KO Tournament 16 Strong UCI engines

[Daniel Charles]

KO Tournament
Xeon 5365 (8*3 Gh)
Time 4 + 2 S ( >360s)
Book Perfect9.ctg
Each Match (20 games)
winners move to next round




Zappa Mexico vs List 5.12 : Zappa is already qualified for the next round

The big surprise for the First round is Alaric who is winning vs Spike 1.2... Match is not finish yet and Spike still come back

Colossus Improve a lot in 1 year , however it is not enought vs Strelka ..

rybka 1.0 is taking the best on Scorpio 4 CPU

Fruit 2.3.3 J beta is almost qualified for the next round

Toga is simply too strong for slowchess

This is a very good fight between Glaurung (4 CPU) and Deep Pharaon

Arasan10 (8 cpu) did what it could versus Naum SMP (8cpu), Naum is almost qualified too

[Daniel Charles]

Top 16 :

The only surprise was Alaric victory over Spike 11-9
As expected Fruit,Toga, Glaurung,Naum SMP,Zappa Mexico, Strelka and Free Rybka 1.0 won their match


Top 8 :

Rybka- Strelka : It was war between Rybka & Strelka ... Rybka 1.0 was victorious with a very small margin 10.5 - 9.5

Zappa Mexico- Alaric : No problem for Zappa Mexico with 15-5

Naum 2.2 Vs Glaurung smp : Naum smp 8 cpu had a lot of difficulties to beat Glaurung SMP 4 cpu ... It was a surprise to see that Glaurung did very well vs a top 5 commercial Chess engine . Glaurung did very well , a lot better than I expected .

Toga - Fruit : Toga won the match between the twin brothers . It was very closed . remember this is Toga Beta version.


Top 4 (still running)

Zappa Mexico - Rybka 1.0 : Zappa Mexico is already qualified with 10.5 - 2.5

[GS]

Hi Daniel,

I have an useful hint for you to save you some work, webspace
and also bandwidth on the server here(Moreover they will look better
and everyone can copy them too).
You can export any kind of table from a CB Gui as pure ascii(text).
Just choose save to clipboard and check filling up with spaces in the
options then. (It might look weird if like to import it into your text editor
e.g. Notepad, but if you import it here and put it here as code it will look
perfectly formatted.)
Here is an example for another kind of special tables:

Code: Select all

RWBC Qualification  2002

                  -------------------  Bace045           7.5/10
                  |   ---------------  SDBC360           7.5/10
                  |   |   -----------  Pentagon12        7.0/10
                  |   |   |   -------  Nero51            3.5/10
                  |   |   |   |   ---  Raffaela005       3.5/10
                  |   |   |   |   |   
 1 Frenzee114     ½1  1½  01  11  11     8.0/10
 2 Greko_038      ½0  10  11  ½0  1½     5.5/10
 3 Lightning201   0½  00  00  11  1½     4.0/10
 4 Beaches092     00  00  00  1½  10     2.5/10
 5 ChessCraft096  00  00  00  ½0  ½0     1.0/10

Guenther

[Daniel Charles]

Hello Gunther

Thanks for your advices

sometimes I don t know what to do with my webspace (250 Gb) and my bandwidth (2500 Gb) like I only use 1 or 2 % , I could offer free space for opening book authors or engines author .... I will try next time with ASCII

Semi Final:

Naum 2.2 SMP vs Toga : 13.5- 6.5 :

I expected that That this Toga beta gave more hard time to Naum SMP after Naum Match vs Glaurung ... Maybe previous Rounf It was Glarung SMP lucky day



Zappa Mexico vs Rybka 1.0 : 16 - 4 :

Like I tested last time , this time again Zappa Mexico scored around 75 % versus the strongest free UCI engine (or one the top 3)

Super Final

Zappa Mexico VS Naum 2.2 SMP

100 Games
4 + 2



Zappa mexico will probably win , However we could see that Naum SMP is very strong and managed to draw a lot of games

Daniel

[Tony Thomas]

Are you going to finish the match even if Zappa scored more than 50 points? Wouldnt that be rather pointless?

[Daniel Charles]

Hello Tony

This is not pointless

This is not a match Between 2 engines with 200 or 300 elo difference

Zappa Mexico and Naum 2.2 Mp are in fact very close in strenght , less than 50 elo probably


90 games right now ...

and it is closer than people think .
If we are testers when we say 100 games this is a test for 100 games match

I don t see testers from CEGT or CCRL when they say 20 games stop their test after just 14 or 15 games ... because one engine score more than 50 %

More we will have games ... less we will have error .


Daniel

[Tony Thomas]

Hello Tony

This is not pointless

This is not a match Between 2 engines with 200 or 300 elo difference

Zappa Mexico and Naum 2.2 Mp are in fact very close in strenght , less than 50 elo probably


90 games right now ...

and it is closer than people think .
If we are testers when we say 100 games this is a test for 100 games match

I don t see testers from CEGT or CCRL when they say 20 games stop their test after just 14 or 15 games ... because one engine score more than 50 %

More we will have games ... less we will have error .


Daniel

I just said it would be pointless because it's a knock-out match. If an engine scores more than 50% of the possible score in a match, it wins by default. You do not see testers from CCRL or CGET stopping a 20 game match after 14 or 15 match because they are not playing knockout matches.

[Daniel Charles]

If you see the others match I didn t stop when an engine has over 50 %
I set up a 20 rounds match
Then I specified each match 20 games


for the final

this is special , I wanted to add more games .
and I set Up a 100 games match

and remember that all games played will be very useful for a rating list

Daniel

[Daniel Charles]

Xeon 5365 (8Cpu)
512 Mo
Egtb 3-4-5
100 games
4 mn + 2s
perfect9.ctg




Naum fought back however It was not enought to beat Zappa Mexico

Zappa Mexico managed to win this match and tournament

every chess engine testers know that Zappa scale very well on big Hardware.

Naum 2.2 smp is very close to the top level now. I could say that it scale efficiently too . It should not be ashame of its defeats versus Zappa because It was very close 47.5 - 52.5

Daniel

[ernest]

Naum 2.2 smp is very close to the top level now. I could say that it scale efficiently too . It should not be ashame of its defeats versus Zappa because It was very close 47.5 - 52.5
Daniel

Yes, but it seems the tournament is only 32-bit.
Zappa scales much better than Naum in 32 --> 64-bit.