The grand unified rating list April 2012

[Vinvin]

Time control is about 30 sec to 2 min per moves on an AMD 2 Ghz.

Games from CEGT, CCRL, WBEC, SWCR and Chess War.
I removed IPON because TC is too fast.

1'307'909 games.

Don't forget to look at the error bar (+ and -) !

http://home.scarlet.be/vincentlejeune/r ... 120402.txt

- If there's no "n-CPU" info, it's 1 CPU.
- If there's no "n-bit" info, it's 32 bit.

[IWB]

Thx Vincent,

Regards
Ingo

[Jouni]

Hmm: Critter 1.2 is the strongest Critter Richard back to drawing board?

[gleperlier]

[quote="Vinvin"]Time control is about 30 sec to 2 min per moves on an AMD 2 Ghz.

Games from CEGT, CCRL, WBEC, SWCR and Chess War.
I removed IPON because TC is too fast.

1'307'909 games.

Don't forget to look at the error bar (+ and -) !

http://home.scarlet.be/vincentlejeune/r ... 120402.txt

- If there's no "n-CPU" info, it's 1 CPU.
- If there's no "n-bit" info, it's 32 bit.[/quote

Hi,

Can you make a kind of Pure List too with only let say 1 or maximum 2 version of the same engine maximum ?

Best regards,

Gab

[lkaufman]

Thanks for your work on this, it's nice to have a rating list that combines all the results at longer time controls to get big enough sample sizes. I would suggest you specify a minimum number of games to appear on the list; I notice you have Strelka 5.1 on the list with just 14 games, which is rather silly. The list confirms that Houdini 2 is not an improvement over 1.5 except perhaps on 4 cores. I was much more surprised to see Critter 1.4 below 1.2 on 2 and 4 cores, though higher on 1. In general it seems progress at the top has been slow recently.

[beachknight]

great work, vincent. Thx.

[Terry McCracken]

Time control is about 30 sec to 2 min per moves on an AMD 2 Ghz.

Games from CEGT, CCRL, WBEC, SWCR and Chess War.
I removed IPON because TC is too fast.

1'307'909 games.

Don't forget to look at the error bar (+ and -) !

http://home.scarlet.be/vincentlejeune/r ... 120402.txt

- If there's no "n-CPU" info, it's 1 CPU.
- If there's no "n-bit" info, it's 32 bit.

Well done, thanks!

[beram]

Thanks for your work on this, it's nice to have a rating list that combines all the results at longer time controls to get big enough sample sizes. ...
The list confirms that Houdini 2 is not an improvement over 1.5 except perhaps on 4 cores. I was much more surprised to see Critter 1.4 below 1.2 on 2 and 4 cores, though higher on 1. In general it seems progress at the top has been slow recently.

and the list confirms that Komodo 4 (3189) is not such a big improvement over 3 (3178), +11 ....

[Vinvin]

Thanks for your work on this, it's nice to have a rating list that combines all the results at longer time controls to get big enough sample sizes. I would suggest you specify a minimum number of games to appear on the list; I notice you have Strelka 5.1 on the list with just 14 games, which is rather silly. The list confirms that Houdini 2 is not an improvement over 1.5 except perhaps on 4 cores. I was much more surprised to see Critter 1.4 below 1.2 on 2 and 4 cores, though higher on 1. In general it seems progress at the top has been slow recently.

Only players with 100% and 0% where remove (with ordoprep).

Strelka 5.1 have only 14 games but error bar is set in function. Strelka played 21 games now and drop 2% score (about -14 Elo).

[Vinvin]

Time control is about 30 sec to 2 min per moves on an AMD 2 Ghz.

Games from CEGT, CCRL, WBEC, SWCR and Chess War.
I removed IPON because TC is too fast.

1'307'909 games.

Don't forget to look at the error bar (+ and -) !

http://home.scarlet.be/vincentlejeune/r ... 120402.txt

- If there's no "n-CPU" info, it's 1 CPU.
- If there's no "n-bit" info, it's 32 bit.

Little corrections, mainly "_2CPU" -> "2 CPU"
+ latest games from Strelka 5.1 merged
" 20 Strelka 5.1 3214 87 87 35 64% 3137 60% "

List is here : http://home.scarlet.be/vincentlejeune/r ... 120403.txt