FIDE Rules on ICGA - Rybka controversy

[hyatt]

I can give a highly educated guess for the reason. Take, for example, the discussion about semantic equivalence. I explained it carefully. The simplest form is a function is semantically equivalent to another function if both produce the SAME output given the SAME input. A more rigorous definition would be "two functions that produce the same output, using the same exact operations, given the same input."

I went on to take an example from fruit and an example bit of ASM from rybka and showed how they matched up perfectly, the ONLY difference being how specific tests are done. To discover if a pawn is passed in a bit board program, one single AND with the right bit mask will do the trick. For a mailbox program, one has to loop over the squares in the forward direction. To use your AFC idea, one takes those two pieces of code and abstracts away JUST the bit board/mailbox differences and then asks "do these do either (a) exactly the same tests or (b) mostly the exact same tests with just a small part changed or added?"

You took that to a new universe in your top-500 nonsense. You claimed that since Fruit was interpolating evaluation between MG/EG, as were most other programs, there was "semantic equivalence all over the place". NOBODY applies semantic equivalence to one line of code out of 50 or 100. NOBODY.

I would imagine Mark simply became frustrated when every conversation keeps retreating to the same tired old points, and not adding anything new to the discussion. It does get old. A compiler person reading your semantic equivalence everywhere nonsense would simply shake his head and wonder "what is wrong with him?"

BTW _I_ am wondering "what is wrong with him?"

BTW any sort of phone conversation would be HIGHLY ill-advised. Later. "I did NOT say that." "YES you did." Etc. At LEAST email leaves a footprint that can't easily be wiped away. No court would let someone testify by phone for that very reason.

The phone call was obviously for initial discussion purposes, to investigate possible solutions/whatever. It is unlikely Vas was imagining evidence and counter evidence arguing over the phone. Use your head, man.

A common mistake you made. "obviously for initial discussion purposes." How is that "obvious"? Yet another assumption. How do you know it wasn't his intent to discuss this on the phone and resolve it at that time if he could? You guys twist written words enough to make discussion almost impossible. What would happen with a VERBAL discussion? I don't buy this "obvious", "clearly" and such. I buy actual concrete evidence, and I do NOT try to "read between the lines" as that is pure imagination.

This discussion needs a lot more fact, and a lot less opinion. Or it will continue to go nowhere at all, just around and around the same circles forever.

[Chris Whittington]

I do complain about the nonsense that does nothing but waste time. We already know the following:

(1) the "damaged" person here is vas. He is the ONLY person with an interest at stake in this case. He is the ONLY person that can undertake any sort of action, whether it be contacting the ICGA, complaining to FIDE or initiating a legal action. No one else can do so except for him.

(2) The only two people keeping this discussion going is You/CW. Maybe 2-3 on RF in addition, but they only erupt when the various keywords "ICGA', "Levy", "Hyatt", "Lefler", "Williamson" and maybe WCCC appear. I don't see the point because you have not made any progress, and can't without Vas' direct assistance, which seems to NOT be forthcoming.

(3) The ICGA is only going to listen to Vas. The FIDE EC asked (not insisted) that they reconsider the life-time ban. I think it quite likely that the ICGA board might consider modifying that to what the panel recommended, namely "not allowed to participate until the current version of Rybka is judged to be original by an agreeable third party."

"The definition of insanity is doing the same thing over and over and expecting a different result eventually."

This discussion pretty much fits that definition. Nothing will change until Vas acts.

You, Willamson and Levy are all disbarred by bias from deciding to grant an appeal/review or anything else. A verdict reversal opens immediately the question for damages arisng from the earlier false verdict and the name and shame Yoir personal interest is to never grant any form of appeal and carry on claiming matter closed. Your icga has nobody who can be appealed to, nor can your icga, for the same reasons of bias, participate in any ruling over an appeal. The best you can do is contact Vas representatives to determine if there is any way round this little problem that you have created for yourselves. Meanwhile la luta continua.

[Rebel]

Now 58% similarity is a very poor number for 2 identical evaluations.

The actual numbers, as noted in the middle of page 19, were 56% and 60% similarity [see my email mid-April 2012]. Further, again as noted in the paper, having identical evaluation functions is only part of the picture, as by making search changes the move selections could vary (cf. Thinker).

Sure, percentages may fluctuate a few percent max due to the PC speed and the time control used. So I will give you the 2 percent similarity. And 60% is way too less for 2 identical evaluations, the Loop example (BTW, I didn't get Adam's extreme %) and the various Ippo derived engines, see for example my SIMEX pages that evaluates the pro's and con's of Don's (genius) similarity tool and look (for example) at the top-100 list.

Regarding the influence of search, it's surprising low in the sense one would expect and as a next surprise the longer time you give an engine the higher the similarity of a suspect program becomes. I believe Adam concluded similar.

I would love to hear the opinion of other chess programmers if 60% similarity between 2 identical evaluations (following my idea and your work) isn't actually a bomb under the VIG verdict. It's only that after 4 years I wonder if anyone would care.

[Chris Whittington]

I can give a highly educated guess for the reason. Take, for example, the discussion about semantic equivalence. I explained it carefully. The simplest form is a function is semantically equivalent to another function if both produce the SAME output given the SAME input. A more rigorous definition would be "two functions that produce the same output, using the same exact operations, given the same input."

I went on to take an example from fruit and an example bit of ASM from rybka and showed how they matched up perfectly, the ONLY difference being how specific tests are done. To discover if a pawn is passed in a bit board program, one single AND with the right bit mask will do the trick. For a mailbox program, one has to loop over the squares in the forward direction. To use your AFC idea, one takes those two pieces of code and abstracts away JUST the bit board/mailbox differences and then asks "do these do either (a) exactly the same tests or (b) mostly the exact same tests with just a small part changed or added?"

You took that to a new universe in your top-500 nonsense. You claimed that since Fruit was interpolating evaluation between MG/EG, as were most other programs, there was "semantic equivalence all over the place". NOBODY applies semantic equivalence to one line of code out of 50 or 100. NOBODY.

I would imagine Mark simply became frustrated when every conversation keeps retreating to the same tired old points, and not adding anything new to the discussion. It does get old. A compiler person reading your semantic equivalence everywhere nonsense would simply shake his head and wonder "what is wrong with him?"

BTW _I_ am wondering "what is wrong with him?"

BTW any sort of phone conversation would be HIGHLY ill-advised. Later. "I did NOT say that." "YES you did." Etc. At LEAST email leaves a footprint that can't easily be wiped away. No court would let someone testify by phone for that very reason.

The phone call was obviously for initial discussion purposes, to investigate possible solutions/whatever. It is unlikely Vas was imagining evidence and counter evidence arguing over the phone. Use your head, man.

A common mistake you made. "obviously for initial discussion purposes." How is that "obvious"? Yet another assumption. How do you know it wasn't his intent to discuss this on the phone and resolve it at that time if he could? You guys twist written words enough to make discussion almost impossible. What would happen with a VERBAL discussion? I don't buy this "obvious", "clearly" and such. I buy actual concrete evidence, and I do NOT try to "read between the lines" as that is pure imagination.

This discussion needs a lot more fact, and a lot less opinion. Or it will continue to go nowhere at all, just around and around the same circles forever.

Well, i suppose if you had business or commercial experience you would know how dispute negotiations would work. I can't help your lack of experience/knowledge. Email discussions are far too open to misinterpretation and often work out quite unfriendly and break down, particularly if there is a dispute. A telephoen call is often a way to establish the basis for a better communication. Your Levy turned this down. He likes to dispense "justice" by email, it seems.

[hyatt]

Please use the correct term. NOT "disbarred". A definition:

expel (a lawyer) from the Bar, so that they no longer have the right to practice law None of us are lawyers so there is no way our license to practice law can be revoked, since we don't have one.

There will be NO "verdict reversal". There is a possibility of a change in the penalty. Nothing more. The ICGA won't be contacting the "Vas representatives" since there are none legally. Unless you have a signed power of attorney from him. Vas is free to contact the ICGA of course, something he has not done to date.

[hyatt]

I can give a highly educated guess for the reason. Take, for example, the discussion about semantic equivalence. I explained it carefully. The simplest form is a function is semantically equivalent to another function if both produce the SAME output given the SAME input. A more rigorous definition would be "two functions that produce the same output, using the same exact operations, given the same input."

I went on to take an example from fruit and an example bit of ASM from rybka and showed how they matched up perfectly, the ONLY difference being how specific tests are done. To discover if a pawn is passed in a bit board program, one single AND with the right bit mask will do the trick. For a mailbox program, one has to loop over the squares in the forward direction. To use your AFC idea, one takes those two pieces of code and abstracts away JUST the bit board/mailbox differences and then asks "do these do either (a) exactly the same tests or (b) mostly the exact same tests with just a small part changed or added?"

You took that to a new universe in your top-500 nonsense. You claimed that since Fruit was interpolating evaluation between MG/EG, as were most other programs, there was "semantic equivalence all over the place". NOBODY applies semantic equivalence to one line of code out of 50 or 100. NOBODY.

I would imagine Mark simply became frustrated when every conversation keeps retreating to the same tired old points, and not adding anything new to the discussion. It does get old. A compiler person reading your semantic equivalence everywhere nonsense would simply shake his head and wonder "what is wrong with him?"

BTW _I_ am wondering "what is wrong with him?"

BTW any sort of phone conversation would be HIGHLY ill-advised. Later. "I did NOT say that." "YES you did." Etc. At LEAST email leaves a footprint that can't easily be wiped away. No court would let someone testify by phone for that very reason.

The phone call was obviously for initial discussion purposes, to investigate possible solutions/whatever. It is unlikely Vas was imagining evidence and counter evidence arguing over the phone. Use your head, man.

A common mistake you made. "obviously for initial discussion purposes." How is that "obvious"? Yet another assumption. How do you know it wasn't his intent to discuss this on the phone and resolve it at that time if he could? You guys twist written words enough to make discussion almost impossible. What would happen with a VERBAL discussion? I don't buy this "obvious", "clearly" and such. I buy actual concrete evidence, and I do NOT try to "read between the lines" as that is pure imagination.

This discussion needs a lot more fact, and a lot less opinion. Or it will continue to go nowhere at all, just around and around the same circles forever.

Well, i suppose if you had business or commercial experience you would know how dispute negotiations would work. I can't help your lack of experience/knowledge. Email discussions are far too open to misinterpretation and often work out quite unfriendly and break down, particularly if there is a dispute. A telephoen call is often a way to establish the basis for a better communication. Your Levy turned this down. He likes to dispense "justice" by email, it seems.

I have quite a bit of business experience in fact. And one rule we ALWAYS followed was EVERYTHING is done in writing. Court proceedings are done verbally but every word is recorded by the court reporter to provide indisputable proof of who said what and when. Email is "in writing" and is certainly far superior to a phone conversation in terms of accuracy and preservation of exact statements.

[hyatt]

Now 58% similarity is a very poor number for 2 identical evaluations.

The actual numbers, as noted in the middle of page 19, were 56% and 60% similarity [see my email mid-April 2012]. Further, again as noted in the paper, having identical evaluation functions is only part of the picture, as by making search changes the move selections could vary (cf. Thinker).

Sure, percentages may fluctuate a few percent max due to the PC speed and the time control used. So I will give you the 2 percent similarity. And 60% is way too less for 2 identical evaluations, the Loop example (BTW, I didn't get Adam's extreme %) and the various Ippo derived engines, see for example my SIMEX pages that evaluates the pro's and con's of Don's (genius) similarity tool and look (for example) at the top-100 list.

Regarding the influence of search, it's surprising low in the sense one would expect and as a next surprise the longer time you give an engine the higher the similarity of a suspect program becomes. I believe Adam concluded similar.

I would love to hear the opinion of other chess programmers if 60% similarity between 2 identical evaluations (following my idea and your work) isn't actually a bomb under the VIG verdict. It's only that after 4 years I wonder if anyone would care.

Where is this "identical evaluations" coming from? The ICGA report did not claim that Rybka 1.0 beta and Fruit had identical evaluations. The report and evidence showed that parts of the evaluations were identical. And even the report and evidence pointed out where some things were changed. Even the simple PST stuff changed 4 squares in the central files. So why would you want to use that "identical" term here?

[Chris Whittington]

Please use the correct term. NOT "disbarred". A definition:

expel (a lawyer) from the Bar, so that they no longer have the right to practice law None of us are lawyers so there is no way our license to practice law can be revoked, since we don't have one.

There will be NO "verdict reversal". There is a possibility of a change in the penalty. Nothing more. The ICGA won't be contacting the "Vas representatives" since there are none legally. Unless you have a signed power of attorney from him. Vas is free to contact the ICGA of course, something he has not done to date.

We do. Power of attorney, as I understand it, means we can a ct without consulting him. In the UK, those are usually reserved for sons or daughters acting for a parent unable to make decisions, and that is inappropriate for this case. Ed and I are appointed negotiators/advisors for matters arising out of icga-Rybka affair, subject to consulting and referrign back to Vas, obviously. If your principal wishes to negotiate we can provide him with the relevent document. You and Williamson have shown yourselves disabbred as negotiators through lack of trust. You both repeatedly offered an "appeal" if Vas contacted direct, only for us to be told by Levy thee was no possibilty of appeal whether Vas contacted or not. Thus, what you have to say on the matter is not very relevent. Understand?

[Rebel]

You did not comment on my Assange / Snowden post.

Do I have to comment on every fanciful political parallel that you concoct?

I easily justify myself as a whistleblower

Whistleblowers are usually for things in the public interest, often with (gasp) actual governments involved, and affecting significant sections of a population. I fail to see how the ICGA situation (relevant to maybe 100 people at best, of the highly esoteric "computer chess" special-interest group) is much comparable. There is much more of a "public interest" in being trustworthy and keeping private things private.

Apply the same logic to Levy, he (as whistleblower) posted private email of Vas in his "Attack of the Clones" on Chessvibes.

[Chris Whittington]

I can give a highly educated guess for the reason. Take, for example, the discussion about semantic equivalence. I explained it carefully. The simplest form is a function is semantically equivalent to another function if both produce the SAME output given the SAME input. A more rigorous definition would be "two functions that produce the same output, using the same exact operations, given the same input."

I went on to take an example from fruit and an example bit of ASM from rybka and showed how they matched up perfectly, the ONLY difference being how specific tests are done. To discover if a pawn is passed in a bit board program, one single AND with the right bit mask will do the trick. For a mailbox program, one has to loop over the squares in the forward direction. To use your AFC idea, one takes those two pieces of code and abstracts away JUST the bit board/mailbox differences and then asks "do these do either (a) exactly the same tests or (b) mostly the exact same tests with just a small part changed or added?"

You took that to a new universe in your top-500 nonsense. You claimed that since Fruit was interpolating evaluation between MG/EG, as were most other programs, there was "semantic equivalence all over the place". NOBODY applies semantic equivalence to one line of code out of 50 or 100. NOBODY.

I would imagine Mark simply became frustrated when every conversation keeps retreating to the same tired old points, and not adding anything new to the discussion. It does get old. A compiler person reading your semantic equivalence everywhere nonsense would simply shake his head and wonder "what is wrong with him?"

BTW _I_ am wondering "what is wrong with him?"

BTW any sort of phone conversation would be HIGHLY ill-advised. Later. "I did NOT say that." "YES you did." Etc. At LEAST email leaves a footprint that can't easily be wiped away. No court would let someone testify by phone for that very reason.

The phone call was obviously for initial discussion purposes, to investigate possible solutions/whatever. It is unlikely Vas was imagining evidence and counter evidence arguing over the phone. Use your head, man.

A common mistake you made. "obviously for initial discussion purposes." How is that "obvious"? Yet another assumption. How do you know it wasn't his intent to discuss this on the phone and resolve it at that time if he could? You guys twist written words enough to make discussion almost impossible. What would happen with a VERBAL discussion? I don't buy this "obvious", "clearly" and such. I buy actual concrete evidence, and I do NOT try to "read between the lines" as that is pure imagination.

This discussion needs a lot more fact, and a lot less opinion. Or it will continue to go nowhere at all, just around and around the same circles forever.

Well, i suppose if you had business or commercial experience you would know how dispute negotiations would work. I can't help your lack of experience/knowledge. Email discussions are far too open to misinterpretation and often work out quite unfriendly and break down, particularly if there is a dispute. A telephoen call is often a way to establish the basis for a better communication. Your Levy turned this down. He likes to dispense "justice" by email, it seems.

I have quite a bit of business experience in fact. And one rule we ALWAYS followed was EVERYTHING is done in writing. Court proceedings are done verbally but every word is recorded by the court reporter to provide indisputable proof of who said what and when. Email is "in writing" and is certainly far superior to a phone conversation in terms of accuracy and preservation of exact statements.

How do you negotiate "off the record"? How do you negotaite "what if" situations? I won't accuse you of being silly and claiming to know everything about everything, when you don't because that would be rude, but you get the idea. Your posts like this are just wasting space.