereby also h2 in case you make that free eventually) is important for all commercial programs regardless of whether a later commercial engine is stronger than others.

User avatar
Houdini
Posts: 1471
Joined: Mon Mar 15, 2010 11:00 pm
Contact:

Re: CEGT - rating lists April 08th 2012

Post by Houdini » Tue Apr 10, 2012 9:03 pm

lkaufman wrote:Well, you were clearly wrong in claiming that H2 was stronger than H1.5 (except in blitz or using multi-cores), so maybe you will again be proven wrong. Your claim of +60 over H1.5 and +40 over H2 is already suspicious, unless you are talking about blitz, since (40 + 0 = 60) is false.
It will be my pleasure to ease your suspicious mind at the time of the Houdini 3 release. ;)

beram
Posts: 1187
Joined: Wed Jan 06, 2010 2:11 pm

Re: CEGT - rating lists April 08th 2012

Post by beram » Tue Apr 10, 2012 9:56 pm

lkaufman wrote:Well, you were clearly wrong in claiming that H2 was stronger than H1.5 (except in blitz or using multi-cores), so maybe you will again be proven wrong. Your claim of +60 over H1.5 and +40 over H2 is already suspicious, unless you are talking about blitz, since (40 + 0 = 60) is false.

In any case, I think it is desirable for any commercial program to be stronger than any free one, so passing h1.5 (and thereby also h2 in case you make that free eventually) is important for all commercial programs regardless of whether a later commercial engine is stronger than others.


Well, SedatChess, Frank Quisinksky gave +20 and Vincent Lejeune list give +9

And btw Komodo 4 above 3 on Vincent Lejeune list = + 11 where you claimed to have +25 before release

Code: Select all

http://sedat-chess.110mb.com/index.php?p=1_8
+20

 Rank     Name                   Elo    +    -  games score oppo. draws 

  01 Houdini 2.0 Pro x64 4c     3358   25   24   430   63%  3272   47% 
  02 Houdini 1.5a x64 4c        3338   17   17   863   62%  3264   49%


http://www.amateurschach.de/main/_swcr-all.txt
+20
           NAME / version of engine       ELO    +    -   GAM    SC   OP     DR
  ------------------------------------------------------------------------------
   1     1 Houdini 2.0c x64               3019   18   18  1400   79%  2777   28% 
   -     2 Houdini 1.5 x64                2999   14   14  2320   78%  2772   29% 


http://home.scarlet.be/vincentlejeune/rating20120402.txt
+9
   2 Houdini 2.0 64-bit 4CPU                3303   12   12  2833   76%  3102   32% 
   3 Houdini 1.5 64-bit 4CPU                3294   10   10  3386   72%  3140   37% 

lkaufman
Posts: 3082
Joined: Sun Jan 10, 2010 5:15 am
Location: Maryland USA
Contact:

Re: CEGT - rating lists April 08th 2012

Post by lkaufman » Tue Apr 10, 2012 10:17 pm

beram wrote:
lkaufman wrote:Well, you were clearly wrong in claiming that H2 was stronger than H1.5 (except in blitz or using multi-cores), so maybe you will again be proven wrong. Your claim of +60 over H1.5 and +40 over H2 is already suspicious, unless you are talking about blitz, since (40 + 0 = 60) is false.

In any case, I think it is desirable for any commercial program to be stronger than any free one, so passing h1.5 (and thereby also h2 in case you make that free eventually) is important for all commercial programs regardless of whether a later commercial engine is stronger than others.


Well, SedatChess, Frank Quisinksky gave +20 and Vincent Lejeune list give +9

And btw Komodo 4 above 3 on Vincent Lejeune list = + 11 where you claimed to have +25 before release

Code: Select all

http://sedat-chess.110mb.com/index.php?p=1_8
+20

 Rank     Name                   Elo    +    -  games score oppo. draws 

  01 Houdini 2.0 Pro x64 4c     3358   25   24   430   63%  3272   47% 
  02 Houdini 1.5a x64 4c        3338   17   17   863   62%  3264   49%


http://www.amateurschach.de/main/_swcr-all.txt
+20
           NAME / version of engine       ELO    +    -   GAM    SC   OP     DR
  ------------------------------------------------------------------------------
   1     1 Houdini 2.0c x64               3019   18   18  1400   79%  2777   28% 
   -     2 Houdini 1.5 x64                2999   14   14  2320   78%  2772   29% 


http://home.scarlet.be/vincentlejeune/rating20120402.txt
+9
   2 Houdini 2.0 64-bit 4CPU                3303   12   12  2833   76%  3102   32% 
   3 Houdini 1.5 64-bit 4CPU                3294   10   10  3386   72%  3140   37% 
I acknowledge that Houdini 2 had improved MP scaling over 1.5. I'm talking about the single core performance. The combined results other than blitz show Houdini 2 to be below 1.5 (marginally).

Post Reply