Similarity tests

[kranium]

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 ?)

Yes it can be done Sedat, but it's not easy (and it shouldn't be easy!)...
for ex: change 1 thing in Ivanhoe and if it doesn't crash it gets significantly weaker.

Concerning Fire: I think your criteria are reasonable.
I'm working hard in that regard and hope to meet your requirements within the next year or so.

[Sedat Canbaz]

Hello dear Norman,

First of all,
I am not against to any chess engine programmer (including Fire)

But really it makes me very sad, when I see injustices...

Especially when I see from well-known tester, for example from Frank

He makes double standards....sad but true....!

Frank claims that Fire (your engine) + Critter are original
But in reality this is not true...we are not so naive...!!!
I don't claim that Fire and Critter are clones
But I strongly believe both engines are derivatives (based mainly on Rybka ideas)

BTW, sometimes I am wondering....
Who cracked Rybka and released as free??
Who is behind of the real author of IPPOLIT ?
Do you have any info about this issue ??

BTW,
If I was 1 year with experience with computer chess, maybe I would believe...

But IPPOLIT is a clone in my eyes, sorry to say that
Exception Don's utility that confirms the high similarity to Rybka,
Elo calculation programs too confirm almost same Elo points to Rybka
And even, many experts believe in that (including me)
Only naive people will have difficulties to understand what I am meaning...

One ting more, If we really want to be professional testers:
-We need to be independent testers!!!

In other words, we should not say:
- I don't like this engine, or this engine has a different playing style etc..

For example,
Tornado is created by Turkish Author...and I should allow only Tornado...but not rest, hmm
Or if I allow only Rybka and to not include any other engine which I don't like....this would be clearly wrong...a history mistakes

In shortly, I mean:
Professional testers should follow strictly the rules...!!!

Otherwise...
If we will follow Frank's methodology that his rules are great, then:
*I suggest to Frank to test all other rest IPPOs (Rybka):
-Vitrivus (according to its author:it has Human playing style)
-Strelka can be another good choice...I think it is also different
-Ivanhoe,it is different in playing style too, not bad idea to be tested
-Heron, a new Stockfish clone...I wonder what will its performance in FCT rating list
-Twinfish...another Stockfish clone, also lets see what will be the performance in FCT
Btw, we should not forget to include: Boucket, Tankist, Naum, Electro etc...
And I should not forget to mention this too:
- my BabyMaster engine should be tested too, otherwise there will be again double standard )))


Hopes helps,

Best,
Sedat



Btw, in the beginning of this year I tested all top 10-12 chess engines,
but I did not make no any double standard to any IPPO (Rybka) authors !!!

[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 ?)

Yes it can be done Sedat, but it's not easy (and it shouldn't be easy!)...
for ex: change 1 thing in Ivanhoe and if it doesn't crash it gets significantly weaker.

Concerning Fire: I think your criteria are reasonable.
I'm working hard in that regard and hope to meet your requirements within the next year or so.

Very good...when you clean Rybka code and when you put +100 Elo over Rybka:
- For me will be honor to test your Fire engine in SCCT

Best,
Sedat

[kranium]

Frank claims that Fire (your engine) + Critter are original

I don't think he has said that or believes it...
Frank is very aware, knowledgeable, and extremely perceptive in this field.

I don't pretend to speak for him but:

I believe he has simply chosen to test Fire in place of Houdini because of the Houdini author's actions
(i.e. also: he is not basing his selection criteria on ELO alone)

I think he has probably chosen Critter 0.90 to test (as a compromise) because of the many 'red flags' now associated with the newest versions and recent events.

Norm

[Sedat Canbaz]

Frank claims that Fire (your engine) + Critter are original

I don't think he has said that or believes it...
Frank is very aware, knowledgeable, and extremely perceptive in this field.

I don't pretend to speak for him but:

I believe he has simply chosen to test Fire in place of Houdini because of the Houdini author's actions
(i.e. also: he is not basing his selection criteria on ELO alone)

I think he has probably chosen Critter 0.90 to test (as a compromise) because of the many 'red flags' now associated with the newest versions and recent events.

Norm

Hello Norman,

Of course of course... )))

But I think you missed one of Frank's statements:

Never, never, never we should support it!!!
The result will be a community full of clone and derivative fans and the development of interesting Computer chess things will be stoped for ever.

That is a red Card to each honest programmer, able to develops new and interesting own things in fair competition / contest.


So...what about Critter and Fire ???
I am shocked that Frank is testing both engines !!
In other words, Frank support derivative engines!!!

And last,
Frank is free and he can test any engine which he likes...and with any opening...

But be sure in that,
Frank's rules are not good...I can say bad example for rest testers

Killing the mother: Rybka (one of the most original engines), and allowing only 2 (two) Rybkanian derivatives (Critter, Fire) it's a HISTORY MISTAKE !!!

Sorry....I dislike such things...


BTW, with Frank's rules:
-We should test all rest IPPOs (Rybkanian) engines too....there is no other explanation!!!



Best,
Sedat

[Frank Quisinsky]

Hi Sedat,

I can't follow your logic, sorry!

You are starting your view in the middle of the story!
Better as to start on the end of the story or we have a Houdini fan more

Rybka isn't the start point ...
Start point is Fruit.

In my older Rybka folder I found 32 Rybka 1.0 Beta x or 1.2x versions. I notice that you are a fan from the clone tool (have a look what Richard Vida wrote to the clone tool, or to Rybka - Fruit). If you like to compare with the clone tool ... do it with Fruit and Rybka and you find the correct startpoint of these story very easy in self work to.

Compare Fruit with all the Rybka 1 versions step by step.
You will not like the results because your heart ...

To the other points I have no interest to write more because I believe you are not able to read all the answers by myself. Again and again the same by yourself and that is very boring. This weekend I have a lot of prof. work and not many time for chess. Absolutely KO (but I will start the next round robin in the next hours) and I am not able for complicated discussion after a very hard day.

The complete topic is very boring Sedat ...
Let us discuss about other things, each other computer chess topic is much more interesting as again and again the same to Rybka-Fruit the IPPs and all the other clones available in the past.

Again ...

I like readme files.
Thanks Fabien ... I am using ...
That's all what I await ...

Best
Frank

[Frank Quisinsky]

I'm fairly sure that Frank doesn't mean "cleaned Rybka code"

Never I wrote that, yes!
Good that you wrote a short statement to this topic.

Best
Frank

[Sedat Canbaz]

Hello dear Frank,

If my questions are boring (which you could not answer), no problem !)
We can talk about different questions, but however it's enough...

A little note more:
-I prefer to work with original creators, including talented authors too!

Btw, do you know why I request + 100 Elo (in case of + 55% similarity)??
-Simply the own original ideas are not so cheap, try to understand this please!

And it's pointless to discus anymore,
If you agree with me is good, if not... it's too bad !


I wish you good luck with you tournaments and with you books too !


Kind regards,
Sedat

[Frank Quisinsky]

Hi Sedat,

possible!

But who say to you that +100 Elo based on own ideas?
A programmer able to use the sources by an other programmer can be used the sources by an third programmer too. That ist more probably as own ideas, sorry!

A proof for an own work can be check by an other programmer, no problem today if anything is a bit critical. If not ... the programmer have to life with the opinion others can build. I like the way the Critter programmer go ... he sent his sources to others which have interest to see it. So big differents to others ... maybe ... Dollar's in the eyes! Can be set as engine logo also Dagobert Duck grafics.

Nice idea I think ...


Carl Barks will give his agreement for it in such a case ... sure with it.

Best
Frank

[Sedat Canbaz]

Dear Frank,

I think my comments are started to be not boring now )))

Ok...lets continuing...

The current used SCCT rules are almost ideal...I suggest you again
Just I want to help you my friend, please don't get me wrong...

My rules are very strict Frank, not like your yours!
At least...I don't make double standards...

SCCT has no any derivative or clone engine to be as participant
Exception Houdini...but you know why...I allow it...
And in my eyes its original, because it's based mainly on own ideas...

Yes... Houdini can be Rybka derivative, but 200 Elo stronger than Rybka !!!
Do you know +200 Elo... how can be improved ??? if it is so easy:
Can you do that or any other engine programmer???

There was no any available any stronger engine during 2011 years,
And firstly only Robert Houdart managed to make stronger (+200 Elo) over than Rybka
That means a lot for me...!!!

The same issue for Rybka,
There was no available any other stronger engine,
And firstly only Vasik Rajlich managed to make stronger (+400 Elo) than Fruit
That means a lot for me too...!!!

Btw, I made some calculations...maybe will be useful

1)Rybka 3 is 48 % similar to Fruit 2.1
2)Rybka 3 is +400 Elo stronger than Fruit 2.1

3)Fire 3 is approx 60 % similar to Rybka 3
3)Fire 3 is approx. 50 Elo stronger than Rybka 3

Note:the great Don Dailey's view: +60% sim results can be considered as clones

5)Houdini 4 is approx 60 % similar to Rybka 3
6)Houdini 4 is approx.200 Elo stronger than Rybka 3

Ok...let's take example Twinfish (a 100% clone, optimized against sim test tool),
7)Twinfish is approx 53 % similar to Stockfish
8)Twinfish is approx. 100 Elo weaker than Stockfish


And now do you see the difference ?

In other words,
-We need to check Elo points, really these values help us much...
-Chess engine expert's advise is very important and should be check too
-Of course the great Don's tool should be as a main part in case of detection...

And all together...our rules will be almost perfection !!!
Otherwise we will make double standards again...RED cart )))
And we will see a lot of same or similar chess engines!

Believe me + 100 Elo is not so easy as it looks...
Only talented programmers can do that !!!

Stockfish team,Komodo team, Robert Houdart, Vasik Rajlich,
Vadim Demichev are managed to that...why not others too ??

Note also
Fire chess engine is released later than Rybka!
And definitelyFire is based mainly on Rybka's ideas!!
That's why Fire's Elo points and its similarity is near to Rybka

Can we say same thing for Rybka vs Fruit ???
+400 Elo difference ...
Or
Can we say same thing for Rybka vs Houdini ???
+200 Elo difference too....

Please, please, please...you should change your view...!!!

Note: I am not from Rybka or from Houdini team, but just I don't like injustices

Give me any other example, where I allow clones or derivatives ???


Btw, see below the table please, tell me which engine is clone and should not be tested ???
Just don't say again Rybka and Houdini...
Because both engines are much stronger and based mainly on own ideas (Note: 400 Elo and 200 Elo)

And last, in shortly:
+55% similarity + 100 Elo

Hopes helps this too...


Best,
Sedat