Similarity tests

[Frank Quisinsky]

Hi Sedat,

I will read it ...
A great invention is ...

Perhaps the most important today ...
Really easy to use ...
Without programming knowledge !!!

The ...
The ...
+++++ README file +++++

Here is nice to read ...
I am using for my first Version the sources by Fabian. I have a lot of additional ideas with the work Fabian do for us. Please have a look in my ideas in combination with the work from Fabian. Fabian gave his sources and all the nice ideas he has and I will give my ideas too. Thanks Fabian.

Such a text is great if a person used the work by an other one.
Sedat ... also in the year 2014 and for sure in the year the first Rybka version are available!

I am a fan from "Energy compansation"

I like that so much!
The boomerang never come back to me and has a nice fly.

You too?

Best
Frank

[Sedat Canbaz]

Dear Frank,

Sometimes I have difficulties to understand you ..really !!
But however, it's your rules...I respect!

BTW, a few notes more by me,

For examples,
After reading your comments, you stated that we should protect originality, right ?
And also you said: if we test clones...this is like RED cart for honest programmers, right ?

But you are missing one small point,
If are against those people who test clones, derivative engines
Why you you you (by yourself) are testing some of them (Critter,Fire)

In my opinion, if you want to be an good example for all testers,
You should not allow Fire and Critter in your tournaments!!!

One thing more,
Why you test only Critter and Fire ?
Why you make double standard to rest IPPO (Rybka) Authors ???
If we will follow your mentality... your rules...
All IPPO (Rybka) engines should be tested, correct me please if I am wrong ???

I see that in your eyes,
IPPO is original and Rybka is clone, right ? hmm...if you say YES, this is absurd...(((
Or maybe only Fire and Critter are 100% original, and all rest IPPO (Rybka) are clones ???

Please, please...you are making BIG mistakes my dear Frank, believe me !!!
1)You make double standart also to rest IPPO (Rybka) authors, because you test only Critter and Fire
2)You gave RED cart for honest programmers and also RED cart to rest IPPO (Rybka) authors!!
2)You made a BIG history mistake with killing Rybka and Houdini, believe me...

Maybe you forgot my story: How to be millionaire ?!
Just I'd like to add to my story:
Exception many years hard own working; lottery; bank robbery
-Also you can be millionaire, if you find treasure )))

IPOOLIT is same issue....they appeared as millionaire, but from unknown authors...!!!

IPPOLIT Authors:
Yakov Petrovich Golyadkin,
Igor Igorovich Igoronov
Roberto Pescatore
Yusuf Ralf Weisskopf
Ivan Skavinsky Skavar
plus Decembrists (all)

Is that possible, from many unknown people to create the world strongest engine ???
Don't get me lough please ))) I should be very naive to believe in that funny story )))

And if IPPPOLIT is original work,
-Why the similarity is approx. 60 % near to Rybka ???
-Why IPPOLIT and Rybka have almost same Elo points ???


And If Rybka is clone of Fruit,
-Why the similarity is approx. 48 % near to Fruit ???
-Why Fruit and Rybka have more than 400 Elo points ???


Explain me those questions please...thanks in advance


BTW,
Where are those IPOOLIT authors ???
We can't see those original IPPOLIT creators, hmm !!!

About Rybka's Author
Vas was here on well-known respected forums...for more than 6 years...!
And after the injustices over Rybka...he is not any more here, sad but true !!

Btw, maybe you missed...so here are some of recent sim test results )!

sim version 3
------ Critter 0.90 64-bit SSE4 (time: 100 ms scale: 1.0) ------
57.47 RobboLito 0.085g3 w32 (time: 100 ms scale: 1.0)
56.92 Fire 3.0 x64 (time: 100 ms scale: 1.0)
56.63 Elektro 1.0 (time: 100 ms scale: 1.0)
56.32 BlackMamba 2.0 x64 (time: 100 ms scale: 1.0)
55.43 Houdini 4 x64 (time: 100 ms scale: 1.0)
55.12 Rybka 3 (time: 100 ms scale: 1.0)


sim version 3
------ Fire 3.0 x64 (time: 100 ms scale: 1.0) ------
63.49 RobboLito 0.085g3 w32 (time: 100 ms scale: 1.0)
60.49 Houdini 4 x64 (time: 100 ms scale: 1.0)
60.32 BlackMamba 2.0 x64 (time: 100 ms scale: 1.0)
58.76 Rybka 3 (time: 100 ms scale: 1.0)
58.27 Equinox 3.20 x64mp (time: 100 ms scale: 1.0)
56.92 Critter 0.90 64-bit SSE4 (time: 100 ms scale: 1.0)


sim version 3
------ Rybka 3 (time: 100 ms scale: 1.0) ------
61.79 RobboLito 0.085g3 w32 (time: 100 ms scale: 1.0)
59.18 Elektro 1.0 (time: 100 ms scale: 1.0)
58.76 Fire 3.0 x64 (time: 100 ms scale: 1.0)
58.45 BlackMamba 2.0 x64 (time: 100 ms scale: 1.0)
57.65 Equinox 3.20 x64mp (time: 100 ms scale: 1.0)
57.00 Naum 4.6 (time: 100 ms scale: 1.0)
56.98 Murka 3 x64 UCI (time: 100 ms scale: 1.0)
56.48 Houdini 4 x64 (time: 100 ms scale: 1.0)
55.12 Critter 0.90 64-bit SSE4 (time: 100 ms scale: 1.0)

Note: I used time: 100 ms scale: 1.0,
I mean with Pia's method is expecting the results to be approx. 5% higher



Best,
Sedat

[Sedat Canbaz]

Here are the latest Fire sim test results (including Elektro)

sim version 3
------ Fire 3.0 x64 (time: 100 ms scale: 1.0) ------
64.54 Elektro 1.0 (time: 100 ms scale: 1.0)
63.49 RobboLito 0.085g3 w32 (time: 100 ms scale: 1.0)
60.49 Houdini 4 x64 (time: 100 ms scale: 1.0)
60.32 BlackMamba 2.0 x64 (time: 100 ms scale: 1.0)
58.76 Rybka 3 (time: 100 ms scale: 1.0)
58.27 Equinox 3.20 x64mp (time: 100 ms scale: 1.0)
56.92 Critter 0.90 64-bit SSE4 (time: 100 ms scale: 1.0)

[Frank Quisinsky]

Hi Sedat,

again:
Critter, News 072

For that reason I have a News-Ticker (better as to explain again and again in fora).

Fire, News 012

Why I am not using Rybka ...

1. Based on Fruit
2. To old

Why I am not using Naum ...

1. Based on Fruit
2. Not available, now available but Point 1.

Why I am not using Equinox ...

1. Same style as IPP engines
2. private available (now available, but same style as IPP).
3. Available but no readme file, no information to that important point.

And so one ...

Furthermore,
Fire and Critter goes not with a code from an other one Commercial, Vas and Robert do.

1. Code by Norman was free!
2. Code by Critter programmer was free (he sent his code from Critter to different others, no clone up to Version 0.80! To many own parts are included, proof for all to see, not in the case from Rybka and Houdini (for an example).
Critter programmer gave also a newer work free ...
http://sourceforge.net/projects/critterchess/files/
(v. 1.1.37)
Critter programmer gave a lot of examples to the Rybka - Fruit Story and Houdini Story in the past.

Big differents Sedat!
Very big differents ... to an programmer your heart ... Vas.

Such things I like!!

Again I will read it and all is OK for me!

Best
Frank

[Frank Quisinsky]

In additional ...

Winboard times ...
Crafty code was available, Pepito code, much other sources with around max. 2.450 Elo. No engine with more as 2.650 are available. And the programmer are not stupid at this time.

Later ...
Fruit Code are available, suddenly different others got bigger improvments. Rybka was born, Naum are stronger as in the first versions. In the time before ... not to see.

Later ...
IPP code are available, Glaurung sources other strong stources with around 2.600 Elo ....
Houdini was born, others are clearly stronger as in the time before.
Suddenly 30 IPP clones are available ...

Later ...
Stockfish, based on Glaurung are stronger and stronger ...
Soruces are free ...
Stockfish clones are available ...

Today it is not a wonder if a program with all this material have directly such a strength Houdini have. The Houdini programmer have no Chance to create such an engine for 10 years.

Rybka and Houdini are only results from the work by so many others programmers in the past. But to go commercial without to give also the own idea for free ... if free ideas are used ... isn't right.

Such things we can see in the future with a lot of more examples.
Today we are speaking about Houdini, tomorrow to an other one (programmers like Vas or Robert will come and go. Never heard before but directly an engine same strength as Komodo and Stockfish today. Do you think that all other programmers are stupid in the past?

That is allways the same ...
With the different ...

If the real programmers do nothing all the clones and derivative engine will never get such bigger jumpings.

A world only with clones and derivative engine and the development stopped. After 10 years 200 clones are on the same Level and the programmers of such engines are waiting for a Genius like Fabien or Tord. Sorry, can't have an other opinion if I am looking in past.

200 ELO stronger as Stockfish and Komodo I will see ...
From a Person with used the Stockfish sources in combination with IPP sources and all the other free material.

Vas and Robert are for sure good programmers but both forgot an important point ...

To give more information about the own work in his own ...
README FILES !!

To give his code free, if free works by others are used.

Sources are closed and try to sell ...
I don't like that a lot.
Maybe your like that ... I don't like that.
Nothing for my heart Sedat if I am thinking how many hours I gave for our hobby with all what I do in the past.

Best
Frank

[Frank Quisinsky]

With other words ...

The situation with Vas and Robert ist the same.
That is really a nice copy of a situation we have with both!

I can't see any differents!

And that is the main reason why I don't like different commercial companies in computer chess.

For many years the ChessTiger programmer wrote ...
Never a free engine will be stronger as a commercial engine.
I wrote to that message a lot!

Nothing to see from the ChessTiger programmer ...
That was nonsens.

Many commercials wrote a lot of nonsense in the past.

A group of x programmers can do more as a commercial programmer, sitting alone at home or switch information with a small group of other commercial programmers. A programmer alone have no chance for 20 years and today because a free community will be stronger and stronger with time with the result today ... that this work used by Commercials ... example Houdini and Rybka.

That is fact and our computer chess history is the proof!
But please ... I am waiting ... perhaps a new Genius will give other proofs?

I have time and can wait.

[Uri Blass]

1)It is clear that Vas learned from Fruit and it is the reason that Rybka was stronger.
It does not prove that Rybka is a derivative of the code of Fruit and it is possible to take ideas without taking code.

2)I think that Vas gave a lot to computer chess.
He did not need to release the source for it.

Without Rybka we could not see the Ippolit code and ideas from the ippolit code helped to make stockfish stronger.

For the question if commercial programs have chance to be stronger then free programs then I believe that they clearly have their chances and the question is who is the programmers of the commercial programs and who are the programmers of the free programs.

I believe that there is not a lot of money to earn from computer chess so the best programmers do not work on commercial programs or do not devote enough computer time to compete with the stockfish framework and this is the reason that today stockfish is number 1 or close to be number 1.

[Sedat Canbaz]

Hello dear Frank,

First of all,
You are in wrong way...sorry to say that

After reading your many letters...let me to publish my letter too )!

And I will try to help you (exception on opening books)

Because Critter 0.90 is another Rybka (IPPO ) one,
You made mistake... you should prefer the first Critter releases, not sure, maybe they are also Crafty or Fruit ones

Look...recently I also tested Fritz (v.10),
but I went back to 2006 years, because the latest Fritz versions are also Rybkanians

From your site:
"Robbolito" by Norman Schmidt was the first available engine based on the open source program Ippolit and this programmer spend the

most time for bugfixes and "code cleaning". No doubt about it that the successor of Robbolito, Fire 3.0 AVX x64, will be a worthy

representive engine for all other IPPs in my Continuous Tourney.

Norman is cleaned Rybka code ??? Are you sure of this ?
If Fire is clean, then why the similarity is approx. 60% close to Rybka ???

Maybe you was meaning Gull ?!
Because the previos Gull was releasd as Rybka (IPPO) derivative, but later Vadim cleaned Rybka code

Come on Frank, it's enough for today...I don't want to lough more again ..)))

Fire, Critter (like many other IPPOs) are just based on Rybka with different settings !
Exception Houdini, it's a master chess engine !! based mainly on own ideas !!!
Otherwise...Houdini would not be so strong and to be the 1st World number engine for 2 years!!
*A little note: Houdini 4 MP is approx 200 Elo stronger than Rybka 3 MP (incL.many rest Ippos too)
R. Houdarti's 1st engine maybe is based on Rybka, but he proved us that Rybka can improved to be much better!!!

About Rybka,
It's a just from another planet...I've never seen similar engine before
6 years to be unbeatable..exception via papers!
A real original chess engine by Vasik Rajlich
And there is no doubt that:
-Rybka belongs to the most original chess engines (based mainly on own ideas) !!!

Note also,
Rybka is not based on Fruit, maybe Vas learned , but it's not a Fruit derivative or clone!!
Another little note: No any engine to be 400 Elo stronger...and in the same tiime to be clone
There is no such thing in my mentality...!!! sorry...!!!

One thing more,
I don't relay too much on Human views, I prefer more to relay on Don's detection program!!!
Because Don's detection tool is much better than many chess engine experts, believe me in that !!!
I suggest to all chess engine experts to prove us that IPPOLIT is not clone of Rybka )))

So you are testing the brothers of clone IPPOLIT engine (Fire, Critter), well done to you !!))
I suggest you to use SCCT rules...it's almost perfect !!!

Btw,
Why you test Shredder ? it's also old...
I dislike such things Frank...again double standards by your side

But after all...
I respect your views, we are good chess friends since 2000 years...a lot of time...
Yes... you are very kind and interesting man and of course you are super journalist !)

And I hope one day...
I will meet you in the real life and we can drink together Chivas Regal )!


Have a nice weekend my friend,
Sedat

[Sedat Canbaz]

Just I'd like to add this too, In my mentality:
- 55% similarity and above, plus almost same Elo points (based on original program) is a clone or derivative engine !

So... I request just +100 Elo, not much...

Some chess engine programmers proved that (cleaning other codes, made improvements...)
And it's your turn...why not you ?)

[kranium]

Norman is cleaned Rybka code ??? Are you sure of this ?
If Fire is clean, then why the similarity is approx. 60% close to Rybka ???

Maybe you was meaning Gull ?!
Because the previos Gull was releasd as Rybka (IPPO) derivative, but later Vadim cleaned Rybka code

Come on Frank, it's enough for today...I don't want to lough more again ..)))

Hi Sedat-
I'm fairly sure that Frank doesn't mean "cleaned Rybka code"..., but rather "cleaned up" (fixed up) the source code

Look, for the record: I'm not the bad guy here...

I had nothing to do with the original creation and publication of the Ippolit source code,
but, as you know, I have been working extensively on the source code since it's initial publication in 2009.

Almost all of that work (until very recently) was published open-source on chesslogik.com...
this was simply an effort to distribute and share this new important source of chess-programming knowledge with the community.

My intentions were good, I added a GPL license to most of it, not to claim ownership of Ippolit, but as an effort to prevent it from being misused.
Well that failed, the GPL has been ignored, and there are now many clones/derivatives of Robbolito 085g3 and Fire 2.2.
In fact, there's been so much abuse I almost regret the work I did in this regard.

But the fact remains:
The cat is out of the bag...and Ippolit is one of the most important open-source releases ever and has added greatly to the collective body of chess programming knowledge
(which is now extremely well documented).
i.e. anyone can examine open-source engines like Crafty,Fruit,Ippolit/SF, etc. along with sites like chessprogramming wiki to learn all they need

Many are benefiting:
Houdini started with Robbolito
Stockfish has used ideas from Robbolito
Gull started from Ippolit
Larry K. spent years examining and studying Ippo source in a quest to determine what made it so strong
Chess engine users are benefiting most of all!

I don't think any derivative should be punished, it makes a lot of sense to start with an established codebase...
that's basically how most technology improves (must one re-invent the wheel for each new car model?)
(Stockfish itself is a derivative!)

The chess programming community desperately needs guidelines and criteria in regards to clones/derivatives,
that's why I like your idea about requiring derivatives to demonstrate a substantial ELO increase (along with non-similarity).

Testers and rating lists can simply have 2 (or more) categories: for ex: 'original' and 'derivative'.
Use the sim tool along with other examinations, and require a strength improvement to determine if a derivative has changed enough to be called a new unique program.

This would remove many of the clones/derivatives form the rating lists, benefiting established authors of 'original' engines...
and will also provide the much needed criteria and guidelines for anyone choosing to develop a 'derivative' engine into something unique and interesting.

Best Regards,
Norm