Thoughts on Fruit=Rybka EVAL

[Rebel]

A lawyer with Rybka glasses will read that different. Seriously, check your own words. Here we can talk reasonable although being opponents, in court it's war and perception is all.

Here is the quotation:

Correction: Here is a quotation:

[...] If there were nothing but the PST evidence for Fruit/Rybka, I might say that [if successful in proving this] FL could could claim some piddly (<1%, perhaps <0.1%) amount of Rybka profitability under the heading of copyright infringement. OTOH, I would continue to insist that such "piddly" amounts are indeed "actionable" under the law [...]

When the PST issue started on numerous occassions the PST issue was downplayed by you as being not so important while the document implies the opposite, that is was a major. Allow me to quote from your own document,

2 Outline of the evidence
There are various major points of evidence between Fruit 2.1 and Rybka 1.0
Beta, and a number of minor and/or more circumstantial ones. The major points of evidence include:
ˆ the use of exactly the same evaluation features;
ˆ the identical ordering of operations at the root node in the search;
ˆ the same type of PST-scheme, re-using the identical File/Rank/Line weights.

7 Conclusion
This document has highlighted a number of places where Rybka 1.0 Beta can
be said to have over-stretched an \originality" barrier with respect to Fruit 2.1.
These include a borrowing of arrays in PST,

Why downplay it now ?

==============================================================================

Zach on the PST issue, I am quoting from his report,

Conclusion
We have found that, looking at the PST values of Fruit and Rybka, that Rybka's PSTs can be calculated using Fruit's
code with a minimum of changes. The only differences are the various weights (the constants found near the top of
pst.cpp in Fruit) and the bonuses for center pawns. Because of Fruit's unique PST initialization code, the origin of
Rybka's PSTs in Fruit is clear.

[hyatt]

Second, you really blew it with "A lot of the evidence hasn't even been written up properly". That turns into "A lot of the evidence is improper and worthless."

I agree with you.

The impossible just happened

Zach's honesty is a gift from heaven for any lawyer.

What? That he said "I had a lot more that was _not_ written up, but if you need it I could do so?" Sounds like a lawyer's nightmare rather than his dream, seeing that the pile of evidence can grow even larger.

[Rebel]

Second, you really blew it with "A lot of the evidence hasn't even been written up properly". That turns into "A lot of the evidence is improper and worthless."

I agree with you.

The impossible just happened

Zach's honesty is a gift from heaven for any lawyer.

What? That he said "I had a lot more that was _not_ written up, but if you need it I could do so?" Sounds like a lawyer's nightmare rather than his dream, seeing that the pile of evidence can grow even larger.

Funny, I basically agree what you said, you now disagree with yourself?

You forget that in fora, although being each others opponents, even the die hard VII Rybka fans accepted Zach's posting, showing respect recognizing the right sentiment and tone setting. In a courtroom that is all gone. Fabien will face a supposed angry lawyer with dark pink Rybka glasses and portrait Fabien as a ........ I leave that out ........ and how the hell Fabien could possibly ......... I leave that out .......... use your own imagination. Keywords: In retrospect, I regret, it's fuzzy in many ways, it's circumstantial, not properly, I view this as my own failure in many ways.

The man is doing what he is paid for, to win. Ethics and morals don't count in court, it's about winning or losing and all bets are off.

[hyatt]

It is also about the evidence. More about that than any gerrymandering by an attorney. Zach's statement that he only published a part of what he had prepared doesn't sound like an attorney's dream to me... Sounds like a nightmare.

[veritas]

Second, you really blew it with "A lot of the evidence hasn't even been written up properly". That turns into "A lot of the evidence is improper and worthless."

I agree with you.

The impossible just happened

Zach's honesty is a gift from heaven for any lawyer.

What? That he said "I had a lot more that was _not_ written up, but if you need it I could do so?" Sounds like a lawyer's nightmare rather than his dream, seeing that the pile of evidence can grow even larger.

Funny, I basically agree what you said, you now disagree with yourself?

You forget that in fora, although being each others opponents, even the die hard VII Rybka fans accepted Zach's posting, showing respect recognizing the right sentiment and tone setting. In a courtroom that is all gone. Fabien will face a supposed angry lawyer with dark pink Rybka glasses and portrait Fabien as a ........ I leave that out ........ and how the hell Fabien could possibly ......... I leave that out .......... use your own imagination. Keywords: In retrospect, I regret, it's fuzzy in many ways, it's circumstantial, not properly, I view this as my own failure in many ways.

The man is doing what he is paid for, to win. Ethics and morals don't count in court, it's about winning or losing and all bets are off.

reads like your doing what your paid for and not taking good advice re your needed medication

think its a case of rose colored rybka spectacles , red herrings and straw clucthings as only a retarded pot smoker would fail to understand vas broke simple rules and continued to be an arrogant ignorant fool who thinks the pedastal he was placed on was built on rock , not the sands of time
whats your dream eddie a complete break away , a new wcc where only vas and his minions can play for fake titles ?

[Jeremy Bernstein]

Second, you really blew it with "A lot of the evidence hasn't even been written up properly". That turns into "A lot of the evidence is improper and worthless."

I agree with you.

The impossible just happened

Zach's honesty is a gift from heaven for any lawyer.

What? That he said "I had a lot more that was _not_ written up, but if you need it I could do so?" Sounds like a lawyer's nightmare rather than his dream, seeing that the pile of evidence can grow even larger.

Funny, I basically agree what you said, you now disagree with yourself?

You forget that in fora, although being each others opponents, even the die hard VII Rybka fans accepted Zach's posting, showing respect recognizing the right sentiment and tone setting. In a courtroom that is all gone. Fabien will face a supposed angry lawyer with dark pink Rybka glasses and portrait Fabien as a ........ I leave that out ........ and how the hell Fabien could possibly ......... I leave that out .......... use your own imagination. Keywords: In retrospect, I regret, it's fuzzy in many ways, it's circumstantial, not properly, I view this as my own failure in many ways.

The man is doing what he is paid for, to win. Ethics and morals don't count in court, it's about winning or losing and all bets are off.

reads like your doing what your paid for and not taking good advice re your needed medication

think its a case of rose colored rybka spectacles , red herrings and straw clucthings as only a retarded pot smoker would fail to understand vas broke simple rules and continued to be an arrogant ignorant fool who thinks the pedastal he was placed on was built on rock , not the sands of time
whats your dream eddie a complete break away , a new wcc where only vas and his minions can play for fake titles ?

Um. No. Veritas, you have been warned previously for this kind of stuff. You are hereby banished to the Ippolit forum for the next 2 weeks. *poof*.

jb

[BB+]

[...] Since I've repeatedly made this point to no avail at this point I would suggest that you ask Panel members like Gerd Isenberg, Don Dailey, and Ken Thompson [...] whether there was "too little opposition", and whether they (and others) "overlooked a couple aspects of chess programming".

Why take (almost) everything I say so personal? Why this defensive attitude?

One of the contexts was whether the Panel might have "overlooked a couple aspects of chess programming". You have mused (previously) that I don't understand chess engine development as well as others, since I am not a chess programmer. Thus, I referred you to those who might know more. Similarly, you seem not to believe it when I assert there was notable opposition in the Panel, so I again I referred you to those who might be able to offer another opinion. As I've said before, we may just be having cultural differences in the way that discussions/arguments are carried out.

I will say (again) that I find the evidence with EVAL order to be quite debatable either way. If the evidence were clear, then I agree it would be much more relevant.

But the evidence is clear and not debatable at all. You and Zach have stated Fruit EVAL order is VERY similar to Rybka EVAL order.

I never stated that. I said that one could argue that the EVAL order in Rybka and Fruit is not much different than one would expect from external factors like bitboards and performance issues -- and one could possibly argue the opposite. I then re-stated this (see middle of post: I re-iterate [...] that the evidence says little one way or the other), and then gave rhetorical opposition to your pre-emptive claim/conclusion that they weren't too much alike.

It's a fact the PST's are under pressure, and I am being mild here.

I reiterate my PST challenge, seen here. You can also argue that my challenge is irrelevant or misguided, of course. I personally think that "amount of code changes necessary" is quite a good metric for "substantial similarity" and/or "originality".

Here is the quotation:

Correction: [you meant] Here is a quotation: [...]
When the PST issue started on numerous occassions the PST issue was downplayed by you as being not so important while the document implies the opposite, that is was a major. Allow me to quote from your own document, [...]

I gave you the quotation from which "1% for the purposes of copyright infringement" was derived. You gave me quotations that say PST is a "major" issue for the ICGA "originality" standard. I don't see how these are in conflict, as they are different standards. Copyright law, when infringement is shown, then still has to deal with percentages. The ICGA Rule #2 doesn't necessarily follow this pattern, though "small percentages" could be a mitigating element. A related historical issue is that the Hsu/Berliner "Cray Blitz simulator" dispute was over about 0.3% of the former's code. The EvaluateWinner() code from Crafty that is in Rybka 1.6.1 (to take the most voluminous example there, being ~100 lines of codes) is probably around 1% of either codebase, yet I don't think one would argue that this "smallness" changes the situation much.

If I have "downplayed" PST, it was in the context of Richard Vida's statement (and Miguel Ballicora's seeming agreement) that there was not a vast amount of chess knowledge therein. I agreed that one could argue (as did he) that the ICGA should view PST as so mundane as to irrelevant (like magic bitboards) -- but I also noted that I personally at least considered PST to be sufficient "chess playing code" to count for ICGA Rule #2. Again, other programmers might have a different opinion -- I won't deign to speak for the Panel, and you know who is involved there, so I don't need to "name-drop" them anymore.

[Rebel]

Alright then, let's keep it simple. Same question for you as to Zach in Rybka forum.

1. Do you totally reject Miguel's PST experiments? Is it all bull, or is there some truth in it?

And a few extra ones.

2. Do you still stand by your document that on 2 places states that the PST's are a MAJOR issue?

3. What's your personal opinion, in the light of all that has been said about the PST's here and at Rybka forum, should the panel had to pay more attention to the PST issue?

[Rebel]

One of the contexts was whether the Panel might have "overlooked a couple aspects of chess programming". You have mused (previously) that I don't understand chess engine development as well as others, since I am not a chess programmer. Thus, I referred you to those who might know more. Similarly, you seem not to believe it when I assert there was notable opposition in the Panel, so I again I referred you to those who might be able to offer another opinion. As I've said before, we may just be having cultural differences in the way that discussions/arguments are carried out.

Well, it's my conclusion the panel missed the PST's. And if I label that as not enough opposition it's related to topic and not to people. You picked the latter making it personal. Wrong. Don't do it again.

[hyatt]

Alright then, let's keep it simple. Same question for you as to Zach in Rybka forum.

1. Do you totally reject Miguel's PST experiments? Is it all bull, or is there some truth in it?

And a few extra ones.

2. Do you still stand by your document that on 2 places states that the PST's are a MAJOR issue?

3. What's your personal opinion, in the light of all that has been said about the PST's here and at Rybka forum, should the panel had to pay more attention to the PST issue?

Since I've been involved in the discussion about that with Miguel, I see two takes on the issue.

(1) It is POSSIBLE, that one can use such a simple PST concept that others might not have much trouble duplicating it unintentionally. But with one important qualifier. What is the probability that others will use a purely simple PST concept in the first place? I've given several counter examples to show that simply "move piece toward center" is not the only way. There are others. Move piece towards center and then toward opponent's king. Or move piece toward center of remaining pawn mass. Or move piece toward weak squares it can occupy and not be driven away. Or move piece toward squares where it attacks more squares around the opponent's king. Or any combination of those. So if you buy into the basic premise that "everyone is likely to use the simplest PST idea of all, namely centralization, then Miguel has a point.

In his Stockfish analysis, I disagree with him. Because one component of the stockfish PST initialization is similar. But others are not. For example, try any fudging you want to Fruit's constants to produce the stockfish values. Because stockfish adds piece material values to the squares. That is actually a more important part of the PST value than the centralization bonus. So you have to remove 1/2 of the "content" of each square before you can play with multipliers to get a match. That is far less "close" that what one would find in looking at Rybka. Cray Blitz and early Crafty's (before root preprocessing was eliminated) factored in attracting the pieces toward the king as well as toward the center. Cray Blitz once had 2 16 bit values folded into one PST entry. Use low order 16 bits always, add in high-order 16 bits (shifted right 16 bits obviously) if the square was a good outpost and the opponent couldn't drive the piece away.

In short, I am not a big fan of stripping everything that is different out, just to show that there is this one similar piece (the base centralization values in SF, for example, ignoring the changes that were necessary in the code to add in the piece material scores, and the performance gain that might produce.

(2) I disagree that everyone will use the "simplest PST format possible." Where does that conclusion come from? All of my code serves as counter-examples. One PST is close (different origin/offset). Others are vastly different. So I have not seen anything to support this "many will go simple..." I've asked for an example program, not known to be a fruit derivative, where all the PSTs show the same kind of similarity as is shown in fruit/rybka. That is, the PSTs can be produced ONLY by changing the fruit initialization constants. No code can be added or removed. I'll not even count the occasional minor differences where someone might add a unique bonus for e4/e5, so we will not count such a tiny change as "different" at all. So, is there such a program? If many use the "easy way" there should be "many" that are similar. Where? One will do for a start. The more, the merrier, and the better his point will look. Right now the only such example we have is Fruit/Rybka. The stockfish example is a stretch.

When we had this discussion in the panel, I did look at several programs. I found no example where I could take Fruit's PST initialization code and produce the PSTs for another program by only changing the constant multipliers (not the vectors). Is there such a program? I certainly did not prove there was not because I didn't try to download the source of every program around. All I can say is that I could not find one. That's the reason I had all the fruit PST code ready to post here so quickly, I spent some time looking at that when it first came up from Zach, way back, when he found the "magic numbers" to convert fruit to Rybka. And I then used that code in comparing fruit's values to other programs while the panel discussions moved along. It is simply one damning piece of evidence out of many damning pieces of evidence.

One might argue that "OK, they took Fruit's PST initialization and changed it a bit to obfuscate the origin of their PST values. And that would be possible. But Vas didn't even go that far, as the report shows.