To Christophe Theron

[Dr.Wael Deeb]

[DCCW]
_Deeb’s Computer Chess Workshop

Hardware
Lenovo ThinkCentre M57 6072 - Core 2 Duo E4400 2 GHz
[2.0GHz/6MB L2Cache],1GB DDR2 System Memory [2 Dimm]
Hashtables=256 Mb for each engine!
Software: Windows XP Professional 64,GUI:Arena 1.1,ChessBase GUI!
Nalimov EGTB:all 4 pieces and most of 5 pieces!
Conditions:
Time controls: 40 minutes per game + 40 seconds increment!
Each engine plays 2 games against all other engines!
Ponder off!
Learning enabled!
Own opening book!
No adapters are used!
No frozen,or out of time games!Just clear cut results!


Chess Tiger 2007.1 2700_123/204 games played
Chess Tiger 2007.1 Gambit 2654_129/220 games played
Chess Tiger 2004 2630_130/216 games played
Chess Tiger 15.0.0.1 2616_144/240 games played
Chess Tiger 14.0.0.1 2586_125/208 games played

So after playing more than 200 games at long time controls,I estimated a difference of 114 Elo points between Chess Tiger 14.0.0.1 and the latest version....

[Rolf]

Zach is partly correct. I just don't call this cloning. I meant I could improve Chess Tiger the easy way by taking the source code of Strelka, changing it line by line until it could not be recognized anymore and finally adding a number of search tricks that I have in Chess Tiger and that are not in Strelka.

This would give a program that would approximately by 200 elo points stronger than Chess Tiger 2007.1.

I could also do it in a less easy way by starting from the current source code of Chess Tiger, which would make the result look much less Strelka/Fruit/Rybka-like.

I'm not saying this is ethical or even interesting. It was just to illustrate a point.

I'm not serious about doing this because it's just not interesing at all. With Chess Tiger I have reached, in 1999 and 2001, the #1 rank on the SSDF and it was my source code, which I had been working on for several years, and for the most part without having access to any other chess source code. It does not mean a thing to me to achieve the same with someone else's code.

But I'm confident somebody else will find that this is a worthy goal and will actually do it.

// Christophe

Christophe, here we have a similar case of the one for what you already have apologized to all. You pretend you write such stuff just in joking but as it's factually appearing in a thread about the KN topic against Vas, this is insulting for Vas because you insinuate here but even very clear in the original thread that was moved away.

Cant you see that or what is going on? You practically insinuate that if you would do what you suppose Vas has done, zhis connect is absolutely clear in the original thread, you would also increase 200+ points easily. You may find it tongue in cheek but it's still insultive because it assumes a practice on the side of Vas. And now the same hype as with the KN count. What Vas did isnt dirty at all, it's kosher. Butg if you always describe it as if it were dirty, you insult Vas every time a new.

[Nimzovik]

I dare say that Christophe here trully misses chess programming-- ergo he MUST rail against the opposition and ABSOLUTELY prove he is RIGHT or he might just get re- possessed and start chess programing again! Which we all here KNOW that ends up in a divorce from wife , life and sanity

[tiger]

I dare say that Christophe here trully misses chess programming-- ergo he MUST rail against the opposition and ABSOLUTELY prove he is RIGHT or he might just get re- possessed and start chess programing again! Which we all here KNOW that ends up in a divorce from wife , life and sanity

Some say I have already divorced from sanity.



// Christophe

[Karmazen & Oliver]

Also, the major improvement in the new version was a much better branching factor. It means that at fast time controls it was approximately 2 to 4 times faster than the previous versions, but at long time controls it meant 8 to 16 times faster (to reach the same depth).

From this I expected CT2007 to perform even better at longer time controls, and so to show a gain of more then 80 elo points. However we were only a week before the release date and it was impossible to run long time controls tests. It would have taken a month.

So from the data I had I expected a 80 elo points improvement at fast time controls, and more at long time controls.

This is exactly what you could read on the CT2007 product at Lokasoft.

I still do not understand why CT2007 does not show a clearly better improvement at long time controls. You see, time to retire.

// Christophe

uppss ¡

I`m sorry from Spain. Oliver. ( I will use chesstiger today... on analysis mode, always. )