Naum 4 x64 4CPU in CEGT 40/120 repeated

[Heinz Van Kempen]

Hi all ,

here I give a copy of what I posted in the CEGT Forum for those interested in games with long time controls

either my first results with Naum x64 4CPU are overly good or the gap between Rybka 3 and Naum 4 will be much closer compared to CEGT 40/20 or especially CEGT 40/4, like I would expect, because the gap between Rybka 3 and it´s predesessor and the competitors was narrower previously, too, compared to faster time controls.


Conditions:

Core 2 Quad E6600 (5 men) - 4 GB RAM
1024 MB Hash
5 men EGTB
time control 40/120 repeated real time control (not adapted to weaker hardware)
Openings currently played in Super GM practice restricted to 8 moves, all games with colors reversed also

Code: Select all

CEGT 40/120  

                      1234567890
1   Rybka 3 x64 4CPU  ½½½½1½½½½½  5.5/10
2   Naum 4 x64 4CPU   ½½½½0½½½½½  4.5/10

Although the games were hard fought and mostly interesting nine draws out of ten games. This usully happens with long time controls when difference in strength is low (unlike it was previosly for Rybka 3 against top competitors), but at this stage I do not claim that it will continue this way. Rybka 3 did not have problems with too many draws before like you can see from the individual stats of the CEGT 40/120 quad rating list from last week

http://www.husvankempen.de/nunn/40_120_ ... uad/1.html

I have the same tendency in CEGT 40/400 repeated where I am also running matches Rybka 3 x64 4CPU and Rybka 3 Human x64 4CPU vs. Naum 4 x64 4CPU since last Thursday like this ones. The first games so far all ended in draws.

Code: Select all

CEGT 40/120 

1   Naum 4 x64 4CPU        ½½1½½½1½0½½½  6.5/12
2   Rybka 2.3.2a x64 4CPU  ½½0½½½0½1½½½  5.5/12

This match will be interesting, because this two engine versions will probably compete for rank 2 in the quad list.

Code: Select all

CEGT 40/120  

1   Deep Fritz 11 4CPU  0½½1½½½½  4.0/8  16.00
2   Naum 4 x64 4CPU     1½½0½½½½  4.0/8  16.00

Although being quite strong I guess that Deep Fritz 11 4CPU will not be able to keep the balance in this match in the long run.

Code: Select all

CEGT 40/120  

1   Naum 4 x64 4CPU        1½1½  3.0/4
2   Zappa Mexico x64 4CPU  0½0½  1.0/4

The matches will finally consist of 150 to 200 games each. Others against Hiarcs 11 and DS will follow.

33 games quad games with tournament time control so far are included in the new download here (link 3 and 4 from above):

http://www.husvankempen.de/nunn/Replay/ ... d40120.htm

Naum will be included in the rating list after first 200 or 300 games in a few weeks.

Best Regards
Heinz