CEGT - rating lists December 13th 2009

[Werner]

Hi all ,
our updated rating lists are now online and can be found under the attached links.

40 / 20:
Today we added 1677 games made with 71 different engines to our list. See more in our list "Games of the week". In total our 40/20 list is based now on 371.695 games.

New engines:

Deep Junior 2010 x64 4CPU has had a good start here too with 2950 elos in 150 games. This is around 100 points in front of version 10.1 in our list. With 2CPU we got a start-rating of 2892/137 games now (around +80 to verison 10.1)
Fruit 090705 x64 2CPU has a start-rating of 2938/186 games - but need more games for a sure rating.
Fruit 090705 x64 1CPU reached 2919 elos in 550 games. This is 30 points over version 2.3.5m 32 bit. Both engines are still private!
Pupsi 2 v.0.08 has a start rating of 2458/ 200 games which is a bit better than version v. 0.07

Updated engines:

Code: Select all

7 Deep Shredder 12 x64 4CPU 3067 +11 -11 2282 games (-2)
26 Spark 0.3MP x64 4CPU 2997 +27 -27 358 games (-2)
84 Thinker 5.4Di x64 1CPU 2921 +38 -38 185 games (+41)
127 Doch 09.980 x64 2890 +17 -17 946 games (-2)

Blitz Update

We will make an update on Tuesday.

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
Games of the week: http://www.husvankempen.de/nunn/40_40%2 ... on/gow.JPG
Elo-comparison: http://www.husvankempen.de/nunn/Replay/ ... arison.htm

Werner
CEGT-Team

[Vinvin]

There is a little error on the page : when I click "best single version 40/40", I get a 4 cpu : Stockfish 1.5.1 x64 4CPU

Thanks !

Vincent

[beachknight]

Both new Fruits and mp version of Spark are private test versions,
I suppose. I would like getting 2cpu and 4cpu versions of them
in order to include in my tournaments. But I think I should wait for
a public release.

Thinker 5.4Di = Thinker 5.4D inert

Hmmm, I should use same taxonomy.

Best,

[Jouni]

This is interesting in 40/20 list:

5 Naum 4 w32 1CPU 2984 17 17 881
6 Naum 4 x64 1CPU 2979 18 18 728

Even after 700 games 32b version is ahead!?

Jouni

[ernest]

Even after 700 games 32b version is ahead!?

Please do a little statistics thinking, Jouni!
On analysis of the start-position (1-cpu for reproducibility), it can be seen that the 64b version is 7% faster than the 32b.
So this means at most +5 to 7 Elo in favor of the 64b

And in the list, it says 2979-2984 = -5 Elo for the 64b,
the 95% probability deviation being 17 or 18 Elo

So you find 10 Elo strange?...