CEGT Marathon 40/400 repeated

[Heinz Van Kempen]

Hi all ,

over the past months I used one of my quad machines to run a new marathon match 40/400 repeated (real time control):

Code: Select all

CEGT Quad Extreme 40/400 repeated  2008

1   Rybka 2.3.2a X64 4CPU  11½½1½1½½½1½½½½½½½0½1½½½½½½½½½½1½½½½0½½½½½½1½1½½½½ 28.5/50
2   Naum 3.1 X64 4CPU      00½½0½0½½½0½½½½½½½1½0½½½½½½½½½½0½½½½1½½½½½½0½0½½½½ 21.5/50

I also extended the previous match to 50 games for better comparison:

Code: Select all

CEGT Quad Extreme 40/400 repeated  2008

                           12345678901234567890123456789012345678901234567890
1   Rybka 2.3.2a X64 4CPU  1½½½½½½1½½½½½1½½½½½½½½½1½½½½½101½½½½0½½½1½11½½1011 29.5/50
2   Naum 3 X64 4CPU        0½½½½½½0½½½½½0½½½½½½½½½0½½½½½010½½½½1½½½0½00½½0100 20.5/50

The matches for Rybka 2.3.2a X64 4CPU against Zappa Mexico II X64 and Zappa Mexico I upd. X64 with currently 42 games each will be expanded also to 50 games over the next two weeks.

All matches CEGT 40/400 can be replayed here:

http://www.husvankempen.de/nunn/Replay/cegtextreme1.htm

A collection of the most interesting games (in my view) can be replayed here:

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

The download for the complete 40/400 games is available here:

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

Best Regards
Heinz

[Norm Pollock]

The problem of "duplicate games" is a problem that exists in all rating databases. Should a game count more than one time when computing the ratings? Even in this very small database there is a pair of duplicate games.

Code: Select all

[Event "CEGT Quad Extreme 40/400 repeated"]
[Site "Engine Match"]
[Date "2008.02.28"]
[Round "5"]
[White "Naum 3 X64 4CPU"]
[Black "Rybka 2.3.2a X64 4CPU"]
[Result "1/2-1/2"]
[ECO "B97"]
[TimeControl "40/24000:0/0:0/0"]

1. e4 c5 2. Nf3 d6 3. d4 cxd4 4. Nxd4 Nf6 5. Nc3 a6 6. Bg5 e6 7. f4 Qb6 8. 
Qd2 Qxb2 9. Rb1 Qa3 10. f5 Nc6 11. fxe6 fxe6 12. Nxc6 bxc6 13. Be2 Be7 14. 
O-O O-O 15. Rb3 Qc5+ 16. Be3 Qe5 17. Bd4 Qa5 18. Bb6 Qe5 19. Bd4 Qa5 20. 
Bb6 Qe5 21. Bd4 1/2-1/2 

[Event "CEGT Quad Extreme 40/400 repeated"]
[Site "Engine Match"]
[Date "2008.05.21"]
[Round "9"]
[White "Naum 3.1 X64 4CPU"]
[Black "Rybka 2.3.2a X64 4CPU"]
[Result "1/2-1/2"]

1. e4 c5 2. Nf3 d6 3. d4 cxd4 4. Nxd4 Nf6 5. Nc3 a6 6. Bg5 e6 7. f4 Qb6 8. 
Qd2 Qxb2 9. Rb1 Qa3 10. f5 Nc6 11. fxe6 fxe6 12. Nxc6 bxc6 13. Be2 Be7 14. 
O-O O-O 15. Rb3 Qc5+ 16. Be3 Qe5 17. Bd4 Qa5 18. Bb6 Qe5 19. Bd4 Qa5 20. 
Bb6 Qe5 21. Bd4 1/2-1/2 

[Heinz Van Kempen]

Hi Norm ,

in all CEGT databases used for rating lists the duplicate games are deleted.

For CEGT 40/400 there will never be a rating list. The reason is simply that it makes no sense to have one here, because a single game averagely lasts around 24 hours and it is not possible to collect enough games for a meaningful rating list.

The only reason for running this matches if to offer something including the best three engines available with extreme long time controls for those interested in games of superb quality.

Currently I used more or less the same sets of openings for the matches involving Naum 3 and Naum 3.1 to see the differences. So it is not surprising that for two exceptionally very short games a "duplicate" is produced. Does not matter in any way.

[Norm Pollock]

Hi Norm ,

in all CEGT databases used for rating lists the duplicate games are deleted.

For CEGT 40/400 there will never be a rating list. The reason is simply that it makes no sense to have one here, because a single game averagely lasts around 24 hours and it is not possible to collect enough games for a meaningful rating list.

The only reason for running this matches if to offer something including the best three engines available with extreme long time controls for those interested in games of superb quality.

Currently I used more or less the same sets of openings for the matches involving Naum 3 and Naum 3.1 to see the differences. So it is not surprising that for two exceptionally very short games a "duplicate" is produced. Does not matter in any way.

Hi Heinz,

I'm glad to know that duplicates are not used in the CEGT rating list process. But I think the first occurrence of the duplicate should be used.

Duplicate games are "unavoidable" in computer chess. They often happen when the opening book repeats the same opening to the same engines. They also occur in grandmaster chess as well.

In the case of GMs, you have to wonder if the duplicate was pre-arranged so as to make it look like they played a "hard" game, when in fact they were playing someone else's game from memory.

-Norm