CEGT - rating lists October 30th 2011

[ethanara]

perhaps they could choose the latest and strongest release...
IvanHoe 999947c?

I recommend IvanHoe 63Mod5-x64 about 5-10 elo more

[kranium]

Johan/Werner, and Graham as well...

Can you guys explain why you test some 'cloned' engines but others?

i.e. in particular:
why Houdini, Rybka, and most recently Strelka (RE'd Houdini)...
but not IvanHoe?

there seems to be no consistent policy, or some double standard at work?

I thought it was simply a decision to test the best of the Ippolit derivatives, which was clearly Houdini. But if they are testing Strelka as well, that seems to contradict this notion. Personally I can't see any reason not to rate one of your engines now. Perhaps you should determine whether Fire or Ivanhoe is stronger and request that one particular version only be tested, if you are the main author of both of these engines.

Nope..
for a brief time, early Firebird versions were the strongest Ippolit engine...
Firebird was the next step in (our) Ippolit development efforts, and came shortly after we released Robbolito 0.085g3 and 0.09.
I remember quite clearly because I communicated to Sentinel that we should keep the source closed until Firebird was no longer #1, and then share it.
(i.e. our intention was to see just how long it would remain on top). I also publicly promised to release the source code eventually...(on open-chess.org).

but Firebird was never tested by the groups mentioned above.

for that matter, no version of FireBird/Fire (or IvanHoe), or any earlier Ippolit: Robbolito 0.09, Igorrit, etc. has ever been tested by them.

Well, Firebird didn't last long as #1, (maybe 6 months or so) because Houdini was released.
but the CCRL, CEGT, and IPON started testing Houdini almost immediately...?

Shortly after Houdini's initial release, we released Firebird source code, and it has remained open-source ever since.

[Adam Hair]

but not IvanHoe?

Frankly, there used to be, and still are so many IvanHoes, that it is indeed difficult to choose... and there is no "author" who can say: this one has added value, is stable, please test that one!

man that's pretty weak.

are you telling me all three: CCRL/CEGT/IPON are so overwhelmed and confused they can't choose one?

perhaps they could choose the latest and strongest release...
IvanHoe 999947c?

or maybe Fire?...I can assure you I'm a real person, and Fire is proven very stable.

fact is: the entire Ippolit family of engines has been blacklisted by them...yet they test all manner of other RE'd and 'cloned' engines.

these unscrupulous 'testers' have abused their influence, and cannot (or will not) admit they have been completely unfair and mistaken...

meanwhile while Ippolit continues to be 'blacklisted', other engine authors are free to rape and plunder the fantastic source code,
even if it's GPL'd, and their creations are warmly embraced!

Hi Norman,

First of all, let me make it clear that I am speaking for myself and for no one else. If you or anybody else has a problem with what I write here, then all criticism should be directed towards me.

I am free to test and submit games involving any engine that I want. Including Ippolit and its derivatives. So could anybody else in the CCRL. However, I will not submit games involving an engine that has no real name attached to it. I have said this before. I say it now. To do so would be to spit in the face of authors that I respect. It does not matter to me what you or anybody else thinks about my personal choice.

That leaves engines such as Houdini, Saros, Tankist, etc... . Houdini first made its way onto the CCRL lists because Rybka was on the lists. I personally tested it and submitted games for that singular reason. By the way, Houdini had been out for 8 1/2 months before the first games were submitted to the CCRL. I have not tested Saros and Tankist, primarily because I have plenty of other engines to test and there is nothing about them that makes them interesting to me.

Fire. That is the one that I keep thinking about. Although I have doubts about it being the top engine for 6 months (in my personal testing, was it no stronger than various IvanHoes), I do believe that you and Milos did enough to it to warrant testing. However, there is your personal history that causes me some consternation. It is not just that you tried to sell a clone of Toga (Xyclops). You did apologize for that and refunded money. If it was only that, then possibly somebody would test Fire. However, then there was Crimson, a clone of Viper. Not that there is a problem with working on open source engines IMHO, but you released it closed source and with no acknowledgement of its relationship to Viper. Finally, there is Fire, whose source was closed at for a time. If you had been nearly as successful as Robert Houdart has been with Houdini, elo-wise, I seriously doubt that you would have opened the sources. More likely, you would have been selling it.

I am sure that you could care less about what I think. There is no compelling reason why you should. But I will state my thoughts anyway. I have a hard time with supporting someone who has breached the trust of the community at least two times. Furthermore, your complaints about Fire not being tested are done in a way that is not conducive to changing minds. It would appear that your biggest "mistake" was to admit, either explicitly or implicitly, that you had lied about the origins of your engines. After all, others have apparently lied about their engines and they are not being treated as you are. However, it is this "mistake" that would lead me to test Fire. Just to show, in my own way, that honesty matters. I just don't think it really matters to you. You believe that you have been treated unjustly, but you don't seem to believe you have done things to warrant your treatment. So I am not inclined to support you.

Adam

[kranium]

Hi Norman,

First of all, let me make it clear that I am speaking for myself and for no one else. If you or anybody else has a problem with what I write here, then all criticism should be directed towards me.

I am free to test and submit games involving any engine that I want. Including Ippolit and its derivatives. So could anybody else in the CCRL. However, I will not submit games involving an engine that has no real name attached to it. I have said this before. I say it now. To do so would be to spit in the face of authors that I respect. It does not matter to me what you or anybody else thinks about my personal choice.

That leaves engines such as Houdini, Saros, Tankist, etc... . Houdini first made its way onto the CCRL lists because Rybka was on the lists. I personally tested it and submitted games for that singular reason. By the way, Houdini had been out for 8 1/2 months before the first games were submitted to the CCRL. I have not tested Saros and Tankist, primarily because I have plenty of other engines to test and there is nothing about them that makes them interesting to me.

Fire. That is the one that I keep thinking about. Although I have doubts about it being the top engine for 6 months (in my personal testing, was it no stronger than various IvanHoes), I do believe that you and Milos did enough to it to warrant testing. However, there is your personal history that causes me some consternation. It is not just that you tried to sell a clone of Toga (Xyclops). You did apologize for that and refunded money. If it was only that, then possibly somebody would test Fire. However, then there was Crimson, a clone of Viper. Not that there is a problem with working on open source engines IMHO, but you released it closed source and with no acknowledgement of its relationship to Viper. Finally, there is Fire, whose source was closed at for a time. If you had been nearly as successful as Robert Houdart has been with Houdini, elo-wise, I seriously doubt that you would have opened the sources. More likely, you would have been selling it.

I am sure that you could care less about what I think. There is no compelling reason why you should. But I will state my thoughts anyway. I have a hard time with supporting someone who has breached the trust of the community at least two times. Furthermore, your complaints about Fire not being tested are done in a way that is not conducive to changing minds. It would appear that your biggest "mistake" was to admit, either explicitly or implicitly, that you had lied about the origins of your engines. After all, others have apparently lied about their engines and they are not being treated as you are. However, it is this "mistake" that would lead me to test Fire. Just to show, in my own way, that honesty matters. I just don't think it really matters to you. You believe that you have been treated unjustly, but you don't seem to believe you have done things to warrant your treatment. So I am not inclined to support you.

Adam

Hi Adam-

May I point out that in both of cases you illustrate above, I immediately stopped working on and retreated the engines...
precisely as the CC community wished...why? because I care about people and what the CC thinks. I did the right thing and moved on.

Adam, neither was in fact a 'clone' (a simple copy)...and both had great potential.
for ex: xyclops was WB, and had unique features like Dynamic Hash, etc., and was stronger than Toga, and I was busy implementing true SMP

I admit, I do have fairly liberal ideas concerning not 'reinventing the wheel', and the utilization of open sources (apparently not unlike some other authors),
and I don't deny (and never have) that for both engines:
I went through the Fruit and Glaurung code forwards and backwards and took many things.

But I was wrong and I understand that a higher standard is expected...
so, since that time, I have worked even harder by developing the Ippolit source code for the benefit of all, and have published everything as GPL'd open-source.

I do take exception to your statement:
"If you had been nearly as successful as Robert Houdart has been with Houdini, elo-wise, I seriously doubt that you would have opened the sources. More likely, you would have been selling it."

...when Houdini was initially released, it was a barely stronger than Firebird 1.0 or Robbolito 0.09, as the author himself admits
(apparently in one of his rare moments when he's not 'dancing with words'):

Strength
Houdini appears to be notably stronger (+75 Elo) than Stockfish 1.7 and Rybka 3, and slightly stronger (+20 Elo) than the Firebird / Ivanhoe / Robbolito gang.
Your mileage may vary, obviously.

All my ChessLogik Ippolit development efforts (3 RobboLites, 13 RobboLitos, Iggorrit, FireBird, Fire, etc.) have been open-source...
except early Firebird for a very short time when it was #1...(3-6months right before Houdini 1 was released)
that's a fact.


But here's the injustice:

What I did was/is absolutely no different than what Vas R. did...
except one big difference:
he didn't and still doesn't give a damn about what the anybody thought, or what evidence was presented...
he never even had the courtesy or bothered to respond...he simply denied and ignored it for years...(and still does!)
in the meantime he got plenty of time (5 years), plenty of money, and resources to push it forward, and he has made oodles of money from his 'clone', and he's still selling it

and what about Robert H.?
he's on the exact same track...and an absolute boatload of evidence has accumulated, with plenty more not even presented yet that he started with
(a cloned? i hate to use that word but you used in your post above, very derogatory, derivative is more appropriate ) one of my and Sentinel's GPL'd RobboLitos.

Why is it OK for them but not for me (or others)...?
What about a 'higher standard' for them?

It seems you are saying:
once is ok, even if you lie thru your teeth for years and make oodles of money from it...
but twice, you're out (and villified/banned for life) even though you may have done the right thing and made a great effort to attone for you past trangressions.
?

Regards-
Norm

[kranium]

I will not submit games involving an engine that has no real name attached to it. I have said this before. I say it now. To do so would be to spit in the face of authors that I respect.

I disagree with this...I don't see authors complaining.

from what I've read, many/most authors are eagerly going thru the code forwards and backwards...
many top engines have already done so, and i think they're pretty happy about and benefited substantially from what they could take.

including Vas, according to some credible posts on Open-Chess...

[Adam Hair]

Hi Adam-

May I point out that in both of cases you illustrate above, I immediately stopped working on and retreated the engines...
precisely as the CC community wished...why? because I care about people and what the CC thinks. I did the right thing and moved on.

Adam, neither was in fact a 'clone' (a simple copy)...and both had great potential.
for ex: xyclops was WB, and had unique features like Dynamic Hash, etc., and was stronger than Toga, and I was busy implementing true SMP

I admit, I do have fairly liberal ideas concerning not 'reinventing the wheel', and the utilization of open sources (apparently not unlike some other authors),
and I don't deny (and never have) that for both engines:
I went through the Fruit and Glaurung code forwards and backwards and took many things.

But I was wrong and I understand that a higher standard is expected...
so, since that time, I have worked even harder by developing the Ippolit source code for the benefit of all, and have published everything as GPL'd open-source.

I do take exception to your statement:
"If you had been nearly as successful as Robert Houdart has been with Houdini, elo-wise, I seriously doubt that you would have opened the sources. More likely, you would have been selling it."

...when Houdini was initially released, it was a barely stronger than Firebird 1.0 or Robbolito 0.09, as the author himself admits
(apparently in one of his rare moments when he's not 'dancing with words'):

Houdini wrote:

Strength
Houdini appears to be notably stronger (+75 Elo) than Stockfish 1.7 and Rybka 3, and slightly stronger (+20 Elo) than the Firebird / Ivanhoe / Robbolito gang.
Your mileage may vary, obviously.



All my ChessLogik Ippolit development efforts (3 RobboLites, 13 RobboLitos, Iggorrit, FireBird, Fire, etc.) have been open-source...
except early Firebird for a very short time when it was #1...(3-6months right before Houdini 1 was released)
that's a fact.


But here's the injustice:

What I did was/is absolutely no different than what Vas R. did...
except one big difference:
he didn't and still doesn't give a damn about what the anybody thought, or what evidence was presented...
he never even had the courtesy or bothered to respond...he simply denied and ignored it for years...(and still does!)
in the meantime he got plenty of time (5 years), plenty of money, and resources to push it forward, and he has made oodles of money from his 'clone', and he's still selling it

and what about Robert H.?
he's on the exact same track...and an absolute boatload of evidence has accumulated, with plenty more not even presented yet that he started with
(a cloned? i hate to use that word but you used in your post above, very derogatory, derivative is more appropriate ) one of my and Sentinel's GPL'd RobboLitos.

Why is it OK for them but not for me (or others)...?
What about a 'higher standard' for them?

It seems you are saying:
once is ok, even if you lie thru your teeth for years and make oodles of money from it...
but twice, you're out (and villified/banned for life) even though you may have done the right thing and made a great effort to attone for you past trangressions.
?

Regards-
Norm

What I was saying is that, unlike other people, you have eventually acknowledged that your engines were not solely your creations. That is why I used quotation marks around the word mistake. You have received more backlash since you have not denied (and thus not given other people the opportunity to deny the facts) anything. Ironic and unfortunate, to a degree. However, you do not seem to acknowledge the basis for which you have been treated. No one can legitimately shun you because you use an open source engines as the basis for your work. No, that is not the problem. It is the habit of claiming, either implicitly or explicitly, other people's work to be your own that is the problem. You were forgiven by most for Xyclops. Crimson turned many against you, rightfully so. Perhaps it is time for that to change, but your approach and attitude won't make that change occur.

A couple of notes:

My use of the word 'clone' may not be totally appropriate in relation to your work. However, there is little evidence that you make sizable changes to the code that forms the basis for each of your engines. Much evidence that you add many things, but little showing you modify the essence of the preexisting code. But still, I probably should use 'derived', if for no other reason than to be less contentious. My purpose is not to start a fight but to address your statements, rather than ignore them.

I will retract my supposition about you making Fire commercial if you had been able to a significant amount of Elos. Perhaps that would not have occurred.

[kranium]

However, there is little evidence that you make sizable changes to the code that forms the basis for each of your engines. Much evidence that you add many things, but little showing you modify the essence of the preexisting code.

Adam-
I don't know how you can say that, that's entirely unfair...have you ever really compared any of the source code closely?
the changes are more than substantial...often even SMP revised. (see Iggorit/Firebird)

RobboLito 0.09 is greatly improved, as strong if not stronger than any original wikispaces Ippolit release...
and it was released by Sentinel and I almost two years ago!

here's an image of the longest running 1 CPU rating list on a well-known Russian site:



And Fire 2.2 w32 is #4 and stronger than IvanHoe...after almost 1200 games on Frank's list

• 1 Houdini 1.5 w32 2987 20 19 1160 80% 2753 29%
  2 Critter 1.2 w32 2936 19 18 1120 75% 2750 35%
  3 Stockfish 2.1.1 JA w32 2904 18 18 1160 71% 2756 39%
  4 Fire 2.2 xTreme w32 2903 18 17 1160 70% 2756 40%
  5 Rybka 4.1 w32 2901 18 18 1160 70% 2756 37%
  6 IvanHoe B47cB w32 2896 18 17 1160 69% 2756 39%
  7 Stockfish 1.9.1 JA w32 2895 20 20 1000 77% 2694 31%
  8 Rybka 4 w32 2893 18 18 1200 76% 2697 32%
  9 Critter 1.01 w32 2891 18 18 1200 72% 2721 33%
  10 Komodo 3.0 w32 2887 18 18 1240 72% 2723 32%

I can send you Fire 2.3 w32 which is proving to be another 20-30 ELO stronger...(after many weeks of testing) would that help?
here's the current ongoing test:

Games Completed = 2836 of 50000 (Avg game length = 14.882 sec)
Settings = RR/64MB/1000ms+100ms/M 1000cp for 12 moves, D 150 moves/
Time = 17536 sec elapsed, 291632 sec remaining
1. Fire 2.3_049 w32 1547.0/2836 1053-795-988 (L: m=451 t=108 i=0 a=236) (D: r=722 i=139 f=64 s=2 a=61) (tpm=105.2 d=8.9 nps=237615)
2. Fire 2.2 xTreme w32 1289.0/2836 795-1053-988 (L: m=598 t=100 i=0 a=355) (D: r=722 i=139 f=64 s=2 a=61) (tpm=105.4 d=9.0 nps=216025)

I'm able to send a version with learning, but i'm not likely to ever release it because the ELO gain if offset by loss of speed the extra code causes.

[Frank Quisinsky]

Hi Norman,

I have interest to test next Fire and IvanHoe with the next SWCR-32 and SWCR-64 update tourneys beginning of the next year. So I am waiting of a new version of Izak. It would be great if your next Fire will be ready end of the year.

So far I have a lot of work with my Oldie-Mix tourney The SWCR TOPs are in waiting position from my point of view.

Keep up your good work.

Best
Frank

[kranium]

Hi Norman,

I have interest to test next Fire and IvanHoe with the next SWCR-32 and SWCR-64 update tourneys beginning of the next year. So I am waiting of a new version of Izak. It would be great if your next Fire will be ready end of the year.

So far I have a lot of work with my Oldie-Mix tourney The SWCR TOPs are in waiting position from my point of view.

Keep up your good work.

Best
Frank

Thanks Frank, I'll have it ready for you for sure.
Sadly enough, I still work on an old 32-bit system, and so can't spend any time optimizing for 64-bit..
but my 32-bit compiles are and have always been pretty good IMO

[Frank Quisinsky]

Hi Norman,

yes, the 32bit compiles are great.
Can see it in your Fire and can see it in your Houdini.
Look in SWCR, Houdini 1.5 w32 and x64.

Good work you do, unfortunately and others used!

Best
Frank