BrainGames Qualifier, about unfair competition
Ed Schroder
Event: Kramnik vs best computer chess program (October 2001)
Organization: Brain Games Network
In order to determine which software will play Kramnik Brain Games is
hosting a Machine vs Machine qualifying challenge in April 2001.
For the Qualifier just 3 chess programs (Fritz, Junior and Shredder)
were invited which are presented as the strongest of the world. Strong
chess programs like REBEL and CHESS TIGER were not invited even though
they shared the 3th place during the last World Championship. However
Junior although ending fifth strangely enough is invited.
In the meantime Shredder has declined to play the Qualifier with as
result the Qualifier will be played between just 2 chess programs! Tries
to get REBEL and CHESS TIGER to play in the Qualifier failed by stubborn
unwillingness of the organizers.
Rebel Company has strongly protested against this arbitrariness and
likes to present the facts about this injustice and unfair competition.
--------------------------------------------------------------------------------
THE NAKED FACTS
On April 8th Rebel Company for the first time heard about the BrainGames
event. Ed Schroder, author of REBEL, was send email by BrainGames that
REBEL and CHESS TIGER were not invited because, they said, REBEL as well
as CHESS TIGER were not able to run on multiple processors.
Furthermore Ed Schroder was strictly forbidden to speak with anyone
about this information not even with his partner Christophe Theron,
author of Chess Tiger. This information boycott (their first mistake) is
essential for exclusion of Chess Tiger from the Qualifier event, this we
had to learn later.
As a matter of fact a so-called DEEP-TIGER version was available but the
organizers just ASSUMED we had not, their second mistake.
Later, much later the information boycott for Ed Schroder to speak with
his partner about the requirements was solved by itself because the
restriction became public information. After several phone calls with
the BrainGames people and with Christophe Theron, author of Chess Tiger
Rebel Company on April 20th offered DEEP TIGER to BrainGames to play in
the Qualifier.
To our great astonishment Rebel Company was shut out again, now with the
excuse that the Qualifier had to start on April 27 and on these short
terms it was impossible to change their schedule. BrainGames furthermore
said we should have informed them on April 8th about the availability of
DEEP-TIGER and then DEEP-TIGER would have played, they said.
We do not understand the BrainGames logic, BrainGames expects Ed
Schroder to reveal secret business information of Chess Tiger to
BrainGames while at the same time BrainGames forbids Ed Schroder to talk
about that with Christophe Theron the author of Chess Tiger himself !!
Furthermore BrainGames did not provide other crucial information like
the dead-line of April 27, the planned start of the Qualifier leaving us
in the dark.
Two days later (April 22) we got a faked invitation from BrainGames.
Enrique Irazoqui who was attracted by BrainGames to organize the
Qualifier proposed, "If you send DEEP-TIGER before April 23, 11:00 we
will let it play in the Qualifier". This is an impossible demand and Mr.
Irazoqui knows it! as no chess programmer is able to make things ready
in just less than one day!
Counter offers from our side failed, we proposed, We will send you the
current DEEP-TIGER before your (impossible) deadline of April 23, 11:00
to check its functionality, on April 27 you start the Qualifier as
planned with the first match and on or before April 30 you will have the
final version of DEEP-TIGER and thus there are no delays for you as on
April 30 the first match is still not finished. A wall of granite was
our reward, no flexibility on their end even though the situation was
created because of their own errors, no invitation, withhold crucial
information, the information boycott which caused a 7 day delay for
Rebel Company to make the needed preparations.
During the negotiations Rebel Company was seriously offended by
BrainGames as BrainGames on multiple occasions made serious doubts on
the availability of a DEEP-TIGER version. This indirectly labels Rebel
Company as a liar and is understood that way. Especially from such
offending questions like these ones we got the impression, Rebel Company
and Chess Tiger were just not welcome and that our talks were doomed to
fail on beforehand.
Yet the worst thing had to come and says it all, the crucial question
simply remained unanswered. Christophe Theron to BrainGames, "Why did
you forbid Ed Schroder to talk with me about the "DEEP" requirement? Was
there something I was not allowed to know?" No answer from BrainGames.
The information boycott has cost Christophe Theron one week, exactly the
week Rebel Company needed to make the software ready.
CONCLUSION
Chess Tiger was not welcome and was banned on beforehand. Now the
Qualifier is played between 2 ChessBase programs. This whole Qualifier
is one big farce.
--------------------------------------------------------------------------------
GIANT PROTESTS BY VARIOUS PEOPLE
Right from the start BrainGames received major criticism in public
forums in proportions of 10 to 1 at least, this regarding the selection
procedure of the Qualifier. Even the major chess programmers have spoken
against this Qualifier event and its unfairness, respectable names like:
Professor Robert Hyatt, Cray Blitz and Crafty
Bruce Moreland, Ferret
Chris Whittington, Chess System Tal
Ulrich Tuerke, Comet
Stefan Meyer-Kahlen, Shredder
Peter McKenzie, LambChop
James Robertson, Insomniac
Dieter Buerssner, Yace
Ed Schroder, Rebel
Christophe Theron, Chess Tiger
All disagreed with the selection procedure, the Qualifier to play
Kramnik.
Other known names have protested too, to name the most important one,
David Levy, chairman of the ICCA. The ICCA is the key organization that
unite and represents chess programmers and is responsible organizing the
yearly official World championship, this successfully over a period for
more than 25 years. The protest is IGNORED !!
More protests are coming in daily from chess journalists all over the
world. Yesterday we received a so-called "open letter" from Jeroen
Noomen a long time dutch chess journalist, computer chess expert since
18 years, one of the leading opening experts in the world. Jeroen Noomen
has earned extra credit being the author of the opening books for Rebel
since 1989. Jeroen has asked us to put his open letter on the Rebel Home
Page and so we do.
Open letter to prof. Enrique Irazoqui
by Jeroen Noomen
Dear prof. Irazoqui,
During the past few weeks all chess enthousiasts and chess computer fans
were delighted to hear about the Braingames qualifier, to produce a
'challenger' to play Vladimir Kramnik in a serious match. Organisation
was given to you, a pretty easy to understand decision as you have a lot
of experience in organizing a chess computer tournament. I know you
since several years as a fine person, unbiased and with a lot of
understanding concerning chess programs. I always have been an admirer
of your work, especially the famous Cadaques Tournament.
Needless to say I was stunned to hear that Chess Tiger, currently one of
the strongest chess programs, was not even invited to participate in
this qualifier. I tried to get information about the 'why', but up till
now nobody has given any specific argument why Chess Tiger could not
play there. To exclude such a strong program is, of course, very strange
(to say the least). Several people have asked me why Chess Tiger is not
participating. I only could say 'I don't know, sorry'.
Suppose there will be a qualifying event amongst human players to bring
about a new challenger for Kramnik. Wouldn't it be very strange if the
organisers of such an event would say to Anand 'sorry, you can't
compete'? Nobody would understand that. As in this case nobody
understands your decision why you refrain from letting Chess Tiger play
in these qualification matches. Therefore I want to ask you, publicly
through this letter, to clarify matters and to present the chess
enthousiasts the answers on the following questions:
1. Why wasn't Chess Tiger invited to this qualifier?
2. Why did the Chess Tiger team only hear about this qualifier some
weeks before the event started?
3. Why did you refuse to let Chess Tiger play, even though the Tiger
team very willingly cooperated to make the participation of Chess Tiger
possible?
4. Why didn't you try to find a positive solution to this matter?
5. Why did you keep on saying 'no', even when it became clear Chess
Tiger could play under the conditions you wanted?
6. Why did you refuse to believe that Chess Tiger could play under your
conditions, the more because you haven't checked this out yourself?
7. Do you think it is fair to exclude very strong chess programs from
such an event?
8. What proposal was made to the IBM team: Could they play on their own
hardware, or did they have to play on specified hardware?
9. What was the reason IBM declined the offer?
10. Knowing you as a fair person, unbiased and with great achievements
for computer chess, can you now look into a mirror and say to yourself
'I have been absolutely fair to all parties and I have been honest to
all the other chess programmers who are not competing now'?
11. I am pretty sure that deep inside you, you will not feel happy about
this. And that you know that programs like Shredder and Chess Tiger
should compete as well. Am I correct in stating this?
12. Now there is a delay in the Fritz-Junior match, is there room for
Chess Tiger in this event?
I think everybody would appreciate it when you give a public statement
and answer those questions in an honest way. Otherwise people will get
the impression that this qualifier is a pre arranged event, that is not
open to all the strongest chess programs, but only to some of them. It
is up to you to clarify matters and to make clear why Chess Tiger is not
playing here. I think you owe us an explanation.
You made a clear mistake by excluding Chess Tiger from this event. That
is my opinion and I know many people will agree with me. Still, making
mistakes is human, it can happen to everybody. Truly great people are
those people who admit they made a mistake and correct it. You are still
able to correct this mistake, too. By allowing Chess Tiger to play in
the qualifier.
I ask you to give the answers to these questions, questions that many
people are wondering about. And that you correct your mistake and let
Chess Tiger play in the qualifier.
Sincerely yours,
Jeroen Noomen
CCCage
fixed. They should call this tournament CB products vs. Kramnik and nothing
more.
Thx,
CCCage
"Ed Schroder" <in...@rebel.nl> wrote in message
news:3AEC03CE...@rebel.nl...
Paul Vosmer
"Ed Schroder" <in...@rebel.nl> schreef in bericht
news:3AEC03CE...@rebel.nl...
Robert Hyatt
> --------------------------------------------------------------------------------
> THE NAKED FACTS
How can they "forbid you to speak with anyone"???
> As a matter of fact a so-called DEEP-TIGER version was available but the
> organizers just ASSUMED we had not, their second mistake.
Sorry, but I don't buy that. I sort of see why BG wanted to only include
SMP programs. They are definitely stronger than their non-SMP counterparts.
Even more importantly, the one sponsor was going to provide an 8-way box
to use for the event. Obviously a program that could use 8-way SMP was
desiarable from their point of view.
The issue of not including Ferret is strange, since Ferret is definitely
one of the very top few programs in the world, again based on its results
at WMCCC/WCCC events. At the last WCCC event, it had a playoff with
Shredder to determine the title.
As far as a "Deep Tiger" goes. It hasn't been seen _anywhere_. Knowing
how difficult a "good" SMP search is to develop, it seems improbable that
a good SMP program would exist that nobody knows of. Their failure to
ask was therefore not surprising. If you had such an engine, then you
should have had it out in the public somewhere to make it known. If you
were "hiding" it until the next WCCC event (or whatever) then that is a
risk you take by trying to be secretive.
I don't see how they can be faulted for not knowing you had a parallel
program. I do see how they can be faulted for (a) holding a qualifier
at all when Shredder should have been taken as the opponent for Kramnik
based on holding both the WMCCC and WCCC titles awarded by the ICCA;
(b) not including Ferret which was SMP long before the others in the
contest.
> Later, much later the information boycott for Ed Schroder to speak with
> his partner about the requirements was solved by itself because the
> restriction became public information. After several phone calls with
> the BrainGames people and with Christophe Theron, author of Chess Tiger
> Rebel Company on April 20th offered DEEP TIGER to BrainGames to play in
> the Qualifier.
> To our great astonishment Rebel Company was shut out again, now with the
> excuse that the Qualifier had to start on April 27 and on these short
> terms it was impossible to change their schedule. BrainGames furthermore
> said we should have informed them on April 8th about the availability of
> DEEP-TIGER and then DEEP-TIGER would have played, they said.
> We do not understand the BrainGames logic, BrainGames expects Ed
> Schroder to reveal secret business information of Chess Tiger to
> BrainGames while at the same time BrainGames forbids Ed Schroder to talk
> about that with Christophe Theron the author of Chess Tiger himself !!
> Furthermore BrainGames did not provide other crucial information like
> the dead-line of April 27, the planned start of the Qualifier leaving us
> in the dark.
Re-read what you are saying above. You think they should _know_ what you
call a "business secret". In this case, you simply screwed up by keeping
the SMP tiger secret. Maybe they should have done a world-wide advertisement.
Maybe not. If they had asked _me_ to do the qualifier, and given me the
requirement that the programs I selected had to be SMP-ready, I would
have chosen Shredder, Ferret, Fritz and Junior. I _know_ those programs
are SMP already. I would have first tried to negotiate for Shredder alone
as it is the _right_ program to challenge Kramnik with _no_ qualifier
needed. But if there _had_ to be a qualifier, those 4 would have been
invited by me. Assuming the time-frame was short (and it was) I would
not have thought it necessary to do a public announcement.
Your decision to keep an SMP program "secret" backfired. In reality,
I consider that "poetic justice" since commercial programs have kept
_everything_ secret over the years, but faithfully studying all the
public chess works to make sure no new ideas are missed.
I don't think you can have it both ways. If you live in secrecy, then
sometimes you die in secrecy.
> Two days later (April 22) we got a faked invitation from BrainGames.
> Enrique Irazoqui who was attracted by BrainGames to organize the
> Qualifier proposed, "If you send DEEP-TIGER before April 23, 11:00 we
> will let it play in the Qualifier". This is an impossible demand and Mr.
> Irazoqui knows it! as no chess programmer is able to make things ready
> in just less than one day!
> Counter offers from our side failed, we proposed, We will send you the
> current DEEP-TIGER before your (impossible) deadline of April 23, 11:00
> to check its functionality, on April 27 you start the Qualifier as
> planned with the first match and on or before April 30 you will have the
> final version of DEEP-TIGER and thus there are no delays for you as on
> April 30 the first match is still not finished. A wall of granite was
> our reward, no flexibility on their end even though the situation was
> created because of their own errors, no invitation, withhold crucial
> information, the information boycott which caused a 7 day delay for
> Rebel Company to make the needed preparations.
I think "because of their own errors" is a bit strong. You aren't exactly
innocent here either... Secrecy can be a bad thing... Expecting someone
to either know something you have kept secret, or else to ask you about
something unknown, is really stretching things...
> During the negotiations Rebel Company was seriously offended by
> BrainGames as BrainGames on multiple occasions made serious doubts on
> the availability of a DEEP-TIGER version. This indirectly labels Rebel
> Company as a liar and is understood that way. Especially from such
> offending questions like these ones we got the impression, Rebel Company
> and Chess Tiger were just not welcome and that our talks were doomed to
> fail on beforehand.
Sorry but I would greet your claim of an SMP tiger with a lot of
skepticism myself. An SMP program takes months to develop. It takes many
more months to debug. It probably takes a year to become competitive.
Where was this program at the last WCCC, the most recent event where it
could have entered? Or why was it kept secret? How does that _possibly_
help your company? Practically all commercial companies have had a
parallel program for at least a year or two. Christophe disdained the
parallel search many times within the last year, claiming it was unnecessary
and was contrary to where mainstream computing was going. I would not have
guessed he was saying one thing and doing another. And the practice of
doing that ought not be rewarded, IMHO.
> Yet the worst thing had to come and says it all, the crucial question
> simply remained unanswered. Christophe Theron to BrainGames, "Why did
> you forbid Ed Schroder to talk with me about the "DEEP" requirement? Was
> there something I was not allowed to know?" No answer from BrainGames.
> The information boycott has cost Christophe Theron one week, exactly the
> week Rebel Company needed to make the software ready.
I repeat, "How did they prevent you from talking to Christophe?" That simply
sounds like something that couldn't be done, unless they had asked you to sign
a non-disclosure agreement _before_ they talked to you. Which I doubt.
> CONCLUSION
> Chess Tiger was not welcome and was banned on beforehand. Now the
> Qualifier is played between 2 ChessBase programs. This whole Qualifier
> is one big farce.
The qualifier is a farce. But not for the reason you state. Your secrecy
undid your chances. You made the business choice to be secret about the SMP
code. Christophe was on record as saying SMP was worthless. You were
"hoist on your own petard". The qualifier is a farce because there is a
current WCCC title-holder named Shredder.
> GIANT PROTESTS BY VARIOUS PEOPLE
> Right from the start BrainGames received major criticism in public
> forums in proportions of 10 to 1 at least, this regarding the selection
> procedure of the Qualifier. Even the major chess programmers have spoken
> against this Qualifier event and its unfairness, respectable names like:
> Professor Robert Hyatt, Cray Blitz and Crafty
> Bruce Moreland, Ferret
> Chris Whittington, Chess System Tal
> Ulrich Tuerke, Comet
> Stefan Meyer-Kahlen, Shredder
> Peter McKenzie, LambChop
> James Robertson, Insomniac
> Dieter Buerssner, Yace
> Ed Schroder, Rebel
> Christophe Theron, Chess Tiger
> All disagreed with the selection procedure, the Qualifier to play
> Kramnik.
Yes... but remember _why_. Not because a secret and unknown program was
not allowed. Rather because there was a qualifier at all, when we have a
WCCC and WMCCC champion in Shredder.
> Other known names have protested too, to name the most important one,
> David Levy, chairman of the ICCA. The ICCA is the key organization that
> unite and represents chess programmers and is responsible organizing the
> yearly official World championship, this successfully over a period for
> more than 25 years. The protest is IGNORED !!
Note again. David thought Shredder alone should play as the current ICCA
titleholder.
> More protests are coming in daily from chess journalists all over the
> world. Yesterday we received a so-called "open letter" from Jeroen
> Noomen a long time dutch chess journalist, computer chess expert since
> 18 years, one of the leading opening experts in the world. Jeroen Noomen
> has earned extra credit being the author of the opening books for Rebel
> since 1989. Jeroen has asked us to put his open letter on the Rebel Home
> Page and so we do.
That protest is less important. It is by someone that works on the book
for Rebel and Tiger. IE essentially another "programmer" who would be
expected to have interest in seeing RT entered.
<snip>
In short, this is one of those places where you make a decision and
then have to live with the consequences for a while. Secrecy has
helped commercial chess development over the years. It has hurt
overall chess development. In this case a policy "came home to
roost". If you keep something secret, you can hardly blame someone
when they are unaware of it.
--
Robert Hyatt Computer and Information Sciences
hy...@cis.uab.edu University of Alabama at Birmingham
(205) 934-2213 115A Campbell Hall, UAB Station
(205) 934-5473 FAX Birmingham, AL 35294-1170
Sam Simpson
Robert,
Out of interest - why is it hard to SMP-enable chess engines? I'd have
thought you could just split the search tree up into the number of threads
necessary to utilise the system (e.g. search threads = number of
processors).
I guess I'm missing something though............
--
Regards,
Sam
http://www.scramdisk.clara.net/
Robert Hyatt <hy...@crafty.cis.uab.edu> wrote in message
news:9ch8o4$29r$1...@juniper.cis.uab.edu...
<SNIP>
> Sorry, but I don't buy that. I sort of see why BG wanted to only include
> SMP programs. They are definitely stronger than their non-SMP
counterparts.
> Even more importantly, the one sponsor was going to provide an 8-way box
> to use for the event. Obviously a program that could use 8-way SMP was
> desiarable from their point of view.
<SNIP>
Robert Hyatt
> (Groups trimmed)
> Robert,
> Out of interest - why is it hard to SMP-enable chess engines? I'd have
> thought you could just split the search tree up into the number of threads
> necessary to utilise the system (e.g. search threads = number of
> processors).
> I guess I'm missing something though............
Lots of issues. Just the problem of splitting the tree into threads
is non-trivial due to shared data, locking, etc. There are other pitfals
as usually the serial algorithms used by many don't work well in a parallel
search. Repetition detection is one example. Many use the hash table.
This won't work well for parallel searches since all engines see all hash
entries.
Debugging is another issue. I am _still_ finding an occasional glitch
in Crafty. I found glitches in Cray Blitz 7 years after it was finished.
Alpha/beta is inherently sequential in nature, which is another problem.
> --
> Regards,
> Sam
> http://www.scramdisk.clara.net/
> Robert Hyatt <hy...@crafty.cis.uab.edu> wrote in message
> news:9ch8o4$29r$1...@juniper.cis.uab.edu...
> <SNIP>
>> Sorry, but I don't buy that. I sort of see why BG wanted to only include
>> SMP programs. They are definitely stronger than their non-SMP
> counterparts.
>> Even more importantly, the one sponsor was going to provide an 8-way box
>> to use for the event. Obviously a program that could use 8-way SMP was
>> desiarable from their point of view.
> <SNIP>
--
Ashton Anderson
is Chess Tiger and it's company's fault for holding their SMP version in
secrecy ... how could the Qualifier people know about it if it's secret?!
The qualifier is a farce, but so is your pathetic whining about how they
were at fault, and not you.
Ashton Anderson