Not logged inRybka Chess Community Forum
Up Topic Rybka Support & Discussion / Rybka Discussion / Rybka can't match both Fruit and Crafty
1 2 Previous Next  
- - By Venator (Silver) Date 2013-09-11 08:53
From another thread.

Bob quote:

Ever looked at Fruit?  Ever looked at Crafty?  Ever compared the two?  Vastly different overall structure.  Rybka can't match BOTH.

Quote by User923005:

Rybka Investigation and Summary of Findings for the ICGA
Mark Lefler, Robert Hyatt, Harvey Williamson and ICGA panel members
12 May 2011
1. Background
1.1 Purpose: To investigate claims that the chess playing program Rybka is a
derivative of the chess programs Fruit and Crafty
and violated International Computer
Games Association (ICGA) Tournament rules.


A perfect example where Hyatt forgot his own statements and got caught. Obviously, when he claimed that Rybka cannot match both Fruit and Crafty, he was not aware of the fact that he completely destroyed the ICGA argument AND verdict that Vas plagiarized BOTH Fruit AND Crafty in Rybka versions that played in ICGA tournaments.
Parent - - By Rebel (****) Date 2013-09-11 09:36
Bob quote: - Ever looked at Fruit?  Ever looked at Crafty?  Ever compared the two?

Is it suddenly allowed to go forward and backward through foreign source code? And even compare them and make notes?

I am puzzled, is it suddenly allowed for aspirant programmers to make notes from Crafty?

Questions, question, questions.........
Parent - - By Venator (Silver) Date 2013-09-11 09:41 Edited 2013-09-11 09:47
Apparently Bob has gone forwards and backwards through Fruit, otherwise he never could make that statement about 'Vastly different overall structure'!
Parent - - By bob (Gold) Date 2013-09-11 13:39
One has HAD to go back and forth through Fruit during this investigation.  Ever think of that?  Feel free to compare fruit and Crafty however, to convince yourself they look nothing alike.
Parent - - By Venator (Silver) Date 2013-09-11 13:51 Edited 2013-09-11 14:09
I am more interested in the detailed notes you made during that process.

And of course I don't believe you: as soon as the Fruit source code became available, you have studied it extensively. Like everybody else.

That there is nothing to be found in Crafty is not such a big deal. We know from your LMR experiments that you didn't know exactly how to make it work. So you removed it. I am sure there are a lot of other tries that didn't deliver what you expected, so you threw them out again and they never made it into an official Crafty version. That this is not a fairy tale or an assumption, is proved by the Fruit bishop PST.
Parent - - By bob (Gold) Date 2013-09-11 15:10
Eh?  Crafty has used LMR for years, still does.  I don't use history info to trigger/disable LMR.  Neither does Rybka, or many other programs.  So your point is what, exactly?  The PST proves exactly nothing.
Parent - By Venator (Silver) Date 2013-09-11 15:43
The PST proves that you took things from Fruit. More than you want to admit.
Parent - - By turbojuice1122 (Gold) Date 2013-09-11 10:29
I merged it here :-)
Parent - - By bob (Gold) Date 2013-09-10 20:24
Feel free to map Crafty's procedure calls into Fruit's code.  Report back after you fail.
Parent - - By user923005 (****) Date 2013-09-10 20:37
If it can't be done, then Vas must logically be innocent of the charge of merging crafty and fruit code.
It looks like you have acquitted him with one logical swoop!
Parent - - By bob (Gold) Date 2013-09-10 22:45
Eh?  No one ever said he copied all of Crafty, now did they?  Not after 1.6.1 anyway.  1.0 beta was derived from Fruit.  They would have the similar function mapping...

Looks like once again you rely on distortion when there are no supporting facts???
Parent - - By Banned for Life (Gold) Date 2013-09-10 23:25
Are you stupid?

Your verdict, delivered to David Levy, claimed that Rybka won its WCCC titles by plagiarizing Crafty and Fruit. Please show us something in any Rybka WCCC entry that used any part of Crafty. Anything at all.
Parent - - By bob (Gold) Date 2013-09-11 00:25
Are YOU stupid?  This discussion was about mapping procedures to procedures.  It works WELL for comparing strelka/fruit and strelka/rybka 1.0.  Has nothing to do with crafty.

So.  Are YOU stupid???

He can't copy the structure of BOTH programs, they are way too different...
Parent - - By Banned for Life (Gold) Date 2013-09-11 04:14
Are you stupid?

Your verdict, delivered to David Levy, claimed that Rybka won its WCCC titles by plagiarizing Crafty and Fruit. Please show us something in any Rybka WCCC entry that used any part of Crafty. Anything at all.
Parent - - By bob (Gold) Date 2013-09-11 13:37
You've already shown your hand here, so it is pointless.  For example, you claim rotated bitboards is an algorithm, as opposed to actual code.  You claim that things like the bishop shift arrays are just a bunch of non-copyrightable numbers.  In your world, NOTHING can be copied because EVERYTHING becomes a formula, or a table, or an algorithm...
Parent - By Venator (Silver) Date 2013-09-11 14:11
Where is the Crafty code in Rybka versions that played in ICGA tournaments? The million dollar question you STILL didn't answer....
Parent - - By user923005 (****) Date 2013-09-10 23:46
Rybka Investigation and Summary of Findings for the ICGA
Mark Lefler, Robert Hyatt, Harvey Williamson and ICGA panel members
12 May 2011
1. Background
1.1 Purpose: To investigate claims that the chess playing program Rybka is a
derivative of the chess programs Fruit and Crafty
and violated International Computer
Games Association (ICGA) Tournament rules. Rybka is a program by Vasik Rajlich.
Fruit was written by Fabien Letouzey. Crafty was written by Robert Hyatt.
Parent - - By bob (Gold) Date 2013-09-11 00:26
Doesn't change a thing.  Ever looked at Fruit?  Ever looked at Crafty?  Ever compared the two?  Vastly different overall structure.  Rybka can't match BOTH.
Parent - - By user923005 (****) Date 2013-09-11 00:27
So the version that competed in ICGA contests did not have any crafty code in it then?
Parent - - By bob (Gold) Date 2013-09-11 04:26
Never seen anyone stoop so low.  Continual distortion.  Etc.

I won't say more, you know what is wrong with your statement.  Again the attempt at being disingenuous does not stand any real scrutiny...
Parent - - By user923005 (****) Date 2013-09-11 05:14
You said that it is too complicated to mix them because of the differing data structures, so I asked if you thought there was not crafty in the ICGA Rybka.
Apparently you are afraid to answer.  The reason is obvious.  If there is crafty in rybka, then it is possible to mix the two systems, which breaks your current argument.  And if there is no crafty in rybka it destroys your previous statement saying that there was crafty in rybka.  So you have nowhere to stand.  So now, we can clearly see that someone is being disingenuous.
Parent - - By bob (Gold) Date 2013-09-11 13:29
I do not believe Rybka 1.0beta and beyond is a direct copy of Crafty, unlike what we saw in Rybka 1.4 - 1.6.1.  Since it is MUCH more like fruit, Fruit's procedures would map to rybka quite well.  Crafty's would not.  The rotated bitboard code in Crafty is something that is inlined at point of use, every time you need the attacked square set for a sliding piece.  No procedures would remain in the binary.  Ergo, mapping MY procedure names to Fruit or Rybka is simply not possible.  Structure is not the same.  Organization is not the same.

You are, as you have been lately, completely wrong.  The only observation I have made about 1.0 beta and beyond was that in one 1.0 beta executable, I had noticed the "bishop shift" arrays.  That led me to speculate that my rotated bitmap code was ALSO present.  But I did not look for it, because, as I have repeatedly mentioned, I did not consider it to be very significant.

Yes, someone is being disingenuous.  You.  And there's no doubt you know this.  If that is the way you want to carry on the argument, feel free to do so.  But you get to argue with yourself.  This is not an ALL or NONE thing.  If one copies rotated bitboard code, there are NO procedures to map.  Seems simple enough.  Sort of like the lame argument Ed tried on Richard, "That initialization is not done in the go_parse() code of Fruit, it is only there in Rybka".  When you look at the binary, you see the final product of the compiler.  If it inlines something, you can't tell whether that was inlined or was directly inserted at that point by the programmer himself.  All you can do is examine the code and determine if both do the SAME things, or not.

If you can't understand the above, you need to change your hobby...
Parent - By Venator (Silver) Date 2013-09-11 13:44
Yes, someone is being disingenuous.  You.  And there's no doubt you know this.  If that is the way you want to carry on the argument, feel free to do so.  But you get to argue with yourself.

If you can't understand the above, you need to change your hobby...

Translation: 'Oops, I have once again been caught contradicting myself. I cannot and will not answer User923005's question. Divert attention. Name calling.'

You are such an open book, Bob :lol:
Parent - By Banned for Life (Gold) Date 2013-09-11 16:27
The verdict you provided to David Levy claims that the Rybka entries to the WCCC plagiarized Crafty, yet you cannot provide a SINGLE example of this?
Parent - - By user923005 (****) Date 2013-09-11 18:55
All I was interested in was your admission that Rybka 1.0 beta and beyond does not have crafty code in it (or, at least it has not been demonstrated that there is any crafty code in Rybka 1.0 beta and beyond).
The reason I brought it up is that means that it makes no sense to punish Vas for it in ICGA events.
Thank you for your expert opinion and clarification.
Parent - - By bob (Gold) Date 2013-09-11 22:49
How can I admit something I do not KNOW?

Sorry...

It does not matter whether he copied Fruit OR Crafty.  One is more than enough.  Give me some time to look at the binary and extract whatever seems "interesting"...

I'll report here as soon as I have the necessary time, regardless of what it shows.
Parent - - By user923005 (****) Date 2013-09-11 22:58
The salient point here is that there is no known infraction with crafty code for the Rybka versions that played in ICGA events.
It seems important to me.
Parent - - By bob (Gold) Date 2013-09-11 23:11
Suppose that after I look, I find NOTHING from Crafty in Rybka 1.0 beta.  Does that change a thing?  Of course not.  There is STILL a ton of fruit there, and that is more than what the rules allow. 

Just hold that thought for a bit...
Parent - By user923005 (****) Date 2013-09-11 23:28
That might be true.  However, thus far nobody has properly demonstrated that.
So as I see it, no crafty at all and no fruit proof.
Parent - - By Venator (Silver) Date 2013-09-12 16:08
I find NOTHING from Crafty in Rybka 1.0 beta.  Does that change a thing?

Yes, of course it does. I think you owe Vas an apology because of the totally false claim that Rybka plagiarized Crafty in ICGA events.
Parent - - By bob (Gold) Date 2013-09-12 18:36
Please quote the entire statement.  I did NOT write what you want to imply.
Parent - - By Venator (Silver) Date 2013-09-13 10:30
You: suppose I find nothing in R1.0 beta. Conclusion (which we already know): you don't have ANYTHING of Crafty code in R1.0 beta, because you haven't looked. Still, this is in the ICGA verdict.

So once again:

I think you owe Vas an apology because of the totally false claim that Rybka plagiarized Crafty in ICGA events.
Parent - - By bob (Gold) Date 2013-09-13 15:56
Sorry.  I looked at _A_ version of 1.0 beta.  I saw the bishop shift tables.  I was not looking for my code specifically so I looked no further, I was looking to answer Ed's "challenge question" specifically.

Really is that simple.
Parent - By Venator (Silver) Date 2013-09-13 18:53
We asked you more than once to come up with proof that there is Crafty code in Rybka versions that played in ICGA tournaments. You failed to do so and you are even ducking the question.

The ICGA claims that Vas plagiarized both Fruit and Rybka in ICGA tournaments.

Now that you are unable to come up with copied Crafty code in a Rybka version that played in an ICGA tournament, you have to apologize to Vas, as the ICGA claim is just bogus and totally false.

It is not that hard. But you are a sissy who prefers to dive, divert attention, or change the subject just not having to admit that you are wrong.
Parent - - By Banned for Life (Gold) Date 2013-09-12 04:24
It is very important when YOU write a verdict that says that Rybka plagiarized Crafty and Fruit that Rybka actually plagiarized Crafty AND Fruit. You can cover yourself on the Fruit aspect, because Rybka clearly uses many ideas taken from Fruit (although you obviously lied about the "large chunks of copied code"), but there is no indication that anything in any Rybka that competed in WCCC events came from Crafty.
Parent - - By turbojuice1122 (Gold) Date 2013-09-12 10:50
Has anyone contacted Vas over the possibility of a libel suit?  Some (much!) of what Bob has said recently has been self-implication--is Vas aware of all of this?  Does he even care??
Parent - - By Banned for Life (Gold) Date 2013-09-12 20:28
I think Vas has considered it, and there is a good chance he could win, but getting a recovery from Bob would be a different matter. As they say, you can't get blood from a turnip. Normally one would go after the parent organization, in this case the ICGA, but this is so-called non-profit is not recognized as a legal entity in any country.
Parent - - By bob (Gold) Date 2013-09-12 21:03
dream on, that is ALL you seem capable of doing...
Parent - - By Banned for Life (Gold) Date 2013-09-13 03:03
As I already stated, you are most likely judgement proof.
Parent - By bob (Gold) Date 2013-09-13 04:04
But NOT for the reason you imagine..

As I said, "dream on"...
Parent - - By bob (Gold) Date 2013-09-12 18:35
1.  We certainly can see the hash table code directly from Rybka, which exactly matches the code from Fruit.  Accidental?  Hardly.  Of course you resort to "that's an algorithm" or "that's a formula" or "that's an idea" or "form follows function" or any of a dozen other worn-out and inapplicable excuses.

2.  I have already posted that I will absolutely look at 1.0 beta to see if the bitboard stuff matches Crafty.  Not that such will mean much here since we go right back to "that is a table, and a formula, and an indexing scheme" and such.
Parent - - By Banned for Life (Gold) Date 2013-09-12 20:24
1.  We certainly can see the hash table code directly from Rybka, which exactly matches the code from Fruit.  Accidental?  Hardly.  Of course you resort to "that's an algorithm" or "that's a formula" or "that's an idea" or "form follows function" or any of a dozen other worn-out and inapplicable excuses.

We've already seen, on multiple occasions, that you are not smart enough to differentiate algorithm from code. This culminated in your moronic assertion that taking notes was equivalent to code copying. The TT code in Rybka is certainly very similar to that in Fruit, with the non-minor exception that Vas broke the processing string up into four separate cases, as detailed by Richard, and only the pv move found case (trans_store_exact) was semantically equivalent. There is too little going on in this code segment to know if the code is actually the same.

It isn't clear if Richard will return, but this is really slim pickings. Where are the "large chunks of copied code" that you lied about finding? You get four Pinocchios for that statement.

It should also be noted that the TT implementation for the WCCC contestants, which Richard also disclosed, is very different, and requires a lot more stuff due to the small (8-byte) records.

2.  I have already posted that I will absolutely look at 1.0 beta to see if the bitboard stuff matches Crafty.

Why bother? You've already made this bogus claim in thousands of forum posts. Why bother checking to see if it is true or not? That's not the way you roll. It's just another example of your seeing "large chunks of copied code" in Rybka that have never been found by anyone else.
Parent - - By bob (Gold) Date 2013-09-12 21:02
As I said, ANYTHING that is copied from Fruit to Rybka is "an algorithm".  Even when the code matches instruction by instruction, it is ALWAYS an algorithm.  Or a formula.  Or a table.  Or an indexing scheme.  Or an idea.  Apparently rybka has ZERO code...
Parent - By Ugh (*****) Date 2013-09-12 21:24
Of course Rybka has code, it's just not the same code that Fruit has.

Bob, you know what goes on in an evaluation function. The software interrogates the board data, it grabs board data, then it processes that data to try and extract some chess knowledge out of it. The chess knowledge is usually some number or other, and that number is translated into a score and added in. Basically, and very generalised.

The problem for your copying theory is the massively different board data representations of Rybka and Fruit. In general, the act of interrogating and grabbing the board data and then processing it for chess information have to be done very differently. You have always claimed that after "abstracting" or removing those differences, the remaining code matches. But you have a problem, in general there is NO remaining code, all the work was being done differently in the abstracted code you had to remove.

The code that did the work was very different, usually algorithmically different as well (loops in Fruit, logic ops in Rybka). For most people this makes them DIFFERENT programs. Do you understand the problem we have with you now? We just don't see this "copied code" of yours, unsurprisingly, because it is not there.
Parent - - By Banned for Life (Gold) Date 2013-09-12 21:46
We're still waiting to see the "huge chunks of copied code" in Rybka. You saw them before the start of the investigation. Where did they go?
Parent - - By bob (Gold) Date 2013-09-12 22:16
The evaluation is "a huge chunk of code."
Parent - By Banned for Life (Gold) Date 2013-09-12 22:51
Yes, and none of it has been shown to be copied.
Parent - By Venator (Silver) Date 2013-09-11 06:54
Never seen anyone stoop so low. Continual distortion.

A perfectly normal question by user923005: show us the Crafty code in Rybka versions that played in ICGA tournaments. A question which you sofar failed to answer.

All you come up with is the above nonsense. You've lost the argument. You have lied. There is no such code. And you know it.
Parent - - By Venator (Silver) Date 2013-09-11 06:57
Ever looked at Fruit?  Ever looked at Crafty?  Ever compared the two?  Vastly different overall structure.  Rybka can't match BOTH.

Then how can it be that Vas plagiarized BOTH Crafty AND Fruit in Rybka versions that played in ICGA tournaments?
Parent - - By bob (Gold) Date 2013-09-11 13:38
So I can't take another program and copy 100 lines of code from one place, and still call that copying?  Don't copy a bunch of procedures, nor the program overall structure, just 100 consecutive lines of code. 

That's such a stupid question...
Up Topic Rybka Support & Discussion / Rybka Discussion / Rybka can't match both Fruit and Crafty
1 2 Previous Next  

Powered by mwForum 2.27.4 © 1999-2012 Markus Wichitill