new/updated engines

[George Speight]

Back up in the threads where Graham said something to the effect that CCRL wasnt testing this engine because it was a clone, or appeared to be so- or something to that effect- what he said was VERY PROPER. Im not posting this to defend him- but to make a point. When an UNKNOWN author pops up with an UNKNOWN program that is most likely in the top 10 in the world- or even the top 20- never heard of before- he is guilty. Meaning he should never have the benefit of the doubt. The onus is on him to prove his engine is legit- to convince everyone. It is not on us to prove anything. Because 99.9999999999999999999999999999999999% of the time it is a clone. That leaves a little wiggle room- but not much. And comparing this to anything Vas has done or come up with is like comparing apples to bricks. So dont start down that rocky road. Take, on the other hand, an unknown author who comes up with an unheard of program that is in the 1900 to 2300 elo range, or anything comparable- and you have a whole diff. scenario. He doesnt have as much he has to prove- meaning we are not watching him with as close an eye as compared to the above example. Anyone who cant see and doesnt understand that..........


Best,

I wouldn't put it quite so harshly. Innocent until proven guilty should always apply. However, we do have to be circumspect, because we don't want to waste our time or money (electricity bills, hardware usage) on illegal clones.
And certainly in the case of a completely unheard of author coming up with a world class engine, careful inspection of the engine is a necessary prerequisite to testing.
I made an error in stating a suspicion without posting the proof, something I criticised others for doing previously. However, fortunately the case was proven in the end.

I make generally 10 to 15 mistakes a day. They arent contagious and they wont kill you.

[Gábor Szots]

When an UNKNOWN author pops up with an UNKNOWN program that is most likely in the top 10 in the world- or even the top 20- never heard of before- he is guilty. Meaning he should never have the benefit of the doubt. The onus is on him to prove his engine is legit- to convince everyone. It is not on us to prove anything. Because 99.9999999999999999999999999999999999% of the time it is a clone. That leaves a little wiggle room- but not much. And comparing this to anything Vas has done or come up with is like comparing apples to bricks.

Well, I believe Vasik Rajlich was an UNKNOWN name when the UNKNOWN Rybka, certainly among the top 10 in the world, appeared... I leave the rest to you...

[George Speight]

When an UNKNOWN author pops up with an UNKNOWN program that is most likely in the top 10 in the world- or even the top 20- never heard of before- he is guilty. Meaning he should never have the benefit of the doubt. The onus is on him to prove his engine is legit- to convince everyone. It is not on us to prove anything. Because 99.9999999999999999999999999999999999% of the time it is a clone. That leaves a little wiggle room- but not much. And comparing this to anything Vas has done or come up with is like comparing apples to bricks.

Well, I believe Vasik Rajlich was an UNKNOWN name when the UNKNOWN Rybka, certainly among the top 10 in the world, appeared... I leave the rest to you...

Except that Vas quietly and without fanfare, in effect, or let others- at least those who would know that i respect and admire- prove his case for him. With this present case, it was "One doth protest too loudly".

[Teemu Pudas]

Well, I believe Vasik Rajlich was an UNKNOWN name when the UNKNOWN Rybka, certainly among the top 10 in the world, appeared... I leave the rest to you...

He was hardly unknown among the programmers.

[Christopher Conkie]

When an UNKNOWN author pops up with an UNKNOWN program that is most likely in the top 10 in the world- or even the top 20- never heard of before- he is guilty. Meaning he should never have the benefit of the doubt. The onus is on him to prove his engine is legit- to convince everyone. It is not on us to prove anything. Because 99.9999999999999999999999999999999999% of the time it is a clone. That leaves a little wiggle room- but not much. And comparing this to anything Vas has done or come up with is like comparing apples to bricks.

Well, I believe Vasik Rajlich was an UNKNOWN name when the UNKNOWN Rybka, certainly among the top 10 in the world, appeared... I leave the rest to you...

Not exactly unknown Gabor. He had an engine that Patrick tested, maybe some others. To be fair, the jump in elo between that and Rybka 1.0 Beta was massive to say the least, which I think is what you mean. From the clouds...so to speak.

The strongest closed source, uci or wb program that was around before Rybka was I think Zappa (...although I'm not sure....). I know that everyone took a good long look at Rybka when it came out.

One thing I can say for certain is that George should mark off at least one mistake as done, from his daily tally.

The figure is more like 50% since 2003 or so. We could count them out if anyone wants.

Ruffian
El Chinito
Zappa
List
Toga
Patriot
Rybka
Strelka

These were all engines that anyone could get their hands on. Maybe I missed a few....no idea.



Christopher

[Matthias Gemuh]

Back up in the threads where Graham said something to the effect that CCRL wasnt testing this engine because it was a clone, or appeared to be so- or something to that effect- what he said was VERY PROPER. Im not posting this to defend him- but to make a point. When an UNKNOWN author pops up with an UNKNOWN program that is most likely in the top 10 in the world- or even the top 20- never heard of before- he is guilty. Meaning he should never have the benefit of the doubt. The onus is on him to prove his engine is legit- to convince everyone. It is not on us to prove anything. Because 99.9999999999999999999999999999999999% of the time it is a clone. That leaves a little wiggle room- but not much. And comparing this to anything Vas has done or come up with is like comparing apples to bricks. So dont start down that rocky road. Take, on the other hand, an unknown author who comes up with an unheard of program that is in the 1900 to 2300 elo range, or anything comparable- and you have a whole diff. scenario. He doesnt have as much he has to prove- meaning we are not watching him with as close an eye as compared to the above example. Anyone who cant see and doesnt understand that..........


Best,

That is 100% my opinion too.
Cloning has to be made unattractive by allowing suspicions to be voiced without reservation. Looking back, none of the last 10 accusations turned out to be false. That alone should be proof that "accusers" don't suspect just for fun.

Matthias.

[Zach Wegner]

The figure is more like 50% since 2003 or so. We could count them out if anyone wants.

Looks a bit more like 75% to me.

[Christopher Conkie]

The figure is more like 50% since 2003 or so. We could count them out if anyone wants.

Looks a bit more like 75% to me.

I think I was trying to say that some, or even, more than a few, get through at that level. Fruit is living proof of it even if it is/was open source.

Of the list i gave 4 have been determined to be clones. 4 not. I've no doubt that it is not a comprehensive list.

Anyway 75% are what? Clones? Or not clones?



I'll let you be the first to name the other 2 whichever way you decide.



Christopher

[Gábor Szots]

Not exactly unknown Gabor. He had an engine that Patrick tested, maybe some others. To be fair, the jump in elo between that and Rybka 1.0 Beta was massive to say the least, which I think is what you mean.

Yes. A jump of 700 or so Elo is equal to emerging out of anonymity. In fact even more suspicious as it suggests the unability to make progress until...
If you are unknown and then emerge with a top class engine you can at least say you have been developing your engine for years but you judged it had not been rife for releasing earlier.

[Olivier Deville]

Hi Gabor

By the way, have you ever heard of an hungarian progammer whose name is Robert Szabo ?

He sent me an engine called Botond the other day... He says Botond's strength is around 2500.

The executable is packed/encrypted, and that indeed arises suspicion.

I'll certainly not take it in "as is"... Let's see if he cares to send me an exe that is not obfuscated.

Olivier

[Gábor Szots]

Hi Gabor

By the way, have you ever heard of an hungarian progammer whose name is Robert Szabo ?

He sent me an engine called Botond the other day... He says Botond's strength is around 2500.

The executable is packed/encrypted, and that indeed arises suspicion.

I'll certainly not take it in "as is"... Let's see if he cares to send me an exe that is not obfuscated.

Olivier

Hi Olivier,

No, I haven't heard of him. I am a member of a correspondence list of Hungarian computer chess programmers and I am sure I have never seen this name pop up. Still, I'm going to ask them.

[Norman Schmidt]

for many years new engines were simply accepted. many new engines (hundreds) were released and quickly became well established...their source code was never (and will never be) demanded for inspection.

thus, i'm confident (and it's very likely) that there are 'accepted' engines out there now utilizing code that is not 100% original...

so, if one engine needs to send source code, than all engines should have do the same...IMHO rybka and other commercials included.

because of your 'activities' against Rybka and against programmers of weaker programs.

Guenther, i never really considered that my effort to develop and improve cyclone was an activity 'against' weaker programs. (it's not mean or ill-spirited in any way). i think i somewhat understand your point, though. I do understand that these programmers have to now compete against a plethora of strong opponents: toga, cyclone, grapefruit, stockfish, etc...which makes the playing field even more difficult and more competitive. (i.,e. if open-source was not allowed, these 'weaker' programs would be more attractive....)

But at the same time, i also know that there are many many happy users out there who get an improved and stronger engine for free. is that a bad thing? it makes everyone work harder. competition provides motivation and benefits the consumer. the source code from 'open-source' engines is 'open' and free for all, i.e. any of the programmers of these 'weaker' engines that you mention could analyze the code, and use ideas from these 'open-source' programs to improve their own programs, if the wanted to...

I have not seen one post so far about programming issues from your side, neither here nor in CCC...

I'm not sure what you mean here...i post a lot on the CCC. are you active there?

Be assured that there were enough people around who cared a lot if programs were clones independent of any strength issue at least since
I follow the scene in 2000/2001.
Thus there were still much more clones of TSCP, Gerbil and Faile discovered then for all much stronger clones together.
You completely failed to understand HG's statistical comment, because it had nada to do with a difference in scrutiny in relation to strength!

With my statement above, I certainly didn't and don't mean to infer that the chess community 'dropped the ball' or was somehow was lax in it's duty to filter out clones in the past. I guess what i'm trying to say is that it's very likely that there are 'accepted' engines out there now utilizing code that is not 100% original. Please don't take that as a criticism of the work that you and many others have done all these years. But it's a difficult process, and is prone to error (like all human endeavor).

Norm

[Christopher Conkie]

Hi Gabor

By the way, have you ever heard of an hungarian progammer whose name is Robert Szabo ?

He sent me an engine called Botond the other day... He says Botond's strength is around 2500.

The executable is packed/encrypted, and that indeed arises suspicion.

I'll certainly not take it in "as is"... Let's see if he cares to send me an exe that is not obfuscated.

Olivier

Hi Olivier,

No, I haven't heard of him. I am a member of a correspondence list of Hungarian computer chess programmers and I am sure I have never seen this name pop up. Still, I'm going to ask them.

Botond has been found to be a compressed/encrypted Rybka Beta 1.0/Strelka/Belka by Michael and I.

No need to bother the programmers I think now Gabor.



Christopher