Frayer's opinion expressed at the Rybka forum....

[Rolf]

I think we have it now. Thanks so far.

I thought that until the fog isnt lifted you were not starting to sue Vas, but after that you would. Now I understand you. This all isnt about a sort of legally relevant failure but some sort of formal inexactitudiness without any legal consequences after such a long time, correct? If seen earlier you might have taken actions but no longer now, ist that true so far?

Again, actually Vas wont talk but as he confirmed others besides me he will talk after he got some things would become clearer.

Couldnt we just give him that time which seems to lay not in his realm and stop this climate of hate against Rybka?

I ask this for the sake of our own peace, not in any way to protect Vas. He has his own remedy and that is ignoring. But we are going nowhhere with this campaign. Also please consider that most here are younger and they might seek a sort of winner which isnt very likely to get. All IMO.

Ok, I am a total outsider, but from all what I know about peace of mind I would support if we would stop such a scapegoating. It shouldnt be allowed on CCC. BTW didnt Vas react positively for talking with you? Let's see what will happen.

All the best for a nice weekend for you all.

Rolf

here's my take on this: He has as long as he wants to make some sort of statement, either about fruit code in Rybka 1 or exactly what was copied from his code and included in robo*. But in fairness, I see absolutely no justification to continue to maintain that robo* is any sort of clone. There are several reasons:

(1) it is stronger (significantly stronger, +70 elo is _not_ easy to do in any program, without a ton of time and a ton of testing);

(2) we have seen absolutely no evidence that it is a clone. Just a statement. And I don't see how anyone could consider it fair that this just drags on and one with no evidence offered.

In light of that, the best solution is to treat Robo* like any other program. If it isn't a clone, good. If it is a clone, that will give Vas incentive to speak up before he drops off the top of all the rating lists. Either way, the truth comes out and that's all anyone really wants at the moment.

ad 1) I doubt the better strength, mostly there are made up tests without meaning because the tests are not made from a neutral angle but always what can actually our baby program do at best, this is such a bias, that it cant be taken for a serious testing result at all

ad 2) the striking argument against the thing is that its authors, if ever singular existences, are anonymous. That speaks for evil-doing in the past because they must hide to not getting into legal difficulties. Otherwise if someone would step forward, I would make a new judgement but actually we here cannot accept that such a good program is based on hiding and hidden authorship. Vas was from his beginning a known figure who stood for the project no matter who collaborated else. Until the Robbos dont open their shields they cant be seriously be taken into test, serverplays and tournaments for any title out there.

ad 2 NB)

if we would accept such anonymous entries we had to face next that say some real talents or titleholders themselves build their second or third families under anonymous accounts, perhaps they steal data from other codes and then participate as manipulative helpers of certain favorites, of themselves, please all may remember the problems a single artist like Bobby Fischer had to face when he had to compete against an armada of Soviet players who didnt play for their own fate so to speak but they were instruments to support the actually best runner for the title; they all resisted to beat each other while they played Fischer for real. If we accepted anonymity in our sport that whole scenario would be repeated. The singular fighters no longer had any chance at all. The Stalinists had their go.

Ad 2 PS)

It is a generally accepted truth that these Robbists are hurting computerchess as a community. And this should be the reason why we all should defend ourselves against these anonymous chicken heads who are dishonest and unethical. I called them chess cyber terrorists. I never saw someone oppose the verdict.

Robbos could become legal but if they insist to be illegal they cant be members of our community.

Rolf GENS UNA SUMUS MINUS CYBER TERRORISTS

[kingliveson]

ad 1) I doubt the better strength, mostly there are made up tests without meaning because the tests are not made from a neutral angle but always what can actually our baby program do at best, this is such a bias, that it cant be taken for a serious testing result at all

I posted 2 results yesterday where Rybka was beaten with different settings. You said you did not like the method, so I agreed to re-run the test. And keep in mind Rybka even had an unfair advantage in the initial tests/results I posted--It used ponder while RobboLito did not and yet RobboLito came out on top.

I ran the test again with your own suggestions and Robbolito came out +63 ELO ahead of Rybka, and now you lay accusation that these test are not fair or made up. I have posted the games and test conditions--anyone can analyze them to see whether there was something mischievous.

I don't have a dog in this fight. I have nothing personal against Vas or Rybka. If Rybka 4 was to be released today and it shows 30-50 ELO strength above Rybka 3 with the bugs fixed, I assure you I would be in line to make the purchase.

My issue with the current situation is that I have not heard Vas come out and say he has looked at these chess engines and have found similarities that make him believe this is a clone, or a copyright infringement--this is issue number 1.

Two, How is it possible that a chess engine (or anyone for that matter) can be condemned (assuming the accusation has been made), when no evidence of wrongdoing has been submitted after several months? How much longer do we wait?

Even Christ offered "doubting Thomas" evidence and proof.

[Rolf]

ad 1) I doubt the better strength, mostly there are made up tests without meaning because the tests are not made from a neutral angle but always what can actually our baby program do at best, this is such a bias, that it cant be taken for a serious testing result at all

I posted 2 results yesterday where Rybka was beaten with different settings. You said you did not like the method, so I agreed to re-run the test. And keep in mind Rybka even had an unfair advantage in the initial tests/results I posted--It used ponder while RobboLito did not and yet RobboLito came out on top.

I ran the test again with your own suggestions and Robbolito came out +63 ELO ahead of Rybka, and now you lay accusation that these test are not fair or made up. I have posted the games and test conditions--anyone can analyze them to see whether there was something mischievous.

I don't have a dog in this fight. I have nothing personal against Vas or Rybka. If Rybka 4 was to be released today and it shows 30-50 ELO strength above Rybka 3 with the bugs fixed, I assure you I would be in line to make the purchase.

My issue with the current situation is that I have not heard Vas come out and say he has looked at these chess engines and have found similarities that make him believe this is a clone, or a copyright infringement--this is issue number 1.

Two, How is it possible that a chess engine (or anyone for that matter) can be condemned (assuming the accusation has been made), when no evidence of wrongdoing has been submitted after several months? How much longer do we wait?

Even Christ offered "doubting Thomas" evidence and proof.

Please dont accuse me of something you dont want to be suffering yourself. I have a longer record than this short period of Robbos life in test design for chess computers and I have a rigid concept for such tests. Most of them are not sound IMHO.

Back in 1996 I showed that the SSDF, a private Swedish test board were NOT following usual rules. They took 16 games results here, some 20 there and another 7 there and then they added the results and got a rating number. IMO this is all wrong. But then like totay people normally could do what they want and after I have made my critic I step back, unless I am suddenly attacked for my criticism, of course then the whole debate goes on. Look, coming from science I know that you cant enforce the truth or best results unless they are accepted in general. So, I know that we must live with much imperfect and untrue stuff. No problem with that. I am not the guardian. If you read my critic and decide to do it like before then this is ok. I have no higher power.

Perhaps we have the chance to talk about it a bit deeper but until then do as usual. It's your hobby. But not science.

What I meant in my critic was mainly that it's ridiculous to weaken Rybka with no pondering only because the cloners have not yet found out how to implement that feature. Know what I mean? So, if you deactivate all the good things in Rybka you might get the results you reported. But they are not valide.

I would only ask you to always honestly report exactly what you have done in details. Then this is ok because others could then see it. Otherwise it were fraud. But until now I have no reason to accuse you of anything. In contrary, I like that you are communicating. That's what such a forum is made for.

All I would say that right now it's premature to say exactly how strong the Robbothing is. But that can change soon. Dont worry. As you know I have a total horror in front of anonymous people behind Robbos. But I would also speak with Stalin or the Mafia if that would go in civilized manner. No problem for me.

[kingliveson]

Ok, we can focus on this issue of testing only.

Please dont accuse me of something you dont want to be suffering yourself. I have a longer record than this short period of Robbos life in test design for chess computers and I have a rigid concept for such tests. Most of them are not sound IMHO.

You were quoted directly.

ad 1) I doubt the better strength, mostly there are made up tests without meaning because the tests are not made from a neutral angle but always what can actually our baby program do at best, this is such a bias, that it cant be taken for a serious testing result at all

And here is the conversation from yesterday regarding my initial test/results and your response.

100 games where Rybka used contempt setting of zero.

Code: Select all

1. RobboLito  +25/=58/-17	54.00%		54.0/100
2. Rybka  3   +17/=58/-25	46.00%		46.0/100

100 games where Rybka used contempt setting of 15 (default).

Code: Select all

1. Robbolito  +24/=53/-23	50.50%		50.5/100
2. Rybka  3   +23/=53/-24	49.50%		49.5/100

Total Result:

Code: Select all

1. RobboLito   +49/=111/-40  52.25%		104.5/200   
2. Rybka   3   +40/=111/-49  47.75%		95.5/200

How often now I have seen such a mess? How do you justify to mix the two samples if you tested with two different, seemingly important, contempt factors? Why not checking if more games with 15 would assure the almost equal result? At least you could have made some reflections and commented on your decisions. Such results are never talking as such, also in that more lower N scale.

No mistake, I dont want to demotivate you but I have learned a different stats. Why not giving something back to you?

Perhaps you have still more results?

How often now I have seen such a mess? How do you justify to mix the two samples if you tested with two different, seemingly important, contempt factors? Why not checking if more games with 15 would assure the almost equal result? At least you could have made some reflections and commented on your decisions. Such results are never talking as such, also in that more lower N scale.

No mistake, I dont want to demotivate you but I have learned a different stats. Why not giving something back to you?

Perhaps you have still more results?

Your criticism is well received. The reason I tested with contempt factor of zero initially is because some have claimed Rybka to be stronger than this engine using that particular setting. Perhaps, I will conduct more tests with default settings and publish the results.

Back in 1996 I showed that the SSDF, a private Swedish test board were NOT following usual rules. They took 16 games results here, some 20 there and another 7 there and then they added the results and got a rating number. IMO this is all wrong. But then like totay people normally could do what they want and after I have made my critic I step back, unless I am suddenly attacked for my criticism, of course then the whole debate goes on. Look, coming from science I know that you cant enforce the truth or best results unless they are accepted in general. So, I know that we must live with much imperfect and untrue stuff. No problem with that. I am not the guardian. If you read my critic and decide to do it like before then this is ok. I have no higher power.

SSDF can answer for themselves.

Perhaps we have the chance to talk about it a bit deeper but until then do as usual. It's your hobby. But not science.

As long as we can keep it on-topic.

What I meant in my critic was mainly that it's ridiculous to weaken Rybka with no pondering only because the cloners have not yet found out how to implement that feature. Know what I mean? So, if you deactivate all the good things in Rybka you might get the results you reported. But they are not valide.

Rybka was not weakened. Both initial tests/results had Rybka using pondering while RobboLito did not. I was under the impression RobboLito was pondering (due to cpu time/cpu activity) but the code itself had not been implemented in that version. One test was done with Rybka using contempt factor of zero because some suggested Rybka showed better performance against RobboLito using that setting. And the other test used Rybka's default settings.

I would only ask you to always honestly report exactly what you have done in details. Then this is ok because others could then see it. Otherwise it were fraud. But until now I have no reason to accuse you of anything. In contrary, I like that you are communicating. That's what such a forum is made for.

I don't only post the games--I include details regarding test methods and conditions as well.

All I would say that right now it's premature to say exactly how strong the Robbothing is. But that can change soon. Dont worry. As you know I have a total horror in front of anonymous people behind Robbos. But I would also speak with Stalin or the Mafia if that would go in civilized manner. No problem for me.

Time will tell.

[kingliveson]

I really hate to have to quote myself, but this was my post after re-running the test:

I posted the results of Rybka vs RobboLito yesterday. I had run the test with ponder on--though it seems RobboLito was pondering, it was made known to me that it was in vain because the code itself had not been implemented--and that gave Rybka an advantage in the test. Yet RobboLito came out ahead. I ran a test overnight with the latest RobboLito which actually supports ponder.

The test was conducted using Fritz 11 GUI on Windows 7 64-bit. The hardware is an AMD Phenom II 940 overclocked to 3.6 GHz.

Both engines used the same book renamed to bookone.ctg (Rybka 3) and booktwo.ctg (Robbolito 0.085e4). The book has no marked/priority moves. The book settings are 2/12 with learning mode disabled. The variety of play was set at approximately 71.4% which is 20/28 right arrow key.

The engines were 1 CPU 64 bit versions with hash size set to 256 MB and pondering/permanent brain was enabled. Rybka used contempt setting of 15 (default). Time control is 4 minutes with no increment.

Result:

Code: Select all

1	RobboLito 	+35/=48/-17	59.00%		59.0/100

2	Rybka 3		+17/=48/-35	41.00%		41.0/100

More games to come...

[Rolf]

I cant see your problem. I saw your results. But they are not valid in my eyes. Anonymous people have allegedly stolen something and now made it better than the weakest version of the known Wch. But look, it became became Wch on way faster hardware and with a different program design. Not with the crippled version you ran your test. I ask myself how I can explain such a beancounter problem to someone else if not even the master himself here understands it. You can test all you want but it has no relevance and I dont believe you that Robbanonymouschicks is stronger than Rybka the Wch. Easy enough for you?

I'm talking here with someone who tests the Wch as a cripple against the stolen stuff improved by anonymous weakis. I get scabies. I loathe it.

[kingliveson]

I cant see your problem. I saw your results. But they are not valid in my eyes. Anonymous people have allegedly stolen something and now made it better than the weakest version of the known Wch. But look, it became became Wch on way faster hardware and with a different program design. Not with the crippled version you ran your test. I ask myself how I can explain such a beancounter problem to someone else if not even the master himself here understands it. You can test all you want but it has no relevance and I dont believe you that Robbanonymouschicks is stronger than Rybka the Wch. Easy enough for you?

I'm talking here with someone who tests the Wch as a cripple against the stolen stuff improved by anonymous weakis. I get scabies. I loathe it.

I don't know what else to say but "We played the flute for you, and you did not dance; we mourned to you, and you did not lament."

[Milos]

Franklin, there is simply no point. Here, only a specialist can help, and certainly not a chess specialist. Most of the people at the forum have already realized it.

[slobo]

Franklin, there is simply no point. Here, only a specialist can help, and certainly not a chess specialist. Most of the people at the forum have already realized it.

In this fellow´s case nothing can be done:
he thinks that a specialist is himself. And more than this:
he used to call himself "the pope".

[bob]

I think we have it now. Thanks so far.

I thought that until the fog isnt lifted you were not starting to sue Vas, but after that you would. Now I understand you. This all isnt about a sort of legally relevant failure but some sort of formal inexactitudiness without any legal consequences after such a long time, correct? If seen earlier you might have taken actions but no longer now, ist that true so far?

Again, actually Vas wont talk but as he confirmed others besides me he will talk after he got some things would become clearer.

Couldnt we just give him that time which seems to lay not in his realm and stop this climate of hate against Rybka?

I ask this for the sake of our own peace, not in any way to protect Vas. He has his own remedy and that is ignoring. But we are going nowhhere with this campaign. Also please consider that most here are younger and they might seek a sort of winner which isnt very likely to get. All IMO.

Ok, I am a total outsider, but from all what I know about peace of mind I would support if we would stop such a scapegoating. It shouldnt be allowed on CCC. BTW didnt Vas react positively for talking with you? Let's see what will happen.

All the best for a nice weekend for you all.

Rolf

here's my take on this: He has as long as he wants to make some sort of statement, either about fruit code in Rybka 1 or exactly what was copied from his code and included in robo*. But in fairness, I see absolutely no justification to continue to maintain that robo* is any sort of clone. There are several reasons:

(1) it is stronger (significantly stronger, +70 elo is _not_ easy to do in any program, without a ton of time and a ton of testing);

(2) we have seen absolutely no evidence that it is a clone. Just a statement. And I don't see how anyone could consider it fair that this just drags on and one with no evidence offered.

In light of that, the best solution is to treat Robo* like any other program. If it isn't a clone, good. If it is a clone, that will give Vas incentive to speak up before he drops off the top of all the rating lists. Either way, the truth comes out and that's all anyone really wants at the moment.

ad 1) I doubt the better strength, mostly there are made up tests without meaning because the tests are not made from a neutral angle but always what can actually our baby program do at best, this is such a bias, that it cant be taken for a serious testing result at all

ad 2) the striking argument against the thing is that its authors, if ever singular existences, are anonymous. That speaks for evil-doing in the past because they must hide to not getting into legal difficulties. Otherwise if someone would step forward, I would make a new judgement but actually we here cannot accept that such a good program is based on hiding and hidden authorship. Vas was from his beginning a known figure who stood for the project no matter who collaborated else. Until the Robbos dont open their shields they cant be seriously be taken into test, serverplays and tournaments for any title out there.

ad 2 NB)

if we would accept such anonymous entries we had to face next that say some real talents or titleholders themselves build their second or third families under anonymous accounts, perhaps they steal data from other codes and then participate as manipulative helpers of certain favorites, of themselves, please all may remember the problems a single artist like Bobby Fischer had to face when he had to compete against an armada of Soviet players who didnt play for their own fate so to speak but they were instruments to support the actually best runner for the title; they all resisted to beat each other while they played Fischer for real. If we accepted anonymity in our sport that whole scenario would be repeated. The singular fighters no longer had any chance at all. The Stalinists had their go.

Ad 2 PS)

It is a generally accepted truth that these Robbists are hurting computerchess as a community. And this should be the reason why we all should defend ourselves against these anonymous chicken heads who are dishonest and unethical. I called them chess cyber terrorists. I never saw someone oppose the verdict.

Robbos could become legal but if they insist to be illegal they cant be members of our community.

Rolf GENS UNA SUMUS MINUS CYBER TERRORISTS

I still don't follow your comments. We know Robo* is stronger. It has been tested in several regular tournaments and the results published. Not on test positions or other odd things, but in real games.

I also don't follow your reasoning. On the one hand you said that even if Vas _did_ copy fruit, he modified it to make it much stronger so that it wasn't the same as a clone. Yet now you claim that Robo* is a copy of Rybka, where the authors modified it to make it much stronger, and that still is a clone? You can't have it both ways.

Either Rybka is a clone, or Robo* is not, based on your logic. But it is an either/or proposition, not an "and".