CEGT - rating lists November 4th 2007

[Werner]

Hi all ,

our new 40/4, 40/20 and 40/120 rating lists are online and can be found under the attached links.

40 / 20:
There are a lot of new engines this week, especially with the 11th series: Chessmaster 11 is out! We´ve got also 2 new free engines: Scorpio 2.0 and Crafty 21.6. We too made more games with Shredder 11 (2CPU and 1CPU), Zappa Mexico w32 1CPU and Colossus 2006e. See more in our list "Games of the week".

In total our 40/20 list is based on 193.761 games!

New engines:
Very new is TheKing350-engine - and not easy to test. We have made 123 games with the 32bit 1CPU-engine and got the first rating with 2694 elos. Hope we can say more next week. I will try to find a way for testing the 2CPU version because I am very interested in the result .
Only start ratings with much to less games we too have with Crafty 21.6 2CPU (could only be tested on a quad as it uses always 1CPU when engine is idle), with Scorpio 2.0 2CPU and Zappa Mexico w32 1CPU. The start rating of Shredder 11 32bit 1CPU is very good after 372 games: 2910 elos which is exactly 100 points more than Shredder 10 has in our list.

updated engines:
DeepShredder x64 2CPU lost 17 points and has now 2944 elos. DeepFritz 10.1 4CPU played another 100 games and has now 2946 points.

40 / 120:
Our 40/120 list was updated too. As we now make matches with 4 CPUs the number of new games is not so high . Interim results (DeepShredder 11 4CPUs included) we publish in our forum:
http://husvankempen.de/nunn/phpBB2/viewtopic.php?t=868
Here you can see the results of the first finished matches. Hopefully we can make a separate list in the next weeks.

40 / 4:
Our blitz list is actuell too:
+90 Elo for DS11 x64 2CPU, +101 for DS11 x64 1 CPU!!!!!!

19 Deep Shredder 11 x64 2CPU : 2946 16 16 1240 57.3 % 2895 34.9 %
47 Deep Shredder 10 x64 2CPU : 2856 13 13 2050 48.1 % 2869 30.5 %

28 Deep Shredder 11 x64 1CPU : 2903 15 15 1400 58.4 % 2844 33.4 %
81 Deep Shredder 10 x64 1CPU : 2802 19 19 840 41.7 % 2860 34.8 %


A big „Thank you“ to all testers as usual!

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

[rdan1987]

Thx for the update ....but why don't you simply use the name "Shredder 11" instead of "Deep Shredder 11 w32 1CPU" ?

[Werner]

Thx for the update ....but why don't you simply use the name "Shredder 11" instead of "Deep Shredder 11 w32 1CPU" ?

Hi, this is only to clarify - we do not use the single version of Shredder as we only have DeepShredder engines. Stefan told us there is no difference to the single version when we use DeepShredder with 1CPU.

[Norm Pollock]

Werner,

Could you clarify what "40/120" in the Event tag actually means?

Does it mean, for example, 40 moves in 120 minutes repeating?

Or does it mean, 40 moves in 120 minutes, followed by 20 moves in 60 minutes, followed by 30 minutes?

Or something else?

Thanks,
Norm

[Heinz Van Kempen]

Hi Norm ,

I am organizing the 40/120 list (adding a quad list in a few weeks).

In CEGT all time controls are adapted to Crafty benchmark like in CCRL also. The repid chess lists from both testing groups use repeated time controls, giving less time on faster machines than the reference one.

In case of 40/120 we give 40 moves in 120 plus 20 moves in 60 plus 30 minutes for the rest on AMD64 x2 4200+. It is necessary to restrict the third time control, because a constantly repeated time control would lead to games lasting 16 hours and more (for the many we have with 150 moves and more). Not doing this leads to gaining relatively few games.
But even this way the differences in rating to the rapid chess lists with repeated time controls are minimal. A good time management can lead to very few ELO points difference afer 1000 games are played,

Most of the games are decided after 80 moves when engines start to play fast, so we can avoid a waste of time with dead drawn positions.

[Norm Pollock]

Hi Heinz,

I understand your reasons for doing what you do for the 40/120 games. But I think that the label "40/120" is therefore misleading because it gives the impression that 40/120 is repeating, which is not the case. An accurate label would be "40/120:20/60:30".

Why not use a "TimeControl" tag? That would avoid any misinterpretation caused by using "40/120" in the "Event" tag. Based on pgn standards, I believe it would be as follows:

[TimeControl "40/7200:20/3600:1800"]

where time is expressed in seconds.
-Norm

[Heinz Van Kempen]

Hi Norm ,

ok we can discuss this if we give it that way.

Will show things more clearly.

[Tony Thomas]

40/120, and 20/60 are pretty much the same thing, only thing that throws it off is the rest in 30 minute time control.