Fritz 11 Update Feb 29 2008

[Ampzilla]

A lot of rumors, but fact is, that we don`t have a new Engine update for the Fritz 11 Engine.


This from Peter at Chessbase

[mike860]

I wish some rumors would get started regarding the release date of Deep Fritz 11.

[Oscar L]

Let's us start the rumors then

http://groups.google.com/group/it.hobby ... 4ca95dffa0

[Yarget]

There seems to be a lot of confusion whether there exist an updated Fritz 11 version (11.1) or not. I can't say that I know the answer to that question but I just read some interesting results in the german CSS Forum. Gerhard Sonnabend (CEGT-Blitz 40/4)) has published these results:

Code: Select all

Fritz 11.1 vs
Deep Shredder 11 w32 1CPU 59.0-41.0 = 59.0%
Fruit 2.3.3f Test Beta    58.0-42.0 = 58.0%
Loop 13.6 w32 1CPU        65.5-34.5 = 65.5%
Naum 3.0 w32 1CPU         57.5-42.5 = 57.5%
Rybka 2.3.2a w32 1CPU     52.0-48.0 = 52.0%
(Gespielt ohne TBs)

zum Vergleich (soweit vorhanden)

Fritz 11.0 vs
Deep Shredder 11 w32 1CPU 87.5-62.5 = 58.3%
Fruit 2.3.3f Test Beta    60.0-40.0 = 60.0%
Naum 3.0 w32 1CPU         93.0-57.0 = 62.0%
Rybka 2.3.2a w32 1CPU     62.5-85.5 = 42.2%

Ich habe diese 500 Spiele mal in die aktuelle CEGT-Blitz-
Ratingliste integriert:
Fritz 11.1      2933   24  24   500    58.4 %   2874   40.0 %
Fritz 11        2908    8   8  4660    59.9 %   2838   38.1 %

Needless to say, 500 games aren't a lot but still..... Another interesting obervation is the results against Rybka 2.3.2a; while Fritz 11 managed to score 42,2% Fritz 11.1 scored 52%! Perhaps Fritz 11.1 is "designed" to beat Rybka?!

Regards
Per

[ml]

There seems to be a lot of confusion whether there exist an updated Fritz 11 version (11.1) or not. I can't say that I know the answer to that question but I just read some interesting results in the german CSS Forum. Gerhard Sonnabend (CEGT-Blitz 40/4)) has published these results:

Code: Select all

Fritz 11.1 vs
Deep Shredder 11 w32 1CPU 59.0-41.0 = 59.0%
Fruit 2.3.3f Test Beta    58.0-42.0 = 58.0%
Loop 13.6 w32 1CPU        65.5-34.5 = 65.5%
Naum 3.0 w32 1CPU         57.5-42.5 = 57.5%
Rybka 2.3.2a w32 1CPU     52.0-48.0 = 52.0%
(Gespielt ohne TBs)

zum Vergleich (soweit vorhanden)

Fritz 11.0 vs
Deep Shredder 11 w32 1CPU 87.5-62.5 = 58.3%
Fruit 2.3.3f Test Beta    60.0-40.0 = 60.0%
Naum 3.0 w32 1CPU         93.0-57.0 = 62.0%
Rybka 2.3.2a w32 1CPU     62.5-85.5 = 42.2%

Ich habe diese 500 Spiele mal in die aktuelle CEGT-Blitz-
Ratingliste integriert:
Fritz 11.1      2933   24  24   500    58.4 %   2874   40.0 %
Fritz 11        2908    8   8  4660    59.9 %   2838   38.1 %

Needless to say, 500 games aren't a lot but still..... Another interesting obervation is the results against Rybka 2.3.2a; while Fritz 11 managed to score 42,2% Fritz 11.1 scored 52%! Perhaps Fritz 11.1 is "designed" to beat Rybka?!

Regards
Per

The discrepancy in the results between Fritz 11.0 and 11.1 could be due to changes in the GUI, different engine parameters, hash tables, the opening book, or some other factor unrelated to the engine itself. We can't be sure if there are differences between the engine version 11.0 and 11.1 unless we run tests with all other factors identical.

[Spock]

I'm not convinced at the moment. +25 ELO is within the statistical margin of error, is it not ? Has anyone analysed some positions and seen different analyses from the two ? That would convince me

[ThatsIt]

Hi Ray !

Yes, i've done some tests:
http://f23.parsimony.net/forum50826/messages/178984.htm

The games for the CEGT-40/4-List were done under the CB-Fritz 10-GUI
with the same set of positions, identical size of hashtables and so on...

Best,
G.S.

[Spock]

Thanks - I can't read the German, but in the examples you give, the engine output on 11 and 11.1 is identical ?

[ThatsIt]

Thanks - I can't read the German, but in the examples you give, the engine output on 11 and 11.1 is identical ?

Yes, nearly.

Best,
G.S.

[ml]

The output is identical, but it looks like version 11.1 can calculate the same number of nodes in a shorter time. They have run these tests more than once, and the differences show up consistently.