Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: A Blast from the past - Feng Hsu Lost On Science

Author: Rolf Tueschen

Date: 03:50:05 04/28/05

Go up one level in this thread


On April 27, 2005 at 22:18:27, Robert Hyatt wrote:

>On April 27, 2005 at 18:06:01, Rolf Tueschen wrote:
>
>>On April 27, 2005 at 11:17:24, Lar Mader wrote:
>>
>>>On April 26, 2005 at 19:20:22, Rolf Tueschen wrote:
>>>
>>>>On April 26, 2005 at 18:13:31, chandler yergin wrote:
>>>>
>>>>>On April 26, 2005 at 17:09:08, Robert Hyatt wrote:
>>>>>
>>>>>>On April 26, 2005 at 16:49:48, chandler yergin wrote:
>>>>>>
>>>>>>>On April 26, 2005 at 16:39:22, Robert Hyatt wrote:
>>>>>>>
>>>>>>>>On April 26, 2005 at 16:01:57, chandler yergin wrote:
>>>>>>>>
>>>>>>>><big snip>
th>>>>>>>>
>>>>>>>>>h
>>>>>>>>>You really don't believe this?
>>>>>>>>>"No contemporary writer can give an accurate view of anything."
>>>>>>>>>
>>>>>>>>>"Only long after purported events as information is accumulated,  and
>>>>>>>>>the Historians assimilate the totality of the evidence, can a more accurate
>>>>>>>>>picture of what really happpened be provided."
>>>>>>>>>
>>>>>>>>>"This is true for War.. Politics, Stock Market, Religious thought,
>>>>>>>>>and 'Cultural' events."
>>>>>>>>>
>>>>>>>>>Then not only am I surprised, I'm appalled.
>>>>>>>>
>>>>>>>>
>>>>>>>>The be appalled.  I want to know what actually happened.  Now what someone
>>>>>>>>"thought" happened based on speculation, conjecture, rumor, fantasy, etc.
>>>>>>>>
>>>>>>>>History is a precisely recorded enumeration of events as they happen.  With no
>>>>>>>>"interpretation" or "justification" built in.  What you are wanting is "not"
>>>>>>>>history.
>>>>>>>>
>>>>>>>>I can figure out what happened by reading an accurate report about Little Big
>>>>>>>>Horn, or the Alamo.  I don't need any "interpretation" or "spin" thrown in to
>>>>>>>>confuse things.  Just an exact account of events.  That is history.
>>>>>>>
>>>>>>>What are 'accurate reports' without the totality of the evidence, and all
>>>>>>>viewpoints considered?
>>>>>>>
>>>>>>>Would you want to be on a Jury deciding life & death, without considering all
>>>>>>>the evidence.
>>>>>>>
>>>>>>>I don't think so.
>>>>>>>If so.. I wouldn't want you on my Jury.. regardless od what I was indicted for.
>>>>>>>
>>>>>>>I can't seriously you believe what you are saying Bob.
>>>>>>
>>>>>>
>>>>>>There are no "viewpoints" in history.
>>>>>
>>>>>Nonsense!
>>>>>
>>>>>
>>>>>> That is what you are missing.
>>>>>
>>>>>
>>>>>No, you are missing it.
>>>>>
>>>>>
>>>>>>"viewpoints" are opinion.  There is no opinion in history.  History is just a
>>>>>>factual recording of events as they take place, no opinion, no speculation, no
>>>>>>nothing.  A video-tape of an automobile wreck is a perfect example.  I don't
>>>>>>want _your_ opinion as to who was at fault, I want accurately recorded data that
>>>>>>I can use to make up my own mind about what I think about the event...
>>>>>>
>>>>>>Your jury analogy is _not_ valid.  Evidence is factual only.  Which is
>>>>>>historical in content.  I don't care what you think, what you thought you saw,
>>>>>>what you conjecture happened, etc.  As a jurist (and yes I have served multiple
>>>>>>times) I care only about facts.  That is what a jury does, "finding facts".  No
>>>>>>room for "opinion" or anything else in the jury room.
>>>>>
>>>>>Nor should there be.
>>>>>
>>>>>The "totality" of the evidence is what should determine a verdict.
>>>>>
>>>>>Until all the evidence is in, conclusions should not be drawn.
>>>>>
>>>>>That is my opinon and I stick with it.
>>>>
>>>>
>>>>You both used a language I wouldn't support, but in that specific problem I
>>>>agree with Chandler. It is very clear that Bob is making a judgement on a
>>>>limited view on the whole topic. In that limited perception Bob is absolutely
>>>>right, but if one is opening an overall, more whole view on the 1997 event, Bob
>>>>is wrong. By all means Bob wants to underestimate the importance of the human
>>>>client for the design Hsu et al had created.
>>>>
>>>>In this light Bob always wants to insinuate that this was "just" a sports event,
>>>>a match. In reality the event was a mutual attempt to evaluate the chess
>>>>capacities of DBII. It's trivial that if Kasparov was disturbed for playing his
>>>>usual chess, that the whole event was spoiled and the result was meaningless. In
>>>>the eyes of Bob Hyatt a game of chess is always of the same quality, no matter
>>>>if it's played for the Wch, during simultaneous exhibitions, in show matches or
>>>>in skittles in chess cafés or on tables in NY Central Park... The perception is
>>>>that Kasparov is _always_ playing the same quality of chess. So, under that
>>>>perspective it wouldn't really matter much how the IBM team treated Kasparov in
>>>>the show event.
>>>>
>>>>It's also reveiling a lot if you read Fernando's message. He's no less than a
>>>>sociologist from Chile. For him someone who's asking questions during a
>>>>show/research event must be paranoid. That is the level how we discuss things
>>>>here in CCC. If the best chessplayer of the time asks suspicious questions in
>>>>computerchess, he must be mad...! Something is going wrong here.
>>>
>>>
>>>0) I think Rolf has described the crux of the issue nicely here.  There are 2
>>>possible positions on the event:
>>>(a) If your expectation for the event was some sort of fair and controlled
>>>scientific experiment, then reasonable people might argue that the event was
>>>unfair.
>>>(b) If, on the other hand, you thing that the event was a "match", in the
>>>competitive sense, then clearly it was quite normal.
>>
>>Although one could be happy that people like Mader give their own analytical
>>input the fun gets a serious blow and is destroyed if someone gives his analysis
>>and comes to the provoking thesis that certain co-authors must be mentally ill
>>or such. We have just one famous member here who's always writing of mental
>>illnesses in opposing people. Now we have at least two, unless Mader isn't a
>>fake. I detest such insulting nonsense by calling other members here mentally
>>ill. But if such a famous man as Fernando can make such nonsense remarks than
>>others are allowed too...
>>
>>Apart from that negative aspect, Mader's analysis is quite interesting although
>>without real class.
>>
>>To begin with there are no such _two_ possibilities. Simply because it isn't
>>important what WE here expected but what Kasparov the important human half in
>>the experiment did expect. And Kasparov expected to attend a scientifical
>>experiment. Period.
>
>
>Can you provide a citation to support that?  _every_ mention I saw by Kasparov
>refered to "the match with Deep Blue".  Not "the experiment with DB" or "the
>scientific experiment with DB."  He _always_ said "the match".

Of course I can explain that. In special for the American market and for the
crowd over there but also worldwide, out of mutual agreement the show was called
a match or the match or the match of century or the match between Machine and
Man. That is forced speech if you attend to something in the USA. Below such
bogus terminology you can't hope for enough attention. You also call it World
Championship if in sports you have an event with Boston, New Orleans and
Milwaukee. You call World Championship when it's just the Californian Masters.
You called it World Championship in computerchess when you organised a weekend
for 5 university teams. So this alone is well known practice. A bit exaggerated
practice but who cares. I'm astonished that you take this at face value.

>
>
>
>>
>>From that point we can follow all the rest. Of course it wasn't Kasparov's duty
>>to care for exact science in the debates about the contract. So also this
>>argument is nonsense that Kasparov by not protesting against such and such did
>>allow such and such. That is nonsense. Science was the duty for the IBM team and
>>NOT Kasparov because Kasparov was the client of the experiment. He wasn't the
>>designer.
>
>No, he was a _participant_ in a chess match against the computer.  Nothing more,
>nothing less...  Again, find one citation where either IBM or Kasparov refered
>to the match as "an experiment" rather than as "the match" or "the match of the
>century."


Again, as a scientist you shouldn't confuse P/R and substantial facts. The first
is balyhoo and the facts are science. And it's true that Kasparov got thousands
of dollars just to make possible a worldwide P/R for the hosting company. But
Hsu and his team could test their machine which they had built up for decades.






>
>
>
>
>>
>>Everyone else than Bob Hyatt can well forget about science. But Bob is NOT
>>allowed to forget about his own field. Also decades of false practice doesn't
>>change the situation for Bob. Bob has to respect the scientific requirements of
>>such a test or call it show event.
>
>
>Or simply call it what it was?  "A chess match between Kasparov and Deep blue"
>just like the other Matches Kasparov played both against humans and other
>computer programs???


Bob, I am also interested in computerchess. I played thousands of games against
chess software on PC and chesscomputers, but I would never confuse human chess
and its tradition of centuries with computerchess emulated by machines... But of
course you have the disadvantage that you spent so many days and hours in
computerchess events so that you feel the need to revalue computerchess when
it's just the little brother or let's say nephew of human chess. I know that
such truths can hurt. But you can't change truths just for your personal benefit
or interest.




>
>
>
>
>> And if the IBM team violated certain basic
>>rules of scientific experiments, then a scientist must accept that as a fact.
>
>
>The problem is that you are assuming something is a fact when it is not.  This
>wasn't an "experiment" as in "a controlled scientific experiment."  If it were,
>they would have played a dozen matches to get the standard deviation down
>(caused by human inconsistency, health issues, etc).  This was a simple match.
>An experiment is repeated many times to confirm results.  A chess match is just
>a match, just like tennis matches, etc...

But this is all true. Who was responsible for the science of it? Kasparov? IBM?
Or the team around Hsu??? Control is not the same as standard deviations to be
right on the point. Control begins with exact protocols and already here we see
the mess. These scientists didn't care for protocols. For good reasons. Because
with too many protocols they would have shown what the were really doing. And
this way a scientific evaluation of DBII chess could have been made. But that
was NOT in the interest of the scientists. But as Kasparov said, he always
suspected that he was attending such an event. He couldn't foresee how it
rreally went and that the team around Hsu perverted the experiment, the pilot
study into something, we must call hoax. Just to win by confusing the opponent
about the identity of the player. This is Las Vegas but not science.


>
>
>
>A
>>scientist can't evade with the idea that this was a match. Of course the
>>experiment was embedded in the form of a short match. And Kasparov was the match
>>opponent. But anyway Kasparov also was the client, the scientific witness of the
>>experiment. He was, so to speak, the judge for the whole chess. And it's a
>>scandal that the team around Hsu insulted Kasparov the moment he asked simple
>>questions after game 2. He asked questions and showed much confusion. In that
>>situation the team should have cared for their human chessplayer. If not they
>>had destroyed the validity of their experiment. In other words the results
>>became meaningless! But NB always the triviality that it was NEVER Kasparov who
>>had to care for this validity. Kasparov was the human chessplayer and in the
>>contract the team had promissed NOT to disturb Kasparov's concentrationence. Keep
>>that in your mind if you read my arguments. All this was declared in advance.
>>And for the value of the science thing the team around Hsu should have behaved
>>like real scientists and NOT like crooks winning with ugly methods. This is all
>>so trivial!
>
>And so wrong...


It's strange that you oversee the science.


>
>
>>
>>
>>
>>
>>>
>>>BTW, Let's leave out the idea that there was actual cheating (human
>>>intervention) as it is unprovable, unknowable, not supported by fact, and
>>>basically boils down to conspiracy theory without evidence.
>>
>>Of course. But for a science experiment or research whatsoever it is desastreous
>>if you can't prove the validity of your results. It's not a desaster for
>>Kasparov who's getting his money all the same. But the science side suddenly no
>>longer has a result! If you know what I mean.
>
>
>Not a problem, since there was no "science" in this.  When I do "science" I
>repeat the tests many times to confirm results.  When I play matches on ICC, I
>just play matches and move on..


Then you have a rather limited view or understanding of scientific experiments.
You seem to be completely unaware of social science experiments which we have
here in the case of machine vs human chess. And perhaps Hsu et al had no clue
either. That would define them as poor scientists. But at least it would excuse
their naivety. The point is that you can't apply your repetitional experimenting
paradigma. Because what is done and over that is past. You can repeat it but it
will never be the same. Normally the human being has advantages no machine has
in such designs. What the team around Hsu was trying to prove that was the
superiority of a machine through the gambling and cheating (science) of the
operating human team. This is so basic! But at that moment the team around Hsu
no longer researched the chess of DBII but the irritation/ confusion of
Kasparov. Fantastic study! Unfortunately false topic!





>
> >
>>
>>
>>>
>>>1)  It is unreasonable to take position (a) above, that this was supposed to be
>>>a fair controlled experiment.  It was by every definition and action a chess
>>>match, with large sums of money involved, with clear upfront rules that defined
>>>an extremely competitive environment.  There has never been a chess match before
>>>that _wasn't_ of this nature, and psyching out your opponent is part of chess.
>>
>>
>>Nonsense. In human chess it's normal, but NOT in machine vs human chess. Because
>>who's psyching out who? Did the machine that job? I doubt that.
>>But since the chess of the machine was being tested, for much money, here I
>>agree, it makes no sense at all to psych out the human player. Because it
>>doesn't help for the evaluation of the chess of the machine. Period.
>
>DB didn't "psych out" anything.  If you look at the match, Kasparov psyched
>himself out.  I don't see how that is something IBM should have anticipated and
>tried to address...


True. DB psyched out nobody. But the DB team around Hsu psyched out their human
chessplayer just with so antique variables as impoliteness and neglection. I
admit that Kasparov was the best candidate for such a show. Because with his ego
he couldn't just continue to play chess but he also wanted to know what this all
had a meaning for...! Apparently he began to reveil the hoax Hsu et al were
into.




>
>
>
>
>>
>>
>>
>>>Kasparov happens to employ this effectively against other humans.  Why would
>>>this be different?  You are getting confused by the fact that scientists built
>>>the machine, and therefore by extension projecting some requirements for fair
>>>science on the match.
>>
>>Also nonsense. Because the scientist shouldn't have behaved properly because
>>they liked Kasparov but because with misbehaving they destroyed the validity of
>>their own results. Nobody without proper science education is forced to
>>understand what I'm teaching here... <cough>
>>
>>
>>
>>
>>> This, btw, was not part of the rules agreed upon.  Where
>>>in the rules is there _any_ mention of the needs of science here?
>>
>>
>>This is a joke. That the results would be hurt by a misbehaving towards the
>>human client in the thing - that is something nobody would expect in the
>>contract. It's selfunderstood for the value of the result. Of course if suddenly
>>the value is meaningless and winning ugly becomes the option THEN nobody must
>>care about this aspect... <cough>
>>
>
>
>cough all you want.  This was a match.  Just like the matches played over the
>last 100 years between top chess players.  No different except one player was a
>silicon-based chess player rather than carbon-based...


It costs me much energy to contradict a man who has shed so much heartblood for
suc events in computerchess. That is making me sad, because I don't want to
insult you. But you are so far from reality in this aspect that I am forced to
destroy your delusions. No, you never were in human chess. You never were in the
tradion of human chess. Simply because the practices in computerchess do violate
the rules of FIDE chess in multiple points. From the perspective of human chess
computerchess practices look like cheating. And you had decades to repair these
birth mistakes. But of course nobody had interests in doing that. The delusion
is so full of happy moments that it would hurt many people in their basics if
suddenly the standards in computerchess would be set straight. I would also
grant you and all computerchess people the benefit of doubt. They don't know
what really seperates them from human chess. But the moment you play human
players and it's really becomming tense - THEN you can see what I'm talking
about.

Just one example. For you over the decades it was good fun to surprise each
other with always new inventions and tricks and versions of the machines. You
played emulated chess and had fun. But human chess is completly different. Human
chess always relied on preparation. Like all human challenges and championships.
There is no such entity which grows up in hiding and then suddenly plays for the
Wch. You have to participate in order to get the chances to win titles. But by
definition you become known if you play in public. And that allows the others to
analyse you. All these analyses are part of human chess. And only human beings
who almost have no disadvantages or weaknesses can concur for titles. - The
point in computerchess is this: If chessplayers could analyse the machines in
details over months, they would discover extreme weaknesses and then they would
exploit these weaknesses. - By concentrating on short show events like 1997 you
in computerchess are hiding your head in the sand if you believe that the
reality you can fake in such events is also reality in the human chess
environment. It is not. It's really funny to assume that it could be reality. -
Of course this is hurting many people in computerchess but still it's true. As a
scientist you must accept that.




>
>
>
>>
>>>In fact, if
>>>this was the case, eveything about the match would have been conducted
>>>differently.  For example, they would have ensured much more rest between games
>>>for the human, or at least provided the option for such a thing should the human
>>>feel the need _in_the_rules_.  You guys need to take a step back and let all of
>>>this sink in.
>>
>>I am convinced that you should better take a step back because you have no clue
>>about science. Kasparov himself had declared that if this had been for the
>>championship that then the rules should have been different. But for a science
>>test the whole thing could well be held this way IF and ONLY IF the team didn't
>>forget about their hosting duties for the human client! Because you want to
>>measure chess strength of the machine against a human. But if you let the humanm
>>player get confused, you no longer measure chess strength of the machine but
>>confusion of the human player. And this would be pure nonsense to do for over a
>>million dollars, not to forget... <cough>
>>
>>
>>
>>>
>>>Here's a quote from Rolf above:
>>>>In the eyes of Bob Hyatt a game of chess is always of the same quality,
>>>>no matter if it's played for the Wch, during simultaneous exhibitions,
>>>>in show matches or in skittles in chess cafés or on tables in NY Central
>>>>Park... The perception is that Kasparov is _always_ playing the same
>>>>quality of chess.
>>>
>>>2) Bob certainly doesn't think that Kasparov is _always_ playing the same level
>>>of chess no matter what the circumstances of conditions, and he has clearly said
>>>so.  Bob simply considers this point somewhat irrelevant to the claims of
>>>unfairness, as he ascribes to position (b) above.
>>>Here is a quote from Bob acknowledging that human frailties are indeed a problem
>>>for the quality of human play:
>>>
>>>>>When you play chess, you have to play no matter how you feel
>>>>>on the day of the match.  Personal problems?  Deal with them.
>>>>>Illness?  play or forfeit.  You simply have to play, period.
>>>>>And that is a part of chess.  I've played tired, because I
>>>>>had to drive a long distance prior to the start of round 1.
>>>>>I've played tired because of staying up too late the night
>>>>>before the round. I've played with a fever.  That's just a
>>>>>part of chess.  Not a part that the computer has to deal
>>>>>with of course, but then computers lose due to remote power
>>>>>failures, or communication failures, or outright program bugs,
>>>>>etc...
>>>
>>>It is truly disingenous for Rolf to say Bob thinks all human chess play is the
>>>same quality, as it is painfully obvious that he repeatedly _does_not_.  In
>>>fact, I think you'd be hard pressed to find anyone on the planet that thinks
>>>humans always perform at the same level.  Again, what is wrong with you?  Let's
>>>think about that for a moment.  Frankly there is a _lot_ of this kind of
>>>misrepresentation going on in these threads by Rolf and chandler?  Are you guys
>>>really interested in the discussion, or just arguing to piss people off?  Seeing
>>>so much of this stuff, one has to wonder if you you are either intentionally
>>>dishonest or mentally ill.  Your credibility suffers with this kind of thing,
>>>and overtime leads one to conclude that there really is something wrong with
>>>you.
>>>
>>
>>
>>THere is a different explanation for the messages of Chandler and Rolf. Namely
>>that both have a deeper reflexion on this 1997 event than Bob who's massively
>>biased in favor of IBM/team around Hsu.
>
>
>No.  It is just that _I_ have participated in these kinds of events.  I played
>such a match against David Levy in 1984.  SO I have _been there_.  I have played
>other matches against strong humans.  All were carried out just like the match
>with DB and Kasparov.  Nothing different at all.  I don't see how someone that
>has never "walked the walk" can go on and on about how things ought to have
>been, without understanding how things _always have been_...


I think I could show you how wrong you are with that belief. All what you did in
the past decades (in computerchess) this has almost nothing to do with human
chess. And I confirm you that this doesn't take away any single detail of your
successes and status. Only now you can't do this: namely confusing your
computerchess practices with human chess and its centuries of experience. The
crime of the century of Hsu et al was that they cheated on science to get an
ugly win against the best human chessplayer. A chessplayer who naively was
believing that he attended a scientifical experiment to find out the chess
strength of a machine. And he wasn't prepared for a hoax where a company had
financial interest in beating him by all means - without any fairness and
scientific moral.






>
>
>
>
>>
>>It's no big deal that Mader diesn't understand the implication of Rolf's
>>statement. But here again an explanation: if Bob is unable to see the importance
>>of unconfused states for the human chessplayer in the experiment, then Bob makes
>>the implication that no matter how Kasparov played his chess would be the
>>correct lackmustest for the chess strength of Deep Blue. Of course such a
>>proposition is absolute nonsense. Because if Kasparov is confused, he could be
>>easily or easier be beaten. But that would tell us much about the chess of the
>>machine. That is all very tricial. Scientist will understand what I'm talking
>>about. And that is all what counts here on CCC. Chapters about mental illnesses
>>might well be written in CTF; BTW I'm writing in CTF myself for years and it's
>>great fun. But most of the time it's no science. It's much too much political
>>propaganda.
>>
>>
>>
>>>3) I don't think this is really germaine to the discussion, but I'd like to
>>>address this discussion on history.  The kind of objective history that Bob
>>>describes is somewhat ideal, and rarely exists in the real world.  The problem
>>>is that for just about any event, the "facts" about what happened may be in
>>>dispute, may be incomplete, human observers are fallible, human memory is
>>>fallible, etc.  If we had complete, provably unedited video footage of all
>>>events from several angles, along with audio, etc, perhaps we could say that
>>>there is a one true history, but even then there will be disputes about the
>>>"facts".  The world, humans, observation, events, etc... are complex things, and
>>>the idea the we can have perfect history is _usually_ not achievable.  That
>>>being said, all one can do is attempt to gather the objective facts such as the
>>>are, and deal with any conflicting versions of events as they may come up.
>>>However, in this DB event, there really aren't any disputes about the facts.
>>>Everyone agrees on what happened, the computer _appeared_ to make some moves
>>>that Kasparov was surprised by.  Now saying that the computer didn't _really_
>>>play these moves and that someone intervened, this is where we reach the edge of
>>>the factual data available.  Unless someone admits to cheating, or is somehow
>>>caught, there is no way to know.  This part of history can't be proven.  But
>>>again, it is also important to note that there is no evidence to support this
>>>claim, and therefore it must be dismissed until there is.
>>>
>>>-= Lar
>>
>>
>>
>>Also in these final paragraphes Mader missed the topic. It's NOT the issue to
>>find answers who had played for DBII. The problem we discussed here was why the
>>team didn't talk to their human client when he had these suspicions. No one, mot
>>Chandler nor Rolf did ever imply theories about the origines of the real player
>>behind DBII. All that has been criticised is the misbehavior towards Kasparovs
>>when the IBM teamsters called Kasparov mad, after he had asked his questions
>>after game 2. And this is a clear violation of all science. Because from that
>>moment on the scientists did no longer research chess strength of DBII but the
>>chess of a confused Kasparov. Honestly. If they would have wanted to research
>>confusion of a human player against a machine, Hsu could well have organised me,
>>Rolf Tueschen, for say 2000 dollars and he would have got the same picture.
>><tongue in cheek>



This page took 0.01 seconds to execute

Last modified: Thu, 15 Apr 21 08:11:13 -0700

Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.