Deep Fritz 11 eval

[BB+]

I have obtained a copy of DF11, and have deciphered some of [what appears to be] the evaluation. My initial opinion is that it [the DF11 evaluation] is not particularly derivative of Fruit and/or Rybka/Strelka. I will post most later [likely a few days], including an example.

[kingliveson]

I take it this is related to postings on talkchess using DD's sim version 3 that supposedly detects "clones." I thought the discussion and conclusion months back was that this tool was not reliable...

[BB+]

I attach the White Rook eval of DF11. Note that I was not overly assiduous in determining every detail, as it was fairly apparent that there were many differences from Fruit and/or Rybka/Strelka. I also did this simply from objdump, w/o actually running the engine (so runtime-determined array elements are not always known). My data are that this Fritz version appeared in Nov 2007, while the Strelka source code was not totally public until Jan 2008.

[Damir Desevac]

May I ask who sent you a copy of the program ? If it is the Fritz team, or somebody close to them, whatever similarites it have/had with Fruit/Strelka/Rybka might have been removed...
Did you check the single version of Fritz 11, that one most likely would have big similarities to Fruit/Strelka/Belka or Rybka, as that was the only version available before DF 11 was released. To some users it appeared to be very Fruitish, Rybkish or Belkish... And many complained that it lost some of it's soul, because it lost its style of play.

[Uly]

Deep Fritz 11 is the successor of Deep Fritz 10.1, which was the version with "soul", and DF11 was the one doing the big leap in elo (and remains strongest Fritz to date over DF12 and F13) while suddenly playing very dull chess when compared.

So BB+ is checking the correct version.

[Damir Desevac]

I think he should give a single version a try...

[hyatt]

I think he should give a single version a try...

On CCC you claimed DF 11 was the wrong version to test, that for some reason, you thought Frans had removed the "code" from Fritz 11 that looked Fruity, so that it would not be in DF 11, something I do not understand conceptually...

[ernest]

Deep Fritz 11 ... (and remains strongest Fritz to date over DF12 and F13)...

What do you mean?
In the rating lists (see CEGT 40/20) DF11 is weaker than the successors...

[BB+]

I did not obtain DF11 from the Fritz people. I do not have the single cpu version. It turns out that DF11 was almost a year later than F11: http://www.chessbase.com/newsdetail.asp?newsid=5027
Note that they call F11 "a complete rewrite" there.

The other piece evaluations show various distinctions in evaluation (e.g., knights have a Nc3/Pc2/Pd4/bPd5 pattern). I have not made any attempt to decipher the search. I can't say that I am all that interested, and am busy with other matters in any case.

[Damir Desevac]

I think he should give a single version a try...

On CCC you claimed DF 11 was the wrong version to test, that for some reason, you thought Frans had removed the "code" from Fritz 11 that looked Fruity, so that it would not be in DF 11, something I do not understand conceptually...

Maybe this would put more light to the matter. Single Version of Fritz 11 ahead of MP engines. I think it is very common in my opinion.

http://www.husvankempen.de/nunn/40_120_ ... liste.html