Strelka 2.0 B running for the IPON ...

[Houdini]

Guillermo,

Today there are about 20 versions of ippo occupying the tops of the tables and therefore this "drop down" the other amateur engines (and also redirects the attention of the general public).

Where do you find "20 version of ippo occupying the tops of the tables"?
For example look at the following 3 lists:
- IPON list at http://inwoba.de/index.html
- SWCR list at http://www.amateurschach.de/swcr/_swcr-all.htm
- SCCT list at http://sedatchess.110mb.com/index.php?p=1_58

Do these lists conform to your statement?

Robert

[guillef]

I got tired of seeing tournaments including multiple versions of ippo, fire, Robbo, Houdini, etc, Now I do not have the time to look, but you can find these tournaments yourself in this forum and the like.

The links you sent me do not support exactly what I wrote, but did not deny in all the general idea of what I attempted to express, which is not precisely what you replicate in that little line between ""

[michiguel]

I think the point about Strelka and many others is quite simple (from my point of view, although Miguel may agree):

Of course, I agree. Even if it is legitimate to test Strelka, it does not mean it is desirable. The trend would be utterly boring. Having many engines that may choose the same moves is a waste of time. What I disagree is that I do not think it matters if they are at the top or not.

Miguel

strong derivative engines filled the top of all existing rating tables. Today there are about 20 versions of ippo occupying the tops of the tables and therefore this "drop down" the other amateur engines (and also redirects the attention of the general public). This is very good for testers (better and better engines to test), but it is not "motivating" for authors who every day have 5 new +3100 ELO engines that move to their own engines more and more down on the table and away from almost all the valuable tests that run the testers (except CCRL, which tests engines throug all the range of ELO).
I addition, I have my reservations on how good it is to try 20 or 50 versions of virtually the same engine to detect after thousands of hours and computer "horse-power" that the performance of all derivatives is about the same +- 10 ELO in limited testing conditions.

Regards,
Guillermo

[Houdini]

I got tired of seeing tournaments including multiple versions of ippo, fire, Robbo, Houdini, etc, Now I do not have the time to look, but you can find these tournaments yourself in this forum and the like.

The links you sent me do not support exactly what I wrote, but did not deny in all the general idea of what I attempted to express, which is not precisely what you replicate in that little line between ""

Your main point was that you got tired of the fact that "strong derivative engines filled the top of all existing rating tables". (again, just quoting what you wrote).

That claim is simply and demonstrably not true. People that organize high-quality rating lists or tournaments (for example TCEC) are usually intelligent enough to avoid including multiple, nearly identical engines.

Robert

[frcha]

Its really simply -- you just try to determine which is the best ivanhoe version to test -- a later version already shown with good results and use only that..

I had a mini tournament won by Rybka 4 with ivanhoe, stockfish, and critter tied for 2nd place.
Then I added Houdini ...
Only Houdini and 1 version of Ivanhoe need to be added -- thats 2 derivatives - 3 including Rybka.

[guillef]

An example of what I say, I seek no more because I have no time or interest, although there are a lot more in these forums.

http://www.talkchess.com/forum/viewtopic.php?t=34001

If you do not share what I say, is OK, I do not think we need to take pieces of my post between "" and try to "prove false" or "true" like Maths proposals. I guess there are people who are "agree" with what I say. I'll try to put it in more general terms: derivatives of ippo series are being carried more and more attention from everyone, and programmers do not even know someone to ask about, and to share the wonderful ideas of his work. This really happened fortunately with Crafty, Fruit, Spike, Glaurung / Stockfish, even TSCP, and have contributed much to this community.

[Houdini]

Guillermo,

I can perfectly understand and sympathize with your point of view, as long as it is expressed with some nuances.

Robert

[IWB]

Strelka finished the first 1900 games.

http://www.inwob.de

2 Elo missing for the best official Fruit spot ... 300 games to go.

Bye
Ingo

[Uri Blass]

Strelka finished the first 1900 games.

http://www.inwob.de

2 Elo missing for the best official Fruit spot ... 300 games to go.

Bye
Ingo

I do not consider strelka as fruit.
The difference between strelka and fruit is clearly bigger than the difference between toga and fruit and I think the difference is big enough to consider them as different programs.

I think that it is clear that strelka is derived from reverse engineering of rybka1 beta so you can consider it as rybka derivative and the similiarity in move choice between strelka and rybka1 beta is clearly higher than the similiarity in move choice between Strelka and one of the versions of fruit or toga.

[Dann Corbit]

Strelka finished the first 1900 games.

http://www.inwob.de

2 Elo missing for the best official Fruit spot ... 300 games to go.

Bye
Ingo

I do not consider strelka as fruit.
The difference between strelka and fruit is clearly bigger than the difference between toga and fruit and I think the difference is big enough to consider them as different programs.

I think that it is clear that strelka is derived from reverse engineering of rybka1 beta so you can consider it as rybka derivative and the similiarity in move choice between strelka and rybka1 beta is clearly higher than the similiarity in move choice between Strelka and one of the versions of fruit or toga.

Here is my understanding:

Yuri Osipov took some of the fruit code as an outline, and put the stuff that he reverse engineered from Rybka into it.

I think it was sort of a protest, because he thought that Rybka was a modified version of Fruit {this is only suspicion on my part, because he has not said that as the reason as far as I know}.

Some of the things in Strelka are obviously fruit derived without thinking about it because they have french root names.

I suspect that it did not bother Fabian because Strelka became open source. However, Fabian could answer more clearly here.

My opinion about Strelka:
It's not fruit, it's not rybka. But it obviously has stuff gathered from both of them. I have no idea if strelka is legal.