Computerchess Misc (1)
Rolf Tueschen
A couple of days ago I saw the following position for the first time.
W Kh1 Qe2 Rf1 Rg3 Pa3 c3 f5 g2 h2
B Kg8 Qc5 Rb8 Bc6 Nd3 Pa7 c4 f7 g6 h7
Questions:
0. Warning! Everybody who's not capable to solve this position either for
White or Black without the help of computers will never more become a
Grandmaster!
1. What is this a position for
a better White?
b better Black?
c won White?
d won Black?
2. What pattern would a *human* see at first?
a for White?
b for Black?
3. What does your favorite computer program see? Evaluation? Now play the
favorite move. Evaluation better or worse?
4. Which move does your program prefer for Black?
5. What do you think who won the game in the end? Note it was not a zeitnot
duell. The outcome is already to be seen for any Grandmaster. Caution:
Please don't ask Roman or Father Bill on ICC --- this would be a clear
cheating!!!
6. What is your guess? Who were the players? Please all GMs, IMs and
database freaks and businessmen, keep the secret to you. Thanks.
7. For those who don't know exactly, what were the players like:
a Elo under 1400?
b Elo 1400-1600
c Elo 1600-1800
d Elo 1800-2000
e Elo 2000-2200
f Elo 2200-2400 USCF Master
g Elo 2400-2500 IM
h Elo >2500 GM
i Elo >2700 Super GM
Many greetings to all
Rolf Tueschen
Rolf Tueschen
>Computerchess Misc (1)
>A couple of days ago I saw the following position for the first time.
>W Kh1 Qe2 Rf1 Rg3 Pa3 c3 f5 g2 h2
>B Kg8 Qc5 Rb8 Bc6 Nd3 Pa7 c4 f7 g6 h7
>Questions:
>0. Warning! Everybody who's not capable to solve this position either for
>White or Black without the help of computers will never more become a
>Grandmaster!
Warning! I will never become a good chess journalist because I forgot to
mention that of course it was Black on the move.
mclane
Main line: 1...Rb2 2.fxg6 hxg6 3.Qd1 Qf2 4.Rg1 Qxg3 5.hxg3 Nf2+ 6.Kh2
Nxd1
after 284 seconds the score drops to 5.96
1...Rb2 2.fxg6 hxg6 3.Rxg6+ fxg6 4.Qe6+ Kg7 5.Rf7+ Kh6 6.Qh3 Qh5 7.Qe3
best wishes
mclane
Chris Whittington
Rolf Tueschen wrote in message <6rl7df$dls$1...@news02.btx.dtag.de>...
>Computerchess Misc (1)
>
>
>A couple of days ago I saw the following position for the first time.
>
>
>W Kh1 Qe2 Rf1 Rg3 Pa3 c3 f5 g2 h2
>B Kg8 Qc5 Rb8 Bc6 Nd3 Pa7 c4 f7 g6 h7
NB black to move ......
>
>Questions:
>
>0. Warning! Everybody who's not capable to solve this position either for
>White or Black without the help of computers will never more become a
>Grandmaster!
>
>1. What is this a position for
> a better White?
> b better Black?
yes on basis of 'just a glance'
> c won White?
> d won Black?
yes ditto
>
>2. What pattern would a *human* see at first?
> a for White?
at a glance:
rook file attacks on white king, possibly hitting stonewall of pawns, no
white pressure on dark squares, so black king probably safe, all white
material committed to attack. nothing further to add, sac possible with
fxg6, then Rxg6. Except it's not white's move.
> b for Black?
at a glance:
dynamic spring in position, still some pieces can maneouvre in and become
more active, potential threats from pin, smothered mate, back rank mate.
plus one more material unit and maneouvre potential probably decisive. Need
to watch out for sac on g6 though.
>
>3. What does your favorite computer program see? Evaluation?
Chess System Tal would play still play Rb2 +8.91 even after about 30 secs on
PP400.
Incedentally CSTal would have factored all the above 'analysis' in at ply
zero evaluation (ie without any search).
>Now play the
>favorite move. Evaluation better or worse?
Gets worse, but still well positive.
>
>4. Which move does your program prefer for Black?
I assume you mean white ? fxg6 after Rb2
>
>5. What do you think who won the game in the end? Note it was not a zeitnot
>duell. The outcome is already to be seen for any Grandmaster. Caution:
>Please don't ask Roman or Father Bill on ICC --- this would be a clear
>cheating!!!
very unclear. whether or not the sac attack would work without taxing my
little brain too much. may turn on white not getting control of dark squares
(maybe), and not having a tempo move because of the white back rank mate
threat, (or smothered threat after white vacates defence of f2)
>
>6. What is your guess? Who were the players? Please all GMs, IMs and
>database freaks and businessmen, keep the secret to you. Thanks.
>
No idea. If white is ok in the position, then he is a strong player
(assuming he saw it all). And if white is ok, then black may well be a
program counting the material beans.
>
>7. For those who don't know exactly, what were the players like:
>
> a Elo under 1400?
> b Elo 1400-1600
> c Elo 1600-1800
> d Elo 1800-2000
> e Elo 2000-2200
>
Above this level at least ......
> f Elo 2200-2400 USCF Master
> g Elo 2400-2500 IM
Possibly above this if white wins and knew it.
> h Elo >2500 GM
>
> i Elo >2700 Super GM
>
>
Ok, I stuck my neck out. Waiting for the chop-chop :)
>
>Many greetings to all
Likewise,
Chris Whittington
>
>Rolf Tueschen
>
Rolf Tueschen
solutions yet. If you could go a litle bit deeper into the position you
would observe that it's a good one for the whole debate human knowledge in
computer programs. I think the whole difference could be well demonstrated
between our human chess playing and the still very weak computerized chess.
As for fca he gave an almost master-like in depth analysis but he forgot to
answer all my questions. But still if he found his lines without computers
he must be a very strong player. One question reveiled the whole difference
between human chess and computer v. computer chess. fca asked that possibly
a computer could find a hole in his lines ...
That's exactly the problem. And that was the point of Kasparov's paranoid
failure against DB. If you think too much about the possible almost perfect
play of a computer in a certain limited space, you become nuts because you
always have to be careful not to miscalculate. The solution of the position
will be a beautiful explanation how humans play chess. Also we could find
some novelties for our main topic here in the computerchess group. Players
who normally visit the analysis group might find different data.
My theory (a weak theory still) says that if you always analyse positions
with computer aid you lose your human advantage to be able to judge.
Because it's a psychological knowledge that all kind of good or casual data
may lead you on wrong paths. Often it's very difficult to leave all these
prepositions behind. I remember that Bob but also Ingo A. have said that
there overall tactical strength was worsened by their working with
computerized chess.
BTW this as a little feedback. Both players were human players, so no
computer involved.
I still wait for a second answer from mclane about something strange things
happening, no? I mean the position isn't that easy?
And Bob Hyatt isn't interested at all? Would be a pity. Because I have good
news from Crafty. Probably the position has to suffer because it was
presented by the devil? GENS UNA SUMUS.
"Chris Whittington" <chr...@cpsoft.demon.co.uk> wrote:
>Rolf Tueschen wrote in message <6rl7df$dls$1...@news02.btx.dtag.de>...
>>Computerchess Misc (1)
>>
>>
>>A couple of days ago I saw the following position for the first time.
>>
>>
>>W Kh1 Qe2 Rf1 Rg3 Pa3 c3 f5 g2 h2
>>B Kg8 Qc5 Rb8 Bc6 Nd3 Pa7 c4 f7 g6 h7
>NB black to move ......
Of course. Proves that you and fca are very strong players. Because this is
not trivial, no?
[..]
Chris Whittington
Rolf Tueschen wrote in message <6rmr2n$9g7$1...@news01.btx.dtag.de>...
>Chris, thanks for the good answer. Although I don't want to give the
>solutions yet. If you could go a litle bit deeper into the position you
>would observe that it's a good one for the whole debate human knowledge in
>computer programs. I think the whole difference could be well demonstrated
>between our human chess playing and the still very weak computerized chess.
It's not just a good one, it got it in one.
And showed the virtue of 'asking questions' (c) Rolf Tueschen
>
>As for fca he gave an almost master-like in depth analysis but he forgot to
>answer all my questions. But still if he found his lines without computers
>he must be a very strong player. One question reveiled the whole difference
>between human chess and computer v. computer chess. fca asked that possibly
>a computer could find a hole in his lines ...
>
>That's exactly the problem. And that was the point of Kasparov's paranoid
>failure against DB. If you think too much about the possible almost perfect
>play of a computer in a certain limited space, you become nuts because you
>always have to be careful not to miscalculate.
Well pointed out. But it also shows another thing or two -
that the bean-counters (sorry, the comp scientists who lack chess
understanding, also many end-user testers) no longer look at the single
position; instead put it (the position) into the program (or programs, if
they have them) and let the programs deliver their verdict.
Programs as substitute for thought.
And worse, the programs only deliver tactic lines of play, as if the only
importance is the tactic lines.
Do you see what I mean now when I say they play/think tactical, they say
'positional is long-range tactics'; that they haven't moved up a level to
'play/think positional' ? For them chess is all 'best lines' (tactical). And
not positions (positional).
They were bad at seeing, so they made/bought programs. And the programs led
them down the best line and blinded them further.
The solution of the position
>will be a beautiful explanation how humans play chess. Also we could find
>some novelties for our main topic here in the computerchess group. Players
>who normally visit the analysis group might find different data.
>
>My theory (a weak theory still) says that if you always analyse positions
>with computer aid you lose your human advantage to be able to judge.
Yes. Also by laziness and lack of practice at thinking it for yourself, too.
>Because it's a psychological knowledge that all kind of good or casual data
>may lead you on wrong paths. Often it's very difficult to leave all these
>prepositions behind. I remember that Bob but also Ingo A. have said that
>there overall tactical strength was worsened by their working with
>computerized chess.
>
Absolutely. We all said it.
>
>
>BTW this as a little feedback. Both players were human players, so no
>computer involved.
>
>
>I still wait for a second answer from mclane about something strange things
>happening, no? I mean the position isn't that easy?
>
>And Bob Hyatt isn't interested at all? Would be a pity. Because I have good
>news from Crafty. Probably the position has to suffer because it was
>presented by the devil? GENS UNA SUMUS.
One reason he stayed out. But not the only one.
Hyatt would not risk his reputation on hazarding chess positional viewpoints
on this position. He lacks the confidence in his chess knowledge judgements.
Normally Hyatt waits until others have pronounced first in these cases. Or
quotes his own program's evaluation/best line combination.
And, the lack of response from any bean-counter is indicative of the level
of chess knowledge within their program evaluation function.
As Vincent has so eloquently pointed out, for the bean-counter paradigm
evaluation function without search, the opening move e4, loses to the
brilliant reposte h5 !!
Doh !
And this, despite the doubling of knowledge and halving of nps of Crafty
over the last year :))))
Chris Whittington
mclane
>that the bean-counters (sorry, the comp scientists who lack chess
>understanding, also many end-user testers) no longer look at the single
>position; instead put it (the position) into the program (or programs, if
>they have them) and let the programs deliver their verdict.
>Programs as substitute for thought.
Even worse, in a famous computer-chess-magazine, they publish
SOLUTIONS of a corner called RAETSELSCHACH...
and give a price if a reader "finds" the solution. But nobody, wether
the author of the column nor the readers really find them by thinking,
they look into sources or use their computers...
I have seen them giving prices for readers who repeated the same wrong
soultion, because they looked up in old books or had the same
computer-program... nobody thinks. They all let "compute".
>And worse, the programs only deliver tactic lines of play, as if the only
>importance is the tactic lines.
>As Vincent has so eloquently pointed out, for the bean-counter paradigm
>evaluation function without search, the opening move e4, loses to the
>brilliant reposte h5 !!
>Doh !
>And this, despite the doubling of knowledge and halving of nps of Crafty
>over the last year :))))
>Chris Whittington
best wishes
mclane
Rolf Tueschen
>that the bean-counters (sorry, the comp scientists who lack chess
>understanding, also many end-user testers) no longer look at the single
>position; instead put it (the position) into the program (or programs, if
>they have them) and let the programs deliver their verdict.
>Programs as substitute for thought.
Again, I still keep the solution secret. Nothing really new appeared after
yesterday's posts. It's a pity, I repeat, that many who normaly jump in
with "their" lines (of their computer programs of course) stay away from
this interesting position.
Let me talk about the reasons *why* human players, who make use of
computerchess "verdicts", gradually are deteriorating their own chess
judgement.
Let me introduce a very interesting knowledge about human learning or
knowledge as such. There is a complete difference between "active" and
so-called "passive" knowledge. We all have seen it.
I.E. you read a book and normally you can say at the end of your lecture
"interesting stuff, *now* I finally understood that XYZ, quite easy,
wonderful, now I will read the other three references the author has given,
now I'm a real expert of XYZ (...) perhaps I will give a lecture about it
in my study class, but before I want to talk about it with my friend".
Now, please, let's take into consideration just for a short moment, that we
were talking about "normal" people and *not* those very few with eidetics
(read other posts I wrote about it).
If you are just a normal guy (or maid of course) then you surely did make
the following astonishing discovery. The moment you wanted to talk with
your friend about it you were very sure that you had understood the new
topic. And now when you wanted to start talking you realize immediately
that you have'nt understood that topic *at all*. You begin to stumble, to
scratch your head and so on. But you were a very clever guy. simply because
you gave yourself the opportunity to detect the difference between active
and passive knowledge. Just having read about a new theory, does mean
you've learned the details of a new theory and you were able to understand
it, but this doesn't mean at all that you are able to explain the new
theory e.g. like the author of the book you just have read.
And the even more shocking news is that the amount of work you had to
dedicate to this task (preparing yourself to be able to explain [active
mode] the new theory) something between the double or triple amount of time
(of the needed time to read and understand passively the new theory).
If this is true, then the bad effect of just combining your own thought
process with the (instant) output of a machine based search is obvious.
To give another example to yours. Each time I read the story about Bob
Hyatt who just had his nightly "talks" with a lot of GMs (!!), I ask myself
why Bob is so naive to believe that he really is "talking" to YZ GM ...
It's my opinion that this is simply not possible. But I don't doubt for a
moment that Bob Hyatt really is "talking" with his interlocutor. Also it's
my opinion that Bob is able to simulate a smart discussion. If you had
years to become adapted to such a simulation you could do the same.
But I don't believe at all that this "talking" is that of a really
"understanding" person with a GM. Sure the simulation is easier if the two
are used to talk together for years (as Bob did it with GMRoman). This is a
well known knowledge in psychology as any marriage guidance coucelor might
explain to you. In the end the conversation of the two concerend is
non-existent although both parties "talk" enough with well known
vocabulary. But both would answer to the councellor "no, he (she) doesn't
understand at all what I'm trying to say".
I simply can't get it that Bob Hyatt always peretends that he's having so
many (meaningful) talks with a lot of GMs. How could a 1400 (ICC 1900) ever
understand what a GM 2500-2600 is talking about? Assumed that the GMs are
able to express their knowledge more or less accurately ... GM Roman e.g.
is not a world-wide-known expert in teaching chess. I think that even
Karpov would do a better job. At least he has written some books about
strategic patterns.
It's so obvious that all this is "poppy cock" (original TM of Bob Hyatt). A
weak amateur trivially can't cope with GM "talking". At best he could
simulate such a mutual talk/understanding.
You spoke about another myst. Looking at the eval of your program, you can
omit the usual crap of an amateur. Crap that could be shown as incorrect at
the forth ply at the latest. That's the whole trick. It's simply dubious to
talk on the base of a program's output as if you had the ability to judge
about chess positions.
Another point is the somewhat cheap talking at home miles away form OTB
chess.
Take for example fca's analysis of the Rb2 ... Rxg6 fg Qe6 Kh7! Qe3 Qh5 Qe3
Nf4!! line. True, Nf4 in the end is the key for the answer if Rb2 is
winning. But do you believe that fca would see Nf4 immediately with Rb2?
But that's not the only question. I posted the position primarily into a
computerchess group.
Now, I still wait for those who inform us about strange things to happen
with computers in certain important positions. *Sure* all the programs *I*
could test showed immediately Rb2 with a overwhelming score --- but it was
my impression *for wrong reasons*!
Take for example mclane's CSTal (Qe6) Kg8. This is a draw! Period.
Fritz5 in CB7 shows up to ply 17/17 (Qe6) Kh8 (which is also a draw!).
Then after some 15 minutes I stopped it. The same with Junior4.6.
Up to ply 19 (we know that it's some strange output) Junior prefers (Qe6)
Kh8 ...
ONLY Crafty shows Kh7 after 2'12! -- -- Let's see if Bob will show his
head right now. :))
Now -- let's think again about a human chessplayer who has to play OTB.
Sure, if you once saw that Nf4 on your board/display you are favoring Rb2.
But who's able to see Nf4 OTB without moving pieces to look if the line is
ok for Black? That was my question for the Elos. And also here you, who
answered the questions as the only one, were quite right. White can't be a
GM because he didn't see early enough that his attack on g6 wouldn't win
the game. But what with Black? -- Therefore I asked you to look a little
bit deeper. perhaps with the commentary I gave here you might try a final
analysis before I give you all the answers?
And you already got the Gold Medal, Chris. Also because you head the guts
(and the knowledge!) to show your head. I think this is one of the reasons
you are an outsider among the weaker chess players who actually still have
more successful machines/babies.
Let me give a final "psycho" statement.
I'm dissapointed about the overall non-smartness of the cc experts. Take
yourself and your image. Only because you (as a comparatively strong
chessplayer) are convinced that computerchess could need a lot more chess
knowledge to be able to *really* cope with GM chess, not with the shown GM
output in funny skittles for some thousand bucks ..., you are an outsider.
On the other side you could easily take Crafty source code, like Bruce (?),
add this or that and then following the known paradigm. *Then* you were a
nice guy. How primitive such a logic. You chose the hard way, the obviously
more difficult one, and for sure, actually the less successful. Instaed of
admiring your guts and firmness they laugh at you ...
The same with me as a newbie. If I would show groupie-like behavior, I
would be accepted as another friendly brown-noser. People simply can't
estimate the value of an opponent who (also as a newbie) could give some
interesting ideas. But he's only able to create them in constant
criticising of the experts (which seems to be a contradiction in itself
because he's judged as a newbie). "Interdisciplinary" approaches seem to be
widely unknown ... Althiough the experts make use of different experts in
their teams. But then they don't regard their helpers as rela newbies of
course. :)
I dream of a brainstorming group where also those could be accepted who can
only make their contributions in a creative opposition to the experts.
That's BTW the usual history in all sciences. Often the direct expert-like
secondary experts are too close to their chiefs to be able to create
totally *new* ideas. Or let's better say "unusual", normally "not-accepted"
ones.
A really superior expert would be someone who were able to tolerate also
the most aggressive critics. Not because he's a weak personality. But
because he's so strong!
When I saw the position in question I had suddenly all such ideas in mind.
It's a very important position.
>And worse, the programs only deliver tactic lines of play, as if the only
>importance is the tactic lines.
>Do you see what I mean now when I say they play/think tactical, they say
>'positional is long-range tactics'; that they haven't moved up a level to
>'play/think positional' ? For them chess is all 'best lines' (tactical). And
>not positions (positional).
>They were bad at seeing, so they made/bought programs. And the programs led
>them down the best line and blinded them further.
Your operator seems to prove the contrary. Way beyond being a GM himself he
also seems to understand that some "strong" programs often play really
stupid chess ... So I don't buy that people like Bob or Bruce have to be
forced to behave the way you described. Also Bob did state very frequently
that Fritz e.g. still is not all a GM player ... Not even close.
>The solution of the position
>>will be a beautiful explanation how humans play chess. Also we could find
>>some novelties for our main topic here in the computerchess group. Players
>>who normally visit the analysis group might find different data.
>>
>>My theory (a weak theory still) says that if you always analyse positions
>>with computer aid you lose your human advantage to be able to judge.
>Yes. Also by laziness and lack of practice at thinking it for yourself, too.
>>Because it's a psychological knowledge that all kind of good or casual data
>>may lead you on wrong paths. Often it's very difficult to leave all these
>>prepositions behind. I remember that Bob but also Ingo A. have said that
>>there overall tactical strength was worsened by their working with
>>computerized chess.
>>
>Absolutely. We all said it.
I read it only from the quoted two.
>>
>>
>>BTW this as a little feedback. Both players were human players, so no
>>computer involved.
>>
>>
>>I still wait for a second answer from mclane about something strange things
>>happening, no? I mean the position isn't that easy?
>>
>>And Bob Hyatt isn't interested at all? Would be a pity. Because I have good
>>news from Crafty. Probably the position has to suffer because it was
>>presented by the devil? GENS UNA SUMUS.
>One reason he stayed out. But not the only one.
>Hyatt would not risk his reputation on hazarding chess positional viewpoints
>on this position. He lacks the confidence in his chess knowledge judgements.
Usually he's not so careful. The most recent example was the Rebel v. Anand
game with the BB unequal material.
>Normally Hyatt waits until others have pronounced first in these cases. Or
>quotes his own program's evaluation/best line combination.
Yes, I knew this before. And to make a more difficult contest I added my
further questions who then had shown the limited views. BTW your operator
followed exactly this line. He gave a short varialtion and didn't see that
it was a draw. So he acted in reality like a "bean-counter" at least in the
sense you tried to explain it.
>And, the lack of response from any bean-counter is indicative of the level
>of chess knowledge within their program evaluation function.
Here I see a compination of a fear to get involved into another Rolf trick.
But I will show that it's not a trick at all.
>As Vincent has so eloquently pointed out, for the bean-counter paradigm
>evaluation function without search, the opening move e4, loses to the
>brilliant reposte h5 !!
>Doh !
>And this, despite the doubling of knowledge and halving of nps of Crafty
>over the last year :))))
I'm more thinking about the chess-freak GMRoman. As we have still 99945
chess patterns up in our sleeves we have a life-long assurence to be
related to the money "involved" constant flow of "communication" ... ;-)
I still ask myself who's KingCrusher on ICC?
Chris Whittington
Rolf Tueschen wrote in message <6rp1f9$prf$1...@news00.btx.dtag.de>...
And the fools did censor Rolf Tueschen ..........
Yup. But it took you to tell us. And the fools censored Rolf .......
>
>To give another example to yours. Each time I read the story about Bob
>Hyatt who just had his nightly "talks" with a lot of GMs (!!), I ask myself
>why Bob is so naive to believe that he really is "talking" to YZ GM ...
>
>It's my opinion that this is simply not possible. But I don't doubt for a
>moment that Bob Hyatt really is "talking" with his interlocutor. Also it's
>my opinion that Bob is able to simulate a smart discussion. If you had
>years to become adapted to such a simulation you could do the same.
>But I don't believe at all that this "talking" is that of a really
>"understanding" person with a GM. Sure the simulation is easier if the two
>are used to talk together for years (as Bob did it with GMRoman). This is a
>well known knowledge in psychology as any marriage guidance coucelor might
>explain to you. In the end the conversation of the two concerend is
>non-existent although both parties "talk" enough with well known
>vocabulary. But both would answer to the councellor "no, he (she) doesn't
>understand at all what I'm trying to say".
Yup. Been there. Done that :)
>
>I simply can't get it that Bob Hyatt always peretends that he's having so
>many (meaningful) talks with a lot of GMs. How could a 1400 (ICC 1900) ever
>understand what a GM 2500-2600 is talking about? Assumed that the GMs are
>able to express their knowledge more or less accurately ... GM Roman e.g.
>is not a world-wide-known expert in teaching chess. I think that even
>Karpov would do a better job. At least he has written some books about
>strategic patterns.
>
>It's so obvious that all this is "poppy cock" (original TM of Bob Hyatt). A
>weak amateur trivially can't cope with GM "talking". At best he could
>simulate such a mutual talk/understanding.
>
>
>You spoke about another myst. Looking at the eval of your program, you can
>omit the usual crap of an amateur. Crap that could be shown as incorrect at
>the forth ply at the latest. That's the whole trick. It's simply dubious to
>talk on the base of a program's output as if you had the ability to judge
>about chess positions.
>
>Another point is the somewhat cheap talking at home miles away form OTB
>chess.
>
>Take for example fca's analysis of the Rb2 ... Rxg6 fg Qe6 Kh7! Qe3 Qh5 Qe3
>Nf4!! line. True, Nf4 in the end is the key for the answer if Rb2 is
>winning. But do you believe that fca would see Nf4 immediately with Rb2?
No, I didn't believe it when I saw it. I thought (and I might be wrong and
misjudging), I thought he's played the line out on a comp and then wrote the
analysis. Nf4 is very difficult, almost out of the question, to see "at a
glance". GM standard at least. Is he a Yotam clone ?
>
>But that's not the only question. I posted the position primarily into a
>computerchess group.
>
>Now, I still wait for those who inform us about strange things to happen
>with computers in certain important positions. *Sure* all the programs *I*
>could test showed immediately Rb2 with a overwhelming score --- but it was
>my impression *for wrong reasons*!
>
>Take for example mclane's CSTal (Qe6) Kg8. This is a draw! Period.
must play Kh7, the comps see this after a while ....
>
>Fritz5 in CB7 shows up to ply 17/17 (Qe6) Kh8 (which is also a draw!).
>Then after some 15 minutes I stopped it. The same with Junior4.6.
>Up to ply 19 (we know that it's some strange output) Junior prefers (Qe6)
>Kh8 ...
>
>ONLY Crafty shows Kh7 after 2'12! -- -- Let's see if Bob will show his
>head right now. :))
>
>Now -- let's think again about a human chessplayer who has to play OTB.
>Sure, if you once saw that Nf4 on your board/display you are favoring Rb2.
>But who's able to see Nf4 OTB without moving pieces to look if the line is
>ok for Black? That was my question for the Elos.
Fantastic high to see it, right.
>And also here you, who
>answered the questions as the only one, were quite right. White can't be a
>GM because he didn't see early enough that his attack on g6 wouldn't win
>the game. But what with Black? -- Therefore I asked you to look a little
>bit deeper. perhaps with the commentary I gave here you might try a final
>analysis before I give you all the answers?
I've seen too much position and looked at it with CSTal, playing out the
lines. If black, say picked up an offered sac (since he was material up)
having seen all this stuff, knowing that has was safe from more saccing on
g6 after Rb2 , then black is seriously strong.
Unless you have a joker up your sleeve and white has something, somehow,
somewhere. Always possible. Anand did against Rebel.
>
>And you already got the Gold Medal, Chris. Also because you head the guts
>(and the knowledge!) to show your head. I think this is one of the reasons
>you are an outsider among the weaker chess players who actually still have
>more successful machines/babies.
Chrisw is permanent outsider just about everywhere :)
>
>Let me give a final "psycho" statement.
>
>I'm dissapointed about the overall non-smartness of the cc experts. Take
>yourself and your image. Only because you (as a comparatively strong
>chessplayer) are convinced that computerchess could need a lot more chess
>knowledge to be able to *really* cope with GM chess, not with the shown GM
>output in funny skittles for some thousand bucks ..., you are an outsider.
>On the other side you could easily take Crafty source code, like Bruce (?),
>add this or that and then following the known paradigm. *Then* you were a
>nice guy. How primitive such a logic. You chose the hard way, the obviously
>more difficult one, and for sure, actually the less successful. Instaed of
>admiring your guts and firmness they laugh at you ...
Can be useful. I use it as a test for who still can use his brain. The one
that Thorsten passes every time. Probably on intuitive grouds rather than
logical ones, but he passes every time. It also detects the real stupid
ones. And also those that keep quiet.
On success .... I think my ideas work. I think their's are doomed. I think
my ideas, put into code by a far better programmer than me, would work very
effectively indeed. But, hey, the real good programmer, won't get off of his
cloud. Not that I asked him yet.
>
>The same with me as a newbie. If I would show groupie-like behavior, I
>would be accepted as another friendly brown-noser. People simply can't
>estimate the value of an opponent who (also as a newbie) could give some
>interesting ideas.
Rolf, without a doubt, has more creative brain power and insight than the
rest of these (two) cc groups put together. And the fools censored him ...
>But he's only able to create them in constant
>criticising of the experts (which seems to be a contradiction in itself
>because he's judged as a newbie). "Interdisciplinary" approaches seem to be
>widely unknown ... Althiough the experts make use of different experts in
>their teams. But then they don't regard their helpers as rela newbies of
>course. :)
>
>
>I dream of a brainstorming group where also those could be accepted who can
>only make their contributions in a creative opposition to the experts.
>That's BTW the usual history in all sciences. Often the direct expert-like
>secondary experts are too close to their chiefs to be able to create
>totally *new* ideas. Or let's better say "unusual", normally "not-accepted"
>ones.
>
>
>A really superior expert would be someone who were able to tolerate also
>the most aggressive critics. Not because he's a weak personality. But
>because he's so strong!
>
>
>When I saw the position in question I had suddenly all such ideas in mind.
>It's a very important position.
>
I know. Your original post was a work of genius.
>
>>And worse, the programs only deliver tactic lines of play, as if the only
>>importance is the tactic lines.
>
>>Do you see what I mean now when I say they play/think tactical, they say
>>'positional is long-range tactics'; that they haven't moved up a level to
>>'play/think positional' ? For them chess is all 'best lines' (tactical).
And
>>not positions (positional).
>
>>They were bad at seeing, so they made/bought programs. And the programs
led
>>them down the best line and blinded them further.
>
>
>Your operator seems to prove the contrary. Way beyond being a GM himself he
>also seems to understand that some "strong" programs often play really
>stupid chess ...
He has good intuition, and lets his emotions play.
>So I don't buy that people like Bob or Bruce have to be
>forced to behave the way you described. Also Bob did state very frequently
>that Fritz e.g. still is not all a GM player ... Not even close.
>
>
>>The solution of the position
>>>will be a beautiful explanation how humans play chess. Also we could find
>>>some novelties for our main topic here in the computerchess group.
Players
>>>who normally visit the analysis group might find different data.
>>>
>>>My theory (a weak theory still) says that if you always analyse positions
>>>with computer aid you lose your human advantage to be able to judge.
>
>>Yes. Also by laziness and lack of practice at thinking it for yourself,
too.
>
>
>>>Because it's a psychological knowledge that all kind of good or casual
data
>>>may lead you on wrong paths. Often it's very difficult to leave all these
>>>prepositions behind. I remember that Bob but also Ingo A. have said that
>>>there overall tactical strength was worsened by their working with
>>>computerized chess.
>>>
>
>>Absolutely. We all said it.
>
>I read it only from the quoted two.
Ed has said it too. Me also (well I think it anyway).
>
>>>
>>>
>>>BTW this as a little feedback. Both players were human players, so no
>>>computer involved.
>>>
>>>
>>>I still wait for a second answer from mclane about something strange
things
>>>happening, no? I mean the position isn't that easy?
>>>
>>>And Bob Hyatt isn't interested at all? Would be a pity. Because I have
good
>>>news from Crafty. Probably the position has to suffer because it was
>>>presented by the devil? GENS UNA SUMUS.
>
>>One reason he stayed out. But not the only one.
>
>>Hyatt would not risk his reputation on hazarding chess positional
viewpoints
>>on this position. He lacks the confidence in his chess knowledge
judgements.
>
>Usually he's not so careful. The most recent example was the Rebel v. Anand
>game with the BB unequal material.
>
>>Normally Hyatt waits until others have pronounced first in these cases. Or
>>quotes his own program's evaluation/best line combination.
>
>Yes, I knew this before. And to make a more difficult contest I added my
>further questions who then had shown the limited views. BTW your operator
>followed exactly this line. He gave a short varialtion and didn't see that
>it was a draw. So he acted in reality like a "bean-counter" at least in the
>sense you tried to explain it.
Arguably. But don't forget he was also replying to the devil :) This clouds
it a little.
>
>>And, the lack of response from any bean-counter is indicative of the level
>>of chess knowledge within their program evaluation function.
>
>Here I see a compination of a fear to get involved into another Rolf trick.
>But I will show that it's not a trick at all.
>
>>As Vincent has so eloquently pointed out, for the bean-counter paradigm
>>evaluation function without search, the opening move e4, loses to the
>>brilliant reposte h5 !!
>
>>Doh !
>
>>And this, despite the doubling of knowledge and halving of nps of Crafty
>>over the last year :))))
>
>I'm more thinking about the chess-freak GMRoman. As we have still 99945
>chess patterns up in our sleeves we have a life-long assurence to be
>related to the money "involved" constant flow of "communication" ... ;-)
>
De Groot's quantisation is off-topic, sorree.
Oh yes, and Rolf is censored.
Chris Whittington
>
>
>I still ask myself who's KingCrusher on ICC?
>
Do you mean egocrusher ?
mclane
>Take for example mclane's CSTal (Qe6) Kg8. This is a draw! Period.
Where do you see Kg8 ??? I said, i quote:
after 284 seconds the score drops to 5.96
1...Rb2 2.fxg6 hxg6 3.Rxg6+ fxg6 4.Qe6+ Kg7 5.Rf7+ Kh6 6.Qh3 Qh5 7.Qe3
why is this a draw ? you play Nf4 Qxf4 Qg5 .
BTW: I don't like you offending comments about bob in the text below,
therefore I have snipped it. I thought you would better let these
insults. Or are you unable to talk about this position with insulting
bob and others ?
best wishes
mclane
Carl Tillotson
in <6rp1f9$prf$1...@news00.btx.dtag.de>, you mentioned:
<a lot of computer related chess stuff snipped>
Please keep the Computer Chess in the Computer Chess newsgroup where it
belongs. Starting off a thread asking for analysis, with the sole
intention of it getting out of hand (as is the case with your normal
threads) will being to upset a few people.
If people want to see how this and that computer is fairing, then they
will go to rgcc.
--
Adios Amigo
Carl Tillotson
Lancashire Chess Association
homepage: http://www.lancashirechess.demon.co.uk/
Virtual Access 4.02 build 244 (32-bit)
Using Win95
Noam D. Elkies
but isn't White prosaically lost even if Black does not go in
for the 1...Rb2 complications? Black is up two pieces for a rook,
dominates the key d3-square, and has an attack. White' counterattack
with the f-pawn looks like an idle threat unless Black is "helpful".
It would seem that 1...Re8 (taking over the e-file, and in the
process covering the potential hole at e6) should win in due course
whether or not White interpolates 2 f:g6 h:g6; e.g. 2 Qd2 Qe5 intending
3...Qe2 (the 3 f6 and 4 Qh6 idea is too slow); if White trades Queens
Black goes after the White a-, c-, and/or g-pawns.
--Noam D. Elkies (USCF 2263)
[Change UK to USA abbr in e-address to reply]
Rolf Tueschen
please keep your defaming statement "with the sole intention and blabla"
for yourself. And keep it out of public newsgroups. If people want to see
this they might enter into email contact with you.
Do you think that your membership in a chess association gives you the
right to publish plain lies? I stongly doubt such a connection.
The position I posted was relevant for computerchess, that's right, but
also it's a very interesting position for human chess. Perhaps you wait
until I give the answers to my questions and the names of the players. Then
also you might be satisfied with my approach.
BTW what were your ideas about that position? Or don't you have a computer?
GENS UNA SUMUS.
Yours in chess
Rolf Tueschen
Rolf Tueschen
>--Noam D. Elkies (USCF 2263)
> [Change UK to USA abbr in e-address to reply]
Noam, could you please enlighten me what I should do to write an email to
you? I see no UK to change into USA. But then I'm very unfamiliar to the
English. Please inform me by email if you want.
Rolf Tueschen
Rolf Tueschen
I'll give the answers to my questions still a little bit later. It's simply
too fascinating. Without telling too much I want to make some comments.
Perhaps we could find even deeper insight.
f...@accountant.com wrote:
>Chris Whittington wrote:
>> Rolf Tueschen wrote in message <6rp1f9$prf$1...@news00.btx.dtag.de>...
>> >"Chris Whittington" <chr...@cpsoft.demon.co.uk> wrote:
>> >>Rolf Tueschen wrote in message <6rmr2n$9g7$1...@news01.btx.dtag.de>...
>> >>>As for fca he gave an almost master-like in depth analysis
>Nope... I've looked again (I've just returned from a day away) this time
>with an electronic friend, and there are other lines (only 15 mins
>spent). But I think I got the main ones. I missed ...Rb1 sacs for
>example, but it seems to end in a draw (per computer - almost impossible
>for a human to see it)
>> >>> but he forgot to answer all my questions.
>I answered the ones I could... Re who or how strong the players were,
>well if the position was one that could be subject to retro-analysis
>(i.e. what were the last moves) then I could speculate. From the
>position alone, I could not, only to say that maybe it was luck for
>Black that ...Rb2 wins. But a draw for Black was always there, even if
>not a win.
Yes, I see what you mean. That was perhaps a little unfair question.
Without the earlier moves you are by force forced to guess. But that
exactly was my idea. I could have given the moves. But your little thought
about the importance of earlier moves is not such a trivial thing. Believe
me. But you saw it. Not bad. No irony BTW involved.
You will be surprised if you read my solution. Oh, I love it to read your
ideas. And also you can be sure that it's not a hidden trick from "Rolf"...
Couldn't you deepen your ideas a little bit? You know the position is
therefore so interesting because it's not yet known in the databases or
books. But it's still a question of time. So I think the secret couldn't be
daved longer than a couple of days or so.
>> >>> But still if he found his lines without computers
>> >>>he must be a very strong player.
>Not necessarily: maybe only a beginner. In chess terms, looking at the
>magnitude of the problem and the extent of our ignorance, we could all
>be regarded as beginners, maybe.
That's true beyond any reasonable doubts.
>But perhaps you overlook something. The position itself was given as a
>problem, so I approached it as a problem i.e. "There must be
>something....".
I agree with you that it might have appeared as a problem. But I did not at
all present it that way. Alaso don't forget the main group. It's for
computerchess. I added the others to receive answers from different
people/experts. If you read the whole post in context I made it very clear
that this was from a game. Let's agree that the direct search for a quick
solution is sort of seduction you can't omit without telling too much about
it.
>As Chris did.
>Then followed the analysis that the WQ *had* to be ejected from where it
>was carrying on its multiple (but not overload) duties... Rb2 is thus
>easy to find as something to try. Whether it works or not - aha - that
>took me some time.
Perhaps you try to remember if you thought about it as a real game? Where
you had to defend and accept the possibly bad consequences of your
solution? Because that was the situuation for the players in real life.
They couldn't move the pieces for guesses and they also had no computer.
Excuse me for behaving like an idiot. Because it could be seen as if I
tried to insinuate that you might have forgotten about that difference. But
my last post was exactly about this tiopic. The dangerous effects of
immediate computer help for your own chess thought process. Therefore in
special I liked your answer very much. Because you exactly followed the
ranking order of my questions. Normally all cc experts, me (!) included,
simply enter the position into their favorite program to get a first
impression. :)
Myself I've become so lazy that I look at first at the easiest material
(you know pgn mode all comme il faut) before I then decide what to look at
first.
>I remind you of my answer:
>------------------------------------------------
>Summary of Pattern
>(ATTACK POTENTIAL) Philidor (N+Q) attack on f2 (Rook on f1 is tied to
>1st rank - back rank mate theme with WK in corner), combined with R+B
>attack on g2 (usual exposed check theme).
>(DEFENCE ISSUES) g6 pinned and attacked, exposed check and promotion of
>f5 pawn possibilities: while Black B protects against Qe8+, and Black Q
>against Qf8 later, e6 may be a dangerous hole to black (once f7 pawn
>moves). Defense must be checked to ensure no perpetual escape (or
>worse).
>[NB: This was the immediate response that Rolf wanted.... ]
>So... Detailed Plan resulting
>White Q protects against both Philidor and R+B so must be chased away
>pronto. Also,without the wQ cover of f1/g1 the WR becomes
>1st-rank-only. One move does that... ;-)
>Two W rooks attacking my king **but** Qh5 fails to a Philidor's Legacy
>attack ...Nf2 Kg1 Nh3+ Kh1 Qg1+ Rxg1 Nf2++ as the white Rook cannot
>leave 1st rank)
>So I play 1....Rb2 and the Queen has nowehere to go. There is no square
>on which it is safe and also protects key square f2: 2.Qe3 fails to
>X-Ray-type Philidor (2...Nf2+ 3.Kg1 Nh3+ 4.gxh3 Qxe3+ 5.Rxe3 Rg2+ (and
>the R+B attack on g2 prevails easily). 2.Qd1 is better but falls to
>2...Qf2! to which I cannot see any reply (3. Rg1 to squash the mating
>threats on g2 leaves White vulnerable to ...Nf2++ after a suitable
>Q-move!).
Exactly. Now let me enter a further complication. Think of being in the
middle of a tournament. And all your patterns (I had entered that notion
for special reasons) won't tell you how to win a game. What I want to say
is this. All that is wonderfully answering my questions. But please now you
are strong enough to go even deeper. *That* is real chess, sorry,
tournament chess, wenn you now could go a little bit deeper. Excuse me for
talking such a sybillinic story. It's because I can't tell you what you
should do but also I want to guide you to the point. I already know that I
possibly will have to excuse myself even more often simply because I'm not
copying material from a clever GM book but found the data and made my own
thoughts about it. Again I also admit that if you had the beginning moves
that you also were able to make a lot of different experiments. But think
of the possibility that you could give me variations/hypotheses about
connections between a certain ante (like Chris' guess that Black might have
accepted a sacc somewhere before...)
Perhaps you also try to give your opinion on a guess of Chris and also me.
We assumed that it were rather improbable that you as a nornmal chessplayer
should have found the most tactical line completely without computer help.
I must admit that I also have little doubts. But we all don't know each
other. Especially if we have to deel with an anonymous. But NB this is not
meant as an insult. I only tried to be openminded and honest. For a
computerchess freak on the side of Bob Hyatt you simply are a huge
exception, believe me. Perhaps you could also give a little thought on that
idea. What were your first ideas? Looking for yourself as *Rolf* had
wanted? Would really surprise me "un tout petit peu". :)
On the other side, *if* you did it you *are* a very strong OTB player by
definition, this time, believe *me*. :)
>Of course we need to check that White's attack does not come first!
>1...Rb2 2.fg must be followed by the prosaic 2...hg (well 2...Nf2+ 3
>Rxf2 Qxf2! (not 3...Rxe2??? where 4. gf+ mates) may win too, if you like
>living dangerously - I leave that to Mr Computer to find) 3. Rxg6+ fg 4.
>Qxe6+ Kh7 5. Rf7+ Kh6 6.Qh3+ Qh5 7.Qe3+ Nf4! 8.Qxf4+ Qg5 and checks are
>now controlled (9. Rh7+ Kxh7 10. Qf7+ Kh6 11.Qf8+ Kh5 12.Qh8+ Qh6
>13.Qe5+ g5) If instead 2...fg??, 3.Qe6+ looks like a draw, and if
>2...Rxe2??? 3. gf+ queens and wins easily for White.
>------------------------------------------------
>The Summary of Pattern was a matter of a few seconds to think (to write
>it took a 10 times longer). Note how Chris refers to smothering and I
>refer to Philidor. ;-)
>For such positions, only f2 is the key: defenders of it (K excepted)
>must be removed. The WR is not a bona fide defender as it is tied to
>back rank once Q is gone.
>> >>>fca asked that possibly a computer could find a hole in his lines ...
>Of course... I have been beaten tactically by comps so often now, even
>playing Levy-style, I am naturally paranoid.
I'd guessed that. :)
>And this is a position
>which has tactica aplenty. There may well be a better line for Black
>(but a win is good enough I think ;-) ).
Don't let you be distracted. Correct wording?
Please tell us what *you* would play if it were *your* position. BTW it
should be clear that Black had to see all of your lines in advance because
now you can't repair if suddenly White had a sudden winner. Here I wanted
further thoughts.
>> >>>That's exactly the problem. And that was the point of Kasparov's paranoid
>> >>>failure against DB.
>Agreed. We are psyched.
And it's not totally for this group, but I'd like to say that I think that
chess between equally strong players is mainly a question of who believes
what of the other's actions. It's all a game about psyching out your
opponent. If you then play someone two levels higher you're seriously
surprised that you can't find something to psych on because always
something extremely annoying is happening ...
Therefore I tried to insist. Your description of patterns here was
sufficiently smart and true. But I doubt that this is enough. Hmmmm.
>> >Take for example fca's analysis of the Rb2 ... Rxg6 fg Qe6 Kh7! Qe3 Qh5 Qe3
>> >Nf4!! line. True, Nf4 in the end is the key for the answer if Rb2 is
>> >winning. But do you believe that fca would see Nf4 immediately with Rb2?
>Of course I did not see it "immediately" or anything close to it. I am
>not a fast player at all, CorrCh rather. Perhaps unlike the computer, I
>did not need to see it in the short term. 1 ...Rb2 fulfilled the overall
>requirements
Yes, but here your life performance has a big hole. If you hadn't Kh7 it
would all be in vain. See what I mean?
>(eject Q) and after 2. fg (which was very easy to see)
>of course 2....fg (conceding e6) would be silly therefore 2...hg. Now
>3. Rxg6+ is the desparate attack,
After you will have seen the moves, you won't tell this anymore. Obviously
that motive was there for a longer time. Perhaps some 6 moves. So it's
nothing even close to a desperate attack. White played for it with full
conscieousness. Well, I couldn't ask him.
>also an obvious for a human to try
>out,
Also this is obsolete after Kasparov forfeit in a drawn position, no?
>and 3..fxg6 is forced. 4.Qe6+ was the point (see how I identified
>this as a hole) of the rook sac. Now I do not know why Rolf gave
>4...Kh7 an "!" - it is obvious as the only way to escape from white...
ROTFL. For this position I probably posted the whole stuff. Here is the
middle of attention for all computerchess experts.
>I can't believe this move gives problems to comps.
It does. I posted two hours ago some data. Sure it depends on hardware, but
the point seems to be clear. Now look. *If* computers calculate on
alternatives of the only winning move why could they decide in seconds for
Rb2? That is the question from a real cc newbie! And it's a pity that all
the usual experts have been digging *that* hole the past two days ... ...
I'd expected a few explanations.
>Then follows three
>quite forced half-moves (W to give perpetual, B to escape it). The line
>is easy to generate - very easy - as White faces sudden death if he
>loses control.
>So I repeat. Once the rook sac is envisaged at move 3, all moves for
>both sides up to W's 7th are obvious.
>So now to B's 7th move. It is clearly the only way, as it wrong-foots
>WQ. Such moves are often thema for endgame problems, btw. But I did not
>see it until the position appeared.
fca, did I understand you right, that you moved the pieces on your board
along your assumed line??
>I did not need to see it earlier
>(Remember the GM reply to "how many moves ahead do you think?" Ans: "One
>- the right one ;-) " ).
You son son of a ------------------- smart farther.
You mistreat an old saying. ;-)
>
>> No, I didn't believe it when I saw it. I thought (and I might be wrong and
>> misjudging), I thought he's played the line out on a comp and then wrote the
>> analysis.
>Nope.
>>Nf4 is very difficult, almost out of the question, to see "at a glance".
>I agree, and as explained above, it was not seen at a glance. The
>"Summary Of Pattern" was seen at a glance (and is pretty much the same
>as Chris's, except a little more detailed).
>The "Detailed Plan resulting" was a check to see whether the
>pattern-solving move ...Rb2 worked. That took me about 15-20 mins on a
>nice wooden chess board, where I usually plan my regicides.
While moving the peices around??
NB that it took you some 20 minutes. Think of the scandal if your pattern
based thinking resulted in a desaster? White could draw or even win?
>> >But that's not the only question. I posted the position primarily into a
>> >computerchess group.
>> >
>> >Now, I still wait for those who inform us about strange things to happen
>> >with computers in certain important positions. *Sure* all the programs *I*
>> >could test showed immediately Rb2 with a overwhelming score --- but it was
>> >my impression *for wrong reasons*!
>> >
>> >Take for example mclane's CSTal (Qe6) Kg8. This is a draw! Period.
>>
>> must play Kh7, the comps see this after a while ....
>>
>> >Fritz5 in CB7 shows up to ply 17/17 (Qe6) Kh8 (which is also a draw!).
>> >Then after some 15 minutes I stopped it. The same with Junior4.6.
>> >Up to ply 19 (we know that it's some strange output) Junior prefers (Qe6)
>> >Kh8 ...
>> >
>> >ONLY Crafty shows Kh7 after 2'12! -- -- Let's see if Bob will show his
>> >head right now. :))
>I do not see what the fuss is about. Kh7 is a very easy move for a
>human.
I want to oppose strongly. Having the position on the board, yes. But Black
had to see it far earlier! And Chris is quite right. Moves like that in
combination with Nf4 even more make a GM out of a blind amateur.
What's the fuss?? Look, we're here in a computerchess group.
Let me digress another time. I had those pains in my stomach when I read
things like "well I don't care about xy because it happens once in a
thousand games. I leave it like this in my program, because any changes in
the direction of xy would cost me some disadvantages overall...(Bob Hyatt
original quoted by heart)"
This is all HOUDINI. This is sufficient for blind amateurs like myself but
GMs can't be fooled like this. That's typical profit-oriented
science/techno. Atomic power stations have to be much safer. And it doesn't
take rocket science to understand why. Excuse me, but I'm already talking
with Bob's words. It's like an infection.
>> >Now -- let's think again about a human chessplayer who has to play OTB.
>> >Sure, if you once saw that Nf4 on your board/display you are favoring Rb2.
>> >But who's able to see Nf4 OTB without moving pieces to look if the line is
>> >ok for Black? That was my question for the Elos.
>2500+ to see it from the start at 40/2, as a guess; without Nf4 there is
>no Black win possible. 1800+ to see it OTB with the position set up.
We will come back on the two interesting numbers. :)
The logic for your second statement is very weak. You simply can't let your
opponent sacc on g6 and you have still not seen Nf4. This doesn't make
sense. On the other side your statement is true. It's BTW the bias of each
test series. If you know as a human that you will have to solve say 20
important positions then you will find many solutions just with the
knowledge (you described at the beginning of your post). But it's not a
good measuremenmt of anything close to your playing chess strength. A
complex topic.
>> Fantastic high to see it, right.
>
>> >And also here you, who
>> >answered the questions as the only one, were quite right. White can't be a
>> >GM because he didn't see early enough that his attack on g6 wouldn't win
>> >the game.
>No. This occurred to me but I refrained from writing it, as there is
>always the possibility that it was the best available. You cannot rule
>out White being a GM from the position alone.
Pardon me because I wrote with the higher wisdom in my mind. :)
>> >But what with Black? -- Therefore I asked you to look a little
>> >bit deeper. perhaps with the commentary I gave here you might try a final
>> >analysis before I give you all the answers?
>>
>> I've seen too much position and looked at it with CSTal, playing out the
>> lines. If black, say picked up an offered sac (since he was material up)
>> having seen all this stuff, knowing that has was safe from more saccing on
>> g6 after Rb2 , then black is seriously strong.
>>
>> Unless you have a joker up your sleeve and white has something, somehow,
>> somewhere. Always possible. Anand did against Rebel.
>That is not very possible, I now (not yesterday) think. Black wins. 99%
>sure.
Because the position is won or why? Because your line wins?
Ok, that was my comment. I promiss that I won't continue to keep suspense
for another week. All wait for the answers. Or give me a sign how to
proceed.
Rolf Tueschen
possible. I'd wished some more answers from computerchess freaks.
f...@accountant.com wrote:
[..]
>> Ok, that was my comment. I promiss that I won't continue to keep suspense
>> for another week. All wait for the answers. Or give me a sign how to
>> proceed.
>Keep proceeding on-topic, and you have my support. :-)
Sounds good and also a bit strange. BTW if you haven't something to hide or
fear you could email me your real adress and name. I assure you that I will
keep it secret here on rgcc.
Also your chess realted stuff is good. Support? I like chess related posts
but I think I don't need any "support". From your side it would be a little
bit suspicious. As Chris has said, there were too many things that live
their own life. They dictate parts of the future.
Another correction. I wrote life performance and it was a typo because it
should read live performance.
Robert Hyatt
: Computerchess Misc (1)
: A couple of days ago I saw the following position for the first time.
: W Kh1 Qe2 Rf1 Rg3 Pa3 c3 f5 g2 h2
: B Kg8 Qc5 Rb8 Bc6 Nd3 Pa7 c4 f7 g6 h7
: Questions:
: 0. Warning! Everybody who's not capable to solve this position either for
: White or Black without the help of computers will never more become a
: Grandmaster!
: 1. What is this a position for
: a better White?
: b better Black?
: c won White?
: d won Black?
From my "human" perspective, black looks better.
: 2. What pattern would a *human* see at first?
: a for White?
: b for Black?
For "this" human, as opposed to "a human" I see two. Black's attack on
f2 and g2, looks pretty strong. Black has a potential "smothered mate"
(ie Rb2 Q moves, Nf2+ Rxf2 (Kg1 Nh3+ Kh1 Nf2#) Qf2 and white has plenty
to be concerned about with the queen and rook in his face, plus the
bishop bearing on g2 as well...
For white, I don't see one pattern, I
see a couple... ie white's rooks are also bearing on the black king, even
though it is btm... playing f6 leaves back-rank problems for black. Playing
fxg6 opens files on the black king... and that last pattern figures in for
analyzing for black moves as well, because opening the kingside increases the
chances for a draw with queens on the board...
: 3. What does your favorite computer program see? Evaluation? Now play the
: favorite move. Evaluation better or worse?
I saw two moves that I would consider at quick analysis... Rb2 and Re8,
both to get the rook into the game. I would play Rb2 myself without a lot
of thought, because it forces a move by white to save the queen, or else white
has to find another sharp tactical alternative that I don't see.
Giving this to Crafty, it instantly finds Rb2, and says +5. It didn't change
after many minutes. The only thing interesting is that after Rb2 it says
fxg6, and it then finds the same idea I found... ie rxe2 for black leads to
grave difficulties and might well lose (Rxe2 gxf7+ Kf8 Rg8+ etc...) Which
is why I suppose Crafty expects hxg6 rather than Rxe2.
only let it search to depth=13 on my notebook... evaluation had not dropped
nor increased...
: 4. Which move does your program prefer for Black?
Rb2, as I said...
: 5. What do you think who won the game in the end? Note it was not a zeitnot
: duell. The outcome is already to be seen for any Grandmaster. Caution:
: Please don't ask Roman or Father Bill on ICC --- this would be a clear
: cheating!!!
I'd expect black should win... but anything could have happened. I'm not
absolutely certain this is winning (Rb2) anyway.. the possible perpetual by
white is very complicated..
: 6. What is your guess? Who were the players? Please all GMs, IMs and
: database freaks and businessmen, keep the secret to you. Thanks.
I'd guess somebody vs fischer. Since you said "father bill" I assume you
mean Bill Lombardy, which means this might have been spassky vs fischer,
since Lombardy was fischer's "second" at that match.. :)
: 7. For those who don't know exactly, what were the players like:
: a Elo under 1400?
: b Elo 1400-1600
: c Elo 1600-1800
: d Elo 1800-2000
: e Elo 2000-2200
: f Elo 2200-2400 USCF Master
: g Elo 2400-2500 IM
: h Elo >2500 GM
: i Elo >2700 Super GM
no way to answer that... without the moves at least..
: Many greetings to all
: Rolf Tueschen
--
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
Rolf Tueschen
>Computerchess Misc (1)
>A couple of days ago I saw the following position for the first time.
It's from a game between Willie Watson v. Michael Adams.
Played at the BCF-ch in 1986. Watson was IM with 2495.
With his 26th move Adams had sacced a rook for a bishop on e3. But not
enough. He could immediately pin the retaking Re3 with his Qc5. This is
already a won position because White has a single move to defend Re3. That
was Nd2-c4. But this Nc4 was immediately threatened by Pb6-b5. And the N
was lost because the white queen had to protect the Re3 and the white K had
to go out of the pin (Kg1-h1).
>W Kh1 Qe2 Rf1 Rg3 Pa3 c3 f5 g2 h2
>B Kg8 Qc5 Rb8 Bc6 Nd3 Pa7 c4 f7 g6 h7
Now Adams played 32.-Re8.
It was Noam Elkies who mentioned this move for the first time. And for me
it's a very interesting problem in the middle of a computerchess group.
When I put the moves of that game into my database I had Fritz in the
background as usual in CB7. He went for Rb2, the move all of you did
mention. And the move looked really strong. But it was one of the reasons I
posted the position. I wanted to see if it were possible to talk about
chess positions in a computerchess group. I wanted to know how many members
would think on their own about the position -- at least for a short moment.
Therefore I wrote the first questions. It was a good thing to see that
we're not at all limited by our constant use of computers. Noam of course
was the most independant author who had the guts to prefer a move normally
no computer proposed at first. Noam found exactly Michael Adams' move. I'm
not a GM but I brought that position into the group because I saw a
principal difference between human chess and computerchess.
As a human player OTB in a tournament like the BCF-ch you simply shouldn't
play 'va banque' in won position into a line where your opponent has a
long-ranged aggressive attack and where you only could defend your victory
by a surprising move 6 moves later ... Note that Michael Adams at that time
was a genial child prodigee but not an IM yet. He was 14 years old and had
ELO 2360. So normally with not so great chances against IMs. I analysed a
lot of these early games of the later super GM and it was quite evident for
me that this sort of players must have that special strength to understand
that a position is strong and that the well could choose the safer road for
the win. The computer however has no problems with tiredness or emotions
and if he sees a forced tactical line he will go for it.
Now the next point of the reasons why I posted exactly that position. When
I looked a little bit deeper into the Rb2 line I discobvered something
strange with the Fritz modul. It played always Kh8 instead of the winning
Kh7. I could understand it. Therefore I started a little study with all the
other programs. I already gave you the results. Fritz was still fixed on
Kh8 although he was deep in 17 ply and some 20 minutes. The same with
Junior who also had Kh8 as favorite drawing move. Of course both Fritz and
Junior still showed a total winning score. Only Crafty presented the
winning move after 2'12. Again, all the programs played Rb2 but all had Kh8
or CSTal Kg7 on their agenda. I mean this is strange. Because if a human
player would play chess like this he would never become a super GM. That
were my thoughts before I posted the position. This Kh8 also reminded me on
the old days of helpless pawn ending playing computers. The old Mephisto of
mine always went away from the site of the interesting quarrel as if he
could hypnotize me of doing the same and forgetting about queening ... :)
I'm not joking. I see a great task for the programmers to make their
creatures as strong as possible also in so-called lost positions. At first
I met that with the endgame CDs. When I read that you still had a lot of
"better" or "optimal" move choice in a totally lost position. The key must
be to reduce the number of winning solutions for the winner. The optimal
position (often invisible for us humans) might be hidden three moves later
so that the right decisions must be made already now. But then at such a
moment where the chances for a bad move (in a winning position it's
sufficient to make a drawing move) are e.g. 50-50. In older CBMs you will
find a lot of sensational material where the best GMs blundered.
A little side note.
When I saw the game of the German TV between Kramnik and Adams after the
normally winning position of White, Adams started his counter attack with
the two Ns and then Qc6 and all that. That was a different game. Because
both players are super mega GMs. And because the TV team around Helmut
Pfleger had begged the two to give a performance of real suspense they
tried to do their best. Suddenly the Fritz score changed dramatically into
the favor of the counter-attacking Adams but who had a little bit less time
to think. This game was played with a lot of lines like Rb2 with the later
Nf4!!. And as usual most lines weren't visible at all but they had to be
calculated by the players.
What is the reason for Re8? This move destroys all White's hopes of the
perhaps drawing Qe6.
33 Qd2 Qe5
34 fxg6 hxg6
35 Qh6 Qf5!
5 years after the game the GM Adams wrote "A neat backrank trick
effectively finishes the game. If 36 Kg1 Qxf1 37 Kxf1 Re1 mate."
BTW that was it what Chris, fca and surely also Noam had in mind when they
talked about the pattern of a weak first rank of White.
36 Ra1 Ne1
37 Kg1 Re2
38 Rd1 Qf2
0-1
>Questions:
>0. Warning! Everybody who's not capable to solve this position either for
>White or Black without the help of computers will never more become a
>Grandmaster!
I wanted to put as many clues as possible. And had an opening motivator in
mind. Sup-optimal as I would say right now. Because we have so many members
who surely did bury their dreams to become GM a longer time ago. Me
included. :)
>1. What is this a position for
> a better White?
> b better Black?
> c won White?
> d won Black?
Chris and fca were right. It#s a totally won position. But not because the
computers showed that score above 4.
>2. What pattern would a *human* see at first?
> a for White?
> b for Black?
I liked most the fca analysis. Also the nice mentioning of a PHILIDOR
showed me that fca must be a very strong player. If he had all that in mind
within seconds, then he surely must be well above 2200 at least.
OTB is another thin of course. This morning finally Bob entered his
thoughts. And I saw that he surely does understand chess good enough to
analyse positions like that. Although his " I would play Rb2 myself without
a lot of thought, because it forces a move by white to save the queen, or
else white has to find another sharp tactical alternative that I don't
see." proves that he never was a real OTB monster. Because Bob
underestimated White's attacking chances. And what if something like Nf4
were not available? Then White could draw. So the problem is first of all
if Rb2 doesn't throw away your winning game! Probably Willie Watson exctly
had hoped for such a play by Mickey Adams. Because *he* was the master and
had the experience. So Bob, I would say, that's our problem, and I would
include Ingo right away. We were only strong (meaning without apparent
stupid moves, omitting the word 'blunder') with the help of computers. Then
we're not a super GM at all, but perhaps a good IM. But don't let 'em take
away that computer, because then we're as helpless as any average amateur.
That's why we love our game. Because in tennis I see no technique that
could make a Wimbledon quarter finalist virtually. But with our helpers we
sometimes hit the ceiling and are able to demonstrate shocking lines out of
the games of superstars. So we are at least able to feel the positions the
great ones were really in. While in tennis, if your partner is unable to
play an ace, you can't feel the desperation of a real player on the
receiving side.
>3. What does your favorite computer program see? Evaluation? Now play the
>favorite move. Evaluation better or worse?
My machines see an oveerwhelming advantage for Black. Then suddenly dropped
their evals. Perhaps someone could explain that?
This is perhaps an interesting point. Programs don't differentiate between
a >+4. that is without the need to find a special difficult move and the
same score with a forced line down to the end. Just a hint. Why not
implementate sort of finger note. If you then use the multi move display
you can understand the difference between certain moves.
(Copyright for fingernotes like "risky" or "dull overall" is claimed by
now. RT.)
BTW I must add something in defense of the Fritz module. If you allow two
candidate moves in CB7 then Fritz changed much faster to Kh7. What's that?
>4. Which move does your program prefer for Black?
Well, all seem to prefer Rb2.
>5. What do you think who won the game in the end? Note it was not a zeitnot
>duell. The outcome is already to be seen for any Grandmaster. Caution:
>Please don't ask Roman or Father Bill on ICC --- this would be a clear
>cheating!!!
Fischer v. Spassky would be a bad choice for posting such a position
because those games are well known. No, I feared someone who had sort of
bulletin based collection of all greater events. A BCF-ch isn't just a
little weekend open ...
>6. What is your guess? Who were the players? Please all GMs, IMs and
>database freaks and businessmen, keep the secret to you. Thanks.
For me it's always an exciting feeling to enjoy the moves of such a young
player who later made it to the top. I remember my own "strength" when I
was 14. And the difference is quite clear.
>7. For those who don't know exactly, what were the players like:
> a Elo under 1400?
> b Elo 1400-1600
> c Elo 1600-1800
> d Elo 1800-2000
> e Elo 2000-2200
> f Elo 2200-2400 USCF Master
> g Elo 2400-2500 IM
> h Elo >2500 GM
> i Elo >2700 Super GM
That question was answered remarkably exact by Chris. The point is the
balance of two strong plans or agendas. Also White had a top attack. And I
strongly doubt that a 1800 is able to build up such a position against such
a strong one of Black. All this is wiped off if you only use computers as
an aiding help. You simply lose the "FEELING" for the evaluation/judgement
of such tactical positions. The pattern of the sac on g7 with later Qe6 is
well known. But Watson must have played for it because he allowed the
quality sac in a position where he had already great problems due to the
strong black N on d3. Only Watson himself could explain it but he must have
thought that he might get good prospects with the f4-f5 counter attack.
With the Q and the two R. Let's assume he had seen the future until the
strong Nf4!! he then might have chosen a different line.
With fca I already discussed the possible similarity of the "hole" in the
else drawing line with Kasparov's missing of the Qe3 draw. Willie Watson
had no other choice in our position here. He had to go for Rg6 and then
hoping for a perpetual. And likewise even an average player like M. Ashley
saw on the stage that Kasparov's last chance could only be a perpetual. But
to look at a draw a little bit closer, Garry had to go for it at first. But
he didn't do that. Folks, Bobby Fischer did never do that. He played till
the end. If there was the slightest chance to win or save a game. Kasparov
behaved like a weak amateur who
Rolf Tueschen
TUESCHEN.MEDIZ...@t-online.de (Rolf Tueschen) wrote:
>With fca I already discussed the possible similarity of the "hole" in the
>else drawing line with Kasparov's missing of the Qe3 draw. Willie Watson
>had no other choice in our position here. He had to go for Rg6 and then
>hoping for a perpetual. And likewise even an average player like M. Ashley
>saw on the stage that Kasparov's last chance could only be a perpetual. But
>to look at a draw a little bit closer, Garry had to go for it at first. But
>he didn't do that. Folks, Bobby Fischer did never do that. He played till
>the end. If there was the slightest chance to win or save a game. Kasparov
>behaved like a weak amateur who
tried to demonstrate his chessic superiority to real ignorants just by
resigning some moves before the final mate. This is a first theory. But
then there must be an even stronger disturbing thought that might have
succeeded in troubling Kasparov's fighting spirit. That was formerly the
position where he was sure that some human influence might have happened.
This question is still not answered. I would even go so far and state that
the strange retirement of the whole DB/IBM team is somewhat suspicious.
This way they omitted to be forced to present all the evidence perhaps with
data from the same hardware. For Kasparov this is a very bad decision
because he now must live with that unanswered question. And a lot of bad
press in the follow-up. As I demonstrated Kasparov still 6 months later
spoke the same words. They did never present the promissed evidence. But
Hsu or was it Tan did promise it. AND Benjamin didn't present any scores or
games of his many games or matches. So his lie on stage of the event is now
proven. So we have *only* losers on both sides.
What could be made better the next time, in case I wanted to post part 2?
Thanks for all who sent their contributions.
Robert Hyatt
: I liked most the fca analysis. Also the nice mentioning of a PHILIDOR
: showed me that fca must be a very strong player. If he had all that in mind
: within seconds, then he surely must be well above 2200 at least.
: OTB is another thin of course. This morning finally Bob entered his
: thoughts. And I saw that he surely does understand chess good enough to
: analyse positions like that. Although his " I would play Rb2 myself without
: a lot of thought, because it forces a move by white to save the queen, or
: else white has to find another sharp tactical alternative that I don't
: see." proves that he never was a real OTB monster.
Note that the above is not exactly the interpretation I meant. When I
said "without a lot of thought" my meaning was "quick chess." IE in a
position where I didn't have a lot of time to calculate. Because it
seems to me to be the most forcing. Re8 was, as I mentioned, my second
choice, because it seems a trifle safer with the perpetual check problem
after black's kingside gets shredded.
However, I wonder what Kasparov would play? I'd bet Rb2... because it
fits his style, *unless* he analyzes this as a forced draw. But I'd bet
he would spend 30 minutes if needed to convince himself that there is an
escape from the checks, and if he was convinced, Rb2 would be his choice.
: Because Bob
: underestimated White's attacking chances.
I'm not sure I "underestimated white's attacking chances" at all. But when
factoring white's attack in with black's attack, I felt that black's attack
was/is stronger. If someone can prove that Rb2 draws, then I'd agree. But
my comment was that the "perpetual problem" around black's king seems to
be serious, but not unsolvable". But I didn't spend an hour tracking down
all the checks...
However, I'd toss in another point... black is at least drawing, and
probably winning, after Rb2. Which is no small point since black is usually
happy for a draw, and gleeful when winning. And it would be interesting to
see if Re8 can actually produce a win also, because black has problems of his
own with the two white rooks, plus the queen...
: And what if something like Nf4
: were not available? Then White could draw. So the problem is first of all
: if Rb2 doesn't throw away your winning game! Probably Willie Watson exctly
: had hoped for such a play by Mickey Adams. Because *he* was the master and
: had the experience. So Bob, I would say, that's our problem, and I would
: include Ingo right away. We were only strong (meaning without apparent
: stupid moves, omitting the word 'blunder') with the help of computers. Then
: we're not a super GM at all, but perhaps a good IM. But don't let 'em take
: away that computer, because then we're as helpless as any average amateur.
There I don't agree. I play lots of chess *without* a computer. I make my
share of tactical mistakes, and positional mistakes... But I can certainly play
without Crafty's help. No doubt I'd be far better with it, of course...
: That's why we love our game. Because in tennis I see no technique that
: could make a Wimbledon quarter finalist virtually. But with our helpers we
: sometimes hit the ceiling and are able to demonstrate shocking lines out of
: the games of superstars. So we are at least able to feel the positions the
: great ones were really in. While in tennis, if your partner is unable to
: play an ace, you can't feel the desperation of a real player on the
: receiving side.
:>3. What does your favorite computer program see? Evaluation? Now play the
:>favorite move. Evaluation better or worse?
: My machines see an oveerwhelming advantage for Black. Then suddenly dropped
: their evals. Perhaps someone could explain that?
I didn't see this myself... so I'm not sure what you are seeing. Crafty stuck
in at +5 or so solidly. Thorsten said CSTal hit +8, later dropped to +5. I'm
not sure how CSTal evaluates things so I don't know what this means..
: This is perhaps an interesting point. Programs don't differentiate between
: a >+4. that is without the need to find a special difficult move and the
: same score with a forced line down to the end. Just a hint. Why not
: implementate sort of finger note. If you then use the multi move display
: you can understand the difference between certain moves.
: (Copyright for fingernotes like "risky" or "dull overall" is claimed by
: now. RT.)
That's a computer "hallmark" and I don't see a problem with it. They are
quite exact calculators... I can show you games that Cray Blitz played
against humans where it walked a very fine tightrope with attack / defense,
and won, although one wrong move at any point would have meant instant
disaster...
Sure, it would be better to play a "safe +4 move" over a "risky +7" move.
But I don't quite know (yet) how to differentiate those "safe/risky" attributes
and factor them into the numeric score as required by alpha/beta...
: BTW I must add something in defense of the Fritz module. If you allow two
:>Many greetings to all
:>Rolf Tueschen
Robert Hyatt
: TUESCHEN.MEDIZ...@t-online.de (Rolf Tueschen) wrote:
:>With fca I already discussed the possible similarity of the "hole" in the
:>else drawing line with Kasparov's missing of the Qe3 draw. Willie Watson
:>had no other choice in our position here. He had to go for Rg6 and then
:>hoping for a perpetual. And likewise even an average player like M. Ashley
:>saw on the stage that Kasparov's last chance could only be a perpetual. But
:>to look at a draw a little bit closer, Garry had to go for it at first. But
:>he didn't do that. Folks, Bobby Fischer did never do that. He played till
:>the end. If there was the slightest chance to win or save a game. Kasparov
:>behaved like a weak amateur who
: tried to demonstrate his chessic superiority to real ignorants just by
: resigning some moves before the final mate. This is a first theory. But
: then there must be an even stronger disturbing thought that might have
: succeeded in troubling Kasparov's fighting spirit. That was formerly the
: position where he was sure that some human influence might have happened.
: This question is still not answered. I would even go so far and state that
: the strange retirement of the whole DB/IBM team is somewhat suspicious.
: This way they omitted to be forced to present all the evidence perhaps with
: data from the same hardware. For Kasparov this is a very bad decision
: because he now must live with that unanswered question. And a lot of bad
: press in the follow-up. As I demonstrated Kasparov still 6 months later
: spoke the same words. They did never present the promissed evidence. But
: Hsu or was it Tan did promise it. AND Benjamin didn't present any scores or
: games of his many games or matches. So his lie on stage of the event is now
: proven. So we have *only* losers on both sides.
I've pointed this out before, so I guess I should point it out again. The
"output" requested by Kasparov was given to him. In fact, for the moves
he questioned, he got exactly what the program produced. For the main move
(where it refused to take a pawn, although I don't remember where it was so
far back in time) the analysis was published in the newspaper, in fact. He
wanted output for a specific couple of moves... He got that... then he
wanted output for the complete games... they refused. I believe they were
entitled to refuse such a request, because I doubt he would have "dumped
core" and explained what he was doing in each of the games... so why should
DB have to reveal what *it* was doing?
Just a classic case of poor sportsmanship by a chess player I *used* to
admire...
: What could be made better the next time, in case I wanted to post part 2?
: Thanks for all who sent their contributions.
:>>Many greetings to all
:>>Rolf Tueschen
Carl Tillotson
In <6rq07r$a3n$1...@news00.btx.dtag.de>, you mentioned:
> please keep your defaming statement "with the sole intention and blabla"
> for yourself. And keep it out of public newsgroups. If people want to see
> this they might enter into email contact with you.
Sorry Rolf, but you are certainly a leopard that would appear to never
change it's spots.
You also mentioned:
>Thanks for the completion. I will surely post the "answers" as quick as
>possible. I'd wished some more answers from computerchess freaks.
There is a place for the chess_computer stuff, that's why there is a
different newsgroup. Why else have you cross-posted all this stuff, either
it's relevant or it is not.
Why else is it that the major contributors are all well known on rgcc but
hardly ever contribute to rgca ?
If this thread actually stayed on the 'analysis' side of things then fine,
but your last contribution started to contain irrelevant attacks against a
certain Mr Hyatt.
Of course, you may have changed your spots, but you are going to have to
spend some time convincing people. I remain one of those to be convinced.
I merely pointed out that there is a place for Computer Chess, and although
the thread contains an interesting position, I would like the computer
analysis to stay in rgcc - which is what the charter was set up for.
And I repeat you said:
>Thanks for the completion. I will surely post the "answers" as quick as
>possible. I'd wished some more answers from computerchess freaks.
And I again repeat my simple request, can you keep computer chess to the
relevant newsgroup, if the thread continues to stay in the analysis vein
then there isn't a problem. My issue is that this thread has been
crossposted, and thus invites the usual rgcc crowd to inject their
'computer analysis' and a thread that started off innocently would risk
getting out of control.
I trust you will actually take my points into consideration, instead of
jumping off at a tangent - which I must add was on-par for the Rolf I used
to know :-)
Rolf Tueschen
>In rec.games.chess.computer Rolf Tueschen <TUESCHEN.MEDIZ...@t-online.de> wrote:
>: Here's the missing rest of my post.
>: TUESCHEN.MEDIZ...@t-online.de (Rolf Tueschen) wrote:
>:>With fca I already discussed the possible similarity of the "hole" in the
>:>else drawing line with Kasparov's missing of the Qe3 draw. Willie Watson
>:>had no other choice in our position here. He had to go for Rg6 and then
>:>hoping for a perpetual. And likewise even an average player like M. Ashley
>:>saw on the stage that Kasparov's last chance could only be a perpetual. But
>:>to look at a draw a little bit closer, Garry had to go for it at first. But
>:>he didn't do that. Folks, Bobby Fischer did never do that. He played till
>:>the end. If there was the slightest chance to win or save a game. Kasparov
>:>behaved like a weak amateur who
>: tried to demonstrate his chessic superiority to real ignorants just by
Let's take a look in a relaxed way. No agitation.
1. For me it was important to quote Kasparov who made exactly that point in
hi Oxford lecture. He didn't get the promissed data. Now I'm not to deep
into that to decide. But until now this statement was not officially
refused by the DB/IBM team or the rest of it. *You* are not a member.
2. The Benjamin point. I already made the statement many times before and
never got an answer. he had answered someone on the stage where they
commented the games. And he said he were not allowed to even talk about the
scores. But the moment the matchj were finished he would certainly publish
all the data about his games. Now I ask you if it's not a little bit
strange if Tan or was it Hsu now state that Benjamin didn't play a single
"match" or ebven a game to the end because all he did were analysis and
taking moves back and so on ...
This is a clear lie in view of Benjamin's own promise who had talked about
results, scores, games, matches.
So far about good sportsmansship and plain lies.
But let's not get hysterical about an event offside our rather cool and
elevated group rgcc.
[..]
Rolf Tueschen
>In rec.games.chess.computer Rolf Tueschen <TUESCHEN.MEDIZ...@t-online.de> wrote:
>: TUESCHEN.MEDIZ...@t-online.de (Rolf Tueschen) wrote:
>: I liked most the fca analysis. Also the nice mentioning of a PHILIDOR
>: showed me that fca must be a very strong player. If he had all that in mind
>: within seconds, then he surely must be well above 2200 at least.
>: OTB is another thin of course. This morning finally Bob entered his
>: thoughts. And I saw that he surely does understand chess good enough to
>: analyse positions like that. Although his " I would play Rb2 myself without
>: a lot of thought, because it forces a move by white to save the queen, or
>: else white has to find another sharp tactical alternative that I don't
>: see." proves that he never was a real OTB monster.
>Note that the above is not exactly the interpretation I meant.
How could this ever happen. :)
No, I made a statement about something I don't have much data. I will tell
you why I assumed that. It's this. You wanted to play Rb2, but then you had
to know what happened after the destruction of your pawn shield. If you're
now saying that you would play this with short time schedule then you would
play all or nothing. Because without Nf4 it's a draw. That of course can't
be your intention because you have a won position.
>When I
>said "without a lot of thought" my meaning was "quick chess." IE in a
>position where I didn't have a lot of time to calculate.
Yes, that was my point. In that case *I* would recommend to try to find the
safer continuation, you see my point? All or nothing is the proof for my
not a OTB monster assumption. Or did you mean zeitnot? Well even then you
should know about your won position and omit exactly the only line where
White gets a huge dangerous attack.
>Because it
>seems to me to be the most forcing. Re8 was, as I mentioned, my second
>choice, because it seems a trifle safer with the perpetual check problem
>after black's kingside gets shredded.
Exactly. I understood you this way. But unfortunatelly you came rather late
with this move after Noam had already posted it. Otherwise *you* were the
"winner" of our little game. But I think in view of your late coming I
praised you still enough, no? No irony.
>However, I wonder what Kasparov would play? I'd bet Rb2... because it
>fits his style, *unless* he analyzes this as a forced draw.
Exactly. I think you made a good point. Let me differentiate. If it's the
most important game for the Wch or the win of a tournament, even Kasparov
would play for a safe point. Unfortunatelly our position is so clear for a
real super GM that he would also play for Rb2 with closed eyes. But we must
always realize that we're talking cheap from a theoretical outsider
position. We two, we were never in such a situation. But that's our good
game. Even patzers could find the gold in post mortem and with computer
help anyway. For all players below 2200 the whole variation isn't trivial.
Depends of your experience and training. That the N is near-by is to be
seen quite easily. And also that it could be given away because Black has
such a won position. But my assumption was that we here don't have too many
2200 players. What do you think? So I waited for the *human* approaches.
This time honestly without a computer. Which is very naive because the
moment you go into newsgroups *with* your computer, your chessprogram is
not that far away ...
>But I'd bet
>he would spend 30 minutes if needed to convince himself that there is an
>escape from the checks,
Bob, believe me this for a change. Kasparov would see this in 30 seconds.
I'm not joking. You seem to be completely unaware of what such players are
capable of. That's BTW the basis of the critics from professionals against
Kasparov's performance against Deep Blue. The stuff of the second game,
believe me also this, is kind of pea-nuts for a "normal" Kasparov. I mean
Qe3 and draw. Bob, that was the reason for me to doubt the overall
smartness of you computerchess experts. Also the even assuming for the
tenth of a second that Kasparov were able to make such a stupid move as h6
in the sixth game. I simply can't believe such a naivety. I'm not talking
of the normal public. I mean the educated folks of the elite of
computerchess in the USA. But perhaps we could come back on this point
later. Let's not do it here for too long.
I hope that you will afree that Re8 was for the young Adams a wonderful
move. In a won position of course. But didn't we all destroy many won
positions?
>and if he was convinced, Rb2 would be his choice.
That's not the deepest sentence, you will surely agree. :)
>: Because Bob
>: underestimated White's attacking chances.
>I'm not sure I "underestimated white's attacking chances" at all. But when
>factoring white's attack in with black's attack, I felt that black's attack
>was/is stronger. If someone can prove that Rb2 draws, then I'd agree.
No, you can't do that. Did you find another line after Rb2 where White
could draw? C'mon, not another torture.
>But
>my comment was that the "perpetual problem" around black's king seems to
>be serious, but not unsolvable". But I didn't spend an hour tracking down
>all the checks...
>However, I'd toss in another point... black is at least drawing, and
>probably winning, after Rb2. Which is no small point since black is usually
>happy for a draw,
What are you talking about. You mean normally Black is happy to omit the KK
56% law? That's right. But here Black has a won game ...
>and gleeful when winning. And it would be interesting to
>see if Re8 can actually produce a win also, because black has problems of his
>own with the two white rooks, plus the queen...
Bob, where were you living today? I already posted the solution of the
game. And Re8 did win. For sure. What's going on with you?
>: And what if something like Nf4
>: were not available? Then White could draw. So the problem is first of all
>: if Rb2 doesn't throw away your winning game! Probably Willie Watson exctly
>: had hoped for such a play by Mickey Adams. Because *he* was the master and
>: had the experience. So Bob, I would say, that's our problem, and I would
>: include Ingo right away. We were only strong (meaning without apparent
>: stupid moves, omitting the word 'blunder') with the help of computers. Then
>: we're not a super GM at all, but perhaps a good IM. But don't let 'em take
>: away that computer, because then we're as helpless as any average amateur.
>There I don't agree. I play lots of chess *without* a computer. I make my
>share of tactical mistakes, and positional mistakes... But I can certainly play
>without Crafty's help. No doubt I'd be far better with it, of course...
Bob, you snipped half of my very profound post. I told the group exactly
this. That you were able to give a good judgement. Do you want another
perhaps even better statement? :)
>: That's why we love our game. Because in tennis I see no technique that
>: could make a Wimbledon quarter finalist virtually. But with our helpers we
>: sometimes hit the ceiling and are able to demonstrate shocking lines out of
>: the games of superstars. So we are at least able to feel the positions the
>: great ones were really in. While in tennis, if your partner is unable to
>: play an ace, you can't feel the desperation of a real player on the
>: receiving side.
>:>3. What does your favorite computer program see? Evaluation? Now play the
>:>favorite move. Evaluation better or worse?
>: My machines see an oveerwhelming advantage for Black. Then suddenly dropped
>: their evals. Perhaps someone could explain that?
>I didn't see this myself... so I'm not sure what you are seeing. Crafty stuck
>in at +5 or so solidly. Thorsten said CSTal hit +8, later dropped to +5. I'm
>not sure how CSTal evaluates things so I don't know what this means..
>: This is perhaps an interesting point. Programs don't differentiate between
>: a >+4. that is without the need to find a special difficult move and the
>: same score with a forced line down to the end. Just a hint. Why not
>: implementate sort of finger note. If you then use the multi move display
>: you can understand the difference between certain moves.
>: (Copyright for fingernotes like "risky" or "dull overall" is claimed by
>: now. RT.)
>That's a computer "hallmark" and I don't see a problem with it. They are
>quite exact calculators...
It's going a bit slow today with you. It's a pity that you didn't talk
about my idea. To explain it with other words.
Take a position x where the program sees a winning >6. or take 9. whatever
you want. Now you will surely agree with me that even such a won position
could still have a little situation where it could be possible to blunder
or whatever. Right? Now the computer sees still 9. because he sees that he
can omit that blunder. Still. I think that a line that still has
opportunities to blunder or a move decision of a certain difficulty is
"better" for the player who can create such situations for his opponent
even if he's in a ,lost position. Was this clear enough?
BTW I'm convinced that at least Fritz does use such procedures when the
time budget shows relevant values. I often saw that Fritz played nasty
moves in my zeitnot. When I had expected a more normal, yes, even stronger
move against that I already had a good reply in mind. You understand me? Is
this enduser mystification or do you use such tricks in crafty too?
Anyway, I wanted to initialize that a program could add some symbols for my
"words" so that the user could see before it happened that this is a line
with stumbling blocks and another the easy straight way for a win. But the
"stumbling" way could have a still higher eval. You get my point?
>I can show you games that Cray Blitz played
>against humans where it walked a very fine tightrope with attack / defense,
>and won, although one wrong move at any point would have meant instant
>disaster...
>Sure, it would be better to play a "safe +4 move" over a "risky +7" move.
>But I don't quite know (yet) how to differentiate those "safe/risky" attributes
>and factor them into the numeric score as required by alpha/beta...
Ah, you already got my point. Well, how to do this? Quite simple. Ahem, you
know for me at least. You have all your analysis stored, right? Now what
you're doing is the following. You remember the abrupt fall of the eval I
mentioned above. If that happens because suddenly the analysis discovers
such a difficult situation, then e.g. find another key move and then goes
up again with the eval, then tht's a line that could need such a Rolf
symbol, a star, or a sun, whatever you want. ;-)
But perhaps I'm not talking with the right guy because all this could
interest the enduser but will not directly influence the strength of a
machine. Although I also see for you, because you want to tune Crafty
against humans, an important point. If you have an opponent who is able to
astonish Crafty with moves, Crafty didn't see or didn't think it could be
played, or wouldn't be that good, but then the move shows a good or better
eval, *then* this discovery should be processed somehow for the rest of the
game or the next games against the same player or as a signal for this
special opening or whatelse. Just one idea out of my bag with another
million ideas ... ...
[..]
Robert Hyatt
: Let's take a look in a relaxed way. No agitation.
: 1. For me it was important to quote Kasparov who made exactly that point in
: hi Oxford lecture. He didn't get the promissed data. Now I'm not to deep
: into that to decide. But until now this statement was not officially
: refused by the DB/IBM team or the rest of it. *You* are not a member.
Then quite simply, Kasparov is a liar. He asked to see output for a
couple of specific moves. It was given to him. It was given to others.
It was even published in the New York Times. And the output has even
been discussed *here* if you recall. He then decided that the couple of
moves were not enough. He wanted it "all". I'd have told him to jump
in a lake, had it been up to me. I guess Hsu and company effectively did
this, which was the correct action...
: 2. The Benjamin point. I already made the statement many times before and
: never got an answer. he had answered someone on the stage where they
: commented the games. And he said he were not allowed to even talk about the
: scores. But the moment the matchj were finished he would certainly publish
: all the data about his games. Now I ask you if it's not a little bit
: strange if Tan or was it Hsu now state that Benjamin didn't play a single
: "match" or ebven a game to the end because all he did were analysis and
: taking moves back and so on ...
Not strange to me. That is *exactly* how I debug crafty. Play, make
a mistake, take it back, because I am not trying to punish my mistakes,
I am trying to see if it "understands" what the position requires. But
in doing so, I learn a *lot* by simply remembering how many moves I make
that I think are good, but which Crafty promptly recognizes as lemons.
A lot can be learned from that...
: This is a clear lie in view of Benjamin's own promise who had talked about
: results, scores, games, matches.
Or perhaps, since he was *paid* by IBM, they said "shhhhh"... He
*was* obligated to do what they said...
Carl Tillotson
In <6rs9ku$8gl$1...@news00.btx.dtag.de>, you mentioned:
> But it was one of the reasons I
> posted the position. I wanted to see if it were possible to talk about
> chess positions in a computerchess group.
Rolf, you have admitted that you posted this in a computerchess group,
why may I ask did you cross-post it.
And you had the gall to actually criticise me for pulling you up on the
specific point in question.
Carl Tillotson
In <6rsjkc$flm$2...@juniper.cis.uab.edu>, you mentioned:
<lots of stuff snipped>
This is exactly the reason I didn't want Rolf starting a cross-post,I
have no object to you flooding rgcc but please don't bring your
arguments over into rgca.
mclane
>Aber aber, haste denn nicht verstanden?
>Der nutzt doch jetzt so ein Schachproblemchen nur als Traeger
>fuer seinen Bullshit. Das ist wie beim radio, nachdem das
>Megaphon nicht mehr reichte, hat man die Hochfrequenz genutzt,
>um Sprache zu uebetragen. He presto ?
>Und haelt sich dabei noch fuer schlau.....
Ja -es ist schade. Er ist einfach nicht in der Lage diese
Schlammschlacht zu beenden.
Pfui - kann ich da nur sagen...
>Tschoe
>Peter
Beste aber Grüße an dich peter...
Mit den Petern habe ich eigentlich im Net immer gute Erfahrungen
gemacht. :-)
best wishes
mclane
Gerry Forbes
TUESCHEN.MEDIZ...@t-online.de (Rolf Tueschen) wrote:
>Here's the missing rest of my post.
>
>TUESCHEN.MEDIZ...@t-online.de (Rolf Tueschen) wrote:
>
>
>>else drawing line with Kasparov's missing of the Qe3 draw. Willie Watson
>>had no other choice in our position here. He had to go for Rg6 and then
>>hoping for a perpetual. And likewise even an average player like M. Ashley
>>saw on the stage that Kasparov's last chance could only be a perpetual. But
>>to look at a draw a little bit closer, Garry had to go for it at first. But
>>he didn't do that. Folks, Bobby Fischer did never do that. He played till
>>the end. If there was the slightest chance to win or save a game. Kasparov
>>behaved like a weak amateur who
>
>resigning some moves before the final mate. This is a first theory. But
>then there must be an even stronger disturbing thought that might have
>succeeded in troubling Kasparov's fighting spirit. That was formerly the
>position where he was sure that some human influence might have happened.
>This question is still not answered. I would even go so far and state that
>the strange retirement of the whole DB/IBM team is somewhat suspicious.
>This way they omitted to be forced to present all the evidence perhaps with
>data from the same hardware. For Kasparov this is a very bad decision
>press in the follow-up. As I demonstrated Kasparov still 6 months later
>spoke the same words. They did never present the promissed evidence. But
>Hsu or was it Tan did promise it. AND Benjamin didn't present any scores or
>games of his many games or matches. So his lie on stage of the event is now
>proven. So we have *only* losers on both sides.
>
Close, Rolf, but there was *one* winner: not Deep Blue, but IBM.
They got exactly what they wanted i.e. $100 million in free
publicity according to the Wall Street Journal (actually it
probably cost them around $1million, but still a good return).
The more you examine it, the more you will be convinced that this
event was 100% business and 0% science.
For instance, put this in Alta Vista: "Jeff Kisseloff" AND "Deep
Blue". See what lengths IBM went to in order to prevent all
discussion (science) while pumping up the PR (business).
And look at Michael Khodarkovsky's (sp?) book, A New Era: How
Garry Kasparov Changed the Chess World (or whatever; fire the guy
who came up with a long, unmemorable, and non-descriptive title)
which covers Kasparov-Anand and both K-DB matches. From Khodarkovsky
one gets the whole view of the "Get Kasparov" attitude of IBM, who
made certain that they took every opportunity to engage in
psychological warfare from the ludicrous set up (Kasparov's rest
area was originally about 1.5 minutes away from the playing area)
to the censorship of Kisseloff to the battle of the printouts and
CJ Tan's lies lies lies lies lies lies lies lies lies lies lies lies
and more lies. If this guy tells you the sky is blue it's probably
snowing. Not only did Tan lie about the printouts, he also lied
about the number of GMs on the Deep Blue team; Khodarkovsky wonders
if the fact that Fedorowicz and Defirmian were on the team (and had
signed secrecy agreements) being leaked during game 3 was also part
of the psych job. Of course, there is no indication of what they
were doing in the hotel across the street during the games.
As for the printouts (Kasparov requested them for three moves:
35.Bxd6, 37.Be4 and 45.Ra6; DB took 15 minutes on 35.Bxd6 and 6
minutes on 36.axb5 but in game 1 and up to move 35 in game 2 it
took 3-4min/move) Tan originally refused to provide them. Kasparov
asked for them through the appeals board and Tan agreed to give
them to Ken Thompson. After game 3 Kasparov asked for the printouts
in front of the audience. Tan said that after the match K could
have printouts of all the games. Kasparov: "Then I have to believe
that it was the hand of God that helped the computer to play so
strongly in game 2." Later that evening Thompson again asked for
the printouts and Tan asked if he still needed them! On the
morning of the day of game 4 the printouts still hadn't been
delivered. At 12:30 pm Kasparov's manager delivered a statement of
fact to Monty Newborn: if Thompson did not receive the printouts,
he would be unable to appear at game 4 as a member of the appeals
board. At 2:30 Newborn called to say that the appeals board had
received the printouts that Kasparov had requested. When the Kasp-
arov team arrived Thompson informed them that only the printout for
37.Qb6 had been sent (not 37.Be4, but only a side-variation). This
is the printout that was in the NY Times (but IBM refused to allow
Kodarkhovsky to reproduce it in his book, even though Tan--who else?-
had told Kodarkhovsky prior to the match that he would be given some
printouts for his book). After game 4 Kasparov asked that printouts
of game 5 and 6 be sealed and locked in a safe. Tan said it was
impossible. After game 5 Kasparov remained at the board waiting for
the printout. He was told that Tan had said this was impossible but
when Tan was called he said, no problem. After being unable to find
Carol Jarecki, Kasparov went to his hotel leaving his team for the
printouts. Jarecki showed up 10 minutes later carrying an envelope
with her signature on it. The envelope wasn't very thick and
Khodarkhovsky asked if he could look inside. Tan said that he
couldn't. Tan didn't sign the envelope. Obviously it didn't contain
the printouts. According to Khodarkhovsky, Ken Thompson felt that
IBM had engaged in psychological warfare against Kasparov and he
said that he would never work with C-J Tan again.
The best part of Khodarkovsky's book concerns an incident that took
place a few days after the match. Getting into the elevator on his
way to an interview with CNN's Larry King, Kasparov encountered
movie star Charles Bronson.
Bronson: It was your fault Garry.
Kasparov: Well, you can only see happy endings in Hollywood. But I
will get my revenge in the next match.
Bronson: You won't get a rematch from them.
Bronson knows show business when he sees it!
Kasparov's post-match comments sound like sour grapes, but they
make a lot of valid points. It's his pre-match comments that are
questionable. Deep Blue was destroyed. I say it was destroyed by
its performance in game one of the first match! The publicity
created by that game (probably DB's best performance, too) would
lead IBM to create the publicity stunt that the rematch became.
That publicity also affected Kasparov. See Hans Ree's theory in
New in Chess that maybe Kasparov made the first match seem closer
than it really was (or should have been) for two reasons: a
lucrative rematch; and to cover the fact that the PCA had no cycle
in place to determine the next challenger. K's statements leading
up to the rematch reinforce this theory; he basically states that
DB is his only worthy challenger. *After* the match, of course,
he says that DB should go through the same process as every other
chess player! But then it is too late because, as far as the media
are concerned there are no other chess players besides K and DB,
and it was K who told them so.
If IBM were interested in science and research, then Kasparov's
post-game comments would have some validity. Yes, he should have
received DB's practice games because IBM would want a full and
complete test of DB's capabilities and its flaws. Yes, DB should
play other players for the same reasons; different styles of play
by the opponent would more quickly test the machine. Yes, Kasparov
should abandon his anti-computer strategy; how would DB do against
a Kasparov innovation? Or how great is its understanding of openings
if it faces an innocuous move that isn't in its opening book (remem-
ber Seirawan ignoring a program's ...a6 in the Benoni because the
program didn't "know" that it was threatening ...b5)?
But Kasparov's pre-match hype served IBM's purpose. Nobody in the
media understood that the match proved absolutely nothing, even if
it was legitimate (and there was obviously a coverup, but a coverup
of what?). First of all it was a six game match, and if a six game
match was a true test of skill then all world championships would be
six game matches. Clearly Kasparov thought he would have an easy
time like in game 6 of the first match. Secondly, what did DB
demonstrate in the games it won. Remove Kasparov's name from game 2
and estimate the rating of the player of the Black pieces. 2400?
2200? Maybe even an A-player could have played that game! Do the same
for game 6. Now you would have trouble convincing anyone that the
loser of that game was rated over 2000! They might guess that he was
over 1600 because of 19...resigns!
IBM knew better than to claim any kind of superiority based on the
result of the match. Look at their full page ad in July 1997: "The
same IBM RS/6000 (tm) technology that, as 'Deep Blue', competed
against world champion Garry Kasparov is also conquering deep space."
Deep space is conquered, but they don't make the same claim about
Kasparov. And when they retired Deep Blue it was "we have climbed
Mt. Everest and now it is time to move on". It all reeks of "on our
lawyer's advice..." But it all works out, because the media will
say it for them. And the public won't know that way, way down in
some articles written after the match there were some questions
about how good DB was. Didn't GM Alburt say he would bet $10,000
on Gulko against DB? And that there were maybe 100 GMs who could
beat the machine (I think Seirawan suggested 50 GMs could do it).
But IBM got what they wanted, so now it's back to sponsoring golf
tournaments, a very low return on their PR dollar (even if they
didn't sponsor a golf tournament, everybody would think that they
probably did).
I know that some chess programmers will feel some "loyalty" to
their "compatriot", Deep Blue, just as chessplayers keep pointing
out all of the publicity Kasparov gets for chess. But Kasparov is
a black hole publicity-wise: the media only know one chessplayer-
Kasparov- and nothing about the whole rest of chess culture. The
same goes for Deep Blue and chess programming -- been there, done
that, why do you guys keep wasting your time? And you'll never
convince anyone otherwise because they think that DB plays perfect
chess. So they'll never believe that there are some important things
to discover in this field.
Expose Deep Blue or be doomed to the fringes of the geekerati of the
chess and programming worlds!
Enough for now.
Cheers,
Gerry
Daniel
<snip>
> >B Kg8 Qc5 Rb8 Bc6 Nd3 Pa7 c4 f7 g6 h7
>
> Now Adams played 32.-Re8.
> When I put the moves of that game into my database I had Fritz in the
> background as usual in CB7. He went for Rb2, the move all of you did
> mention.
It's a human's ability to find these better than perfect moves that a
computer just cannot find for some reason that will always make the human
eye better than any chess algorythm. Whether it be because the move is
winning, albeit unsound as far as chess theory goes or for some other
unforeseen reason; computers simply cannot see some of the correct moves in
a given position, however the moves they find may not be 'incorrect.'
Robert Hyatt
: <snip>
:> >W Kh1 Qe2 Rf1 Rg3 Pa3 c3 f5 g2 h2
:> >B Kg8 Qc5 Rb8 Bc6 Nd3 Pa7 c4 f7 g6 h7
:>
:> Now Adams played 32.-Re8.
: <another snip>
:> When I put the moves of that game into my database I had Fritz in the
:> background as usual in CB7. He went for Rb2, the move all of you did
:> mention.
: <SNIP>
: It's a human's ability to find these better than perfect moves that a
: computer just cannot find for some reason that will always make the human
: eye better than any chess algorythm. Whether it be because the move is
: winning, albeit unsound as far as chess theory goes or for some other
: unforeseen reason; computers simply cannot see some of the correct moves in
: a given position, however the moves they find may not be 'incorrect.'
Note that Rb2 outright wins, while Re8 doesn't. It may win, but it is
nowhere near as good as Rb2. It is just much "safer" than Rb2. But in
this case, Rb2 is smashing, much better than Re8.
--
ilias kastanas 08-14-90
Rolf Tueschen <TUESCHEN.MEDIZ...@t-online.de> wrote:
@TUESCHEN.MEDIZ...@t-online.de (Rolf Tueschen) wrote:
@
@>Computerchess Misc (1)
@
@>A couple of days ago I saw the following position for the first time.
Very nice; thanks for posting it. I'm sorry I saw it so late.
@It's from a game between Willie Watson v. Michael Adams.
@Played at the BCF-ch in 1986. Watson was IM with 2495.
@>W Kh1 Qe2 Rf1 Rg3 Pa3 c3 f5 g2 h2
@>B Kg8 Qc5 Rb8 Bc6 Nd3 Pa7 c4 f7 g6 h7
Extraneous inference: there must be "something" going on in
the ...Rb2 line (otherwise you wouldn't have posted!).
It was clear Wh's only chance was counterattack. But a perpetual,
so swiftly?!... Then I saw the "something"! And it fooled me; I said,
oh-so-clever, "...Nf4 looks like a turn-around, making for a win; but there
must be another turn-around in reply!" Like a miraculous perpetual after
... Nf4. Rh7+ or ...Nf4. Q:f4+, Qg5. Rh7+ ... Well, no, there isn't!
(I was under the influence of having looked at Mecking - Hort,
San Antonio 1972, Kc1, Qe2, Rd1, f3, Bd3, g5, Nb1, Pa2, b2, c2, e4, g2, h2;
Kg7, Qc7, Rc8, f8, Bb5, e7, Nh5, Pa6, b4, e5, e6, g6, h7. It went on
23. B:e7, R:f3 24. Q:f3 What should Black play now? Answer given below.)
@Now Adams played 32.-Re8.
@
@It was Noam Elkies who mentioned this move for the first time. And for me
@it's a very interesting problem in the middle of a computerchess group.
@When I put the moves of that game into my database I had Fritz in the
@background as usual in CB7. He went for Rb2, the move all of you did
@mention. And the move looked really strong. But it was one of the reasons I
@posted the position. I wanted to see if it were possible to talk about
@chess positions in a computerchess group. I wanted to know how many members
@would think on their own about the position -- at least for a short moment.
@Therefore I wrote the first questions. It was a good thing to see that
@we're not at all limited by our constant use of computers. Noam of course
@was the most independant author who had the guts to prefer a move normally
@no computer proposed at first. Noam found exactly Michael Adams' move. I'm
@not a GM but I brought that position into the group because I saw a
@principal difference between human chess and computerchess.
Eh, I used my computer to run the newsreader...
@As a human player OTB in a tournament like the BCF-ch you simply shouldn't
@play 'va banque' in won position into a line where your opponent has a
@long-ranged aggressive attack and where you only could defend your victory
@by a surprising move 6 moves later ... Note that Michael Adams at that time
@was a genial child prodigee but not an IM yet. He was 14 years old and had
@ELO 2360. So normally with not so great chances against IMs. I analysed a
@lot of these early games of the later super GM and it was quite evident for
@me that this sort of players must have that special strength to understand
@that a position is strong and that the well could choose the safer road for
@the win. The computer however has no problems with tiredness or emotions
@and if he sees a forced tactical line he will go for it.
Now it's a different question; not pure analysis, but "how should
Black think, OTB, and what decision should he reach?".
To set the stage, here is the answer to Mecking - Hort. Black
has the prosaic 24... Q:e7, of course; but he sees the sharp and forcing
alternative, 24... Nf4! Try a few lines; with his awkward K position,
Wh is in deep trouble... pretty much lost... Hort almost played it; at
the last moment he noticed 24... Nf4?? 25. Bd6!!, 1-0 in a couple of moves!
He played 24... Q:e7(!), and nursed his small advantage to a win, 60-odd
moves later.
OTB, one quickly sees ...Re8 and ...Rb2 are the candidates.
Brief eval: ...Re8 gives a solid advantage, Wh having hardly any counter-
play. ...Rb2 seems to mate/win material against all defenses. Then, Bl
notices Wh's counterattack, Qe6+ and possible perpetual.
Should he analyze that further? (Of course "va banque", ...Rb2 wi-
thout analysis, is out). The rational decision is No... go back to ...Re8.
Why spend time and possibly come up empty-handed when a good alternative
exists? And even if he has lots of time, and has noticed ...Nf4 very likely
wins, why risk a long, sharp line requiring detailed analysis and calcula-
tion? Realistically, one must take fallibility into account. We are
fallible. (Humans and computers... if for different reasons). Why run
the risk of mistakes, surprise replies (Hort!) etc if it isn't necessary?
So I agree; ...Re8 is the choice. Over a hundred similar-type
decisions, "Re8" 's will score more points than "Rb2" 's.
I did analyze ...Rb2, for the sake of interest; but OTB I would
have thought as above, and played ...Re8.
@Now the next point of the reasons why I posted exactly that position. When
@I looked a little bit deeper into the Rb2 line I discobvered something
@strange with the Fritz modul. It played always Kh8 instead of the winning
@Kh7. I could understand it. Therefore I started a little study with all the
@other programs. I already gave you the results. Fritz was still fixed on
@Kh8 although he was deep in 17 ply and some 20 minutes. The same with
@Junior who also had Kh8 as favorite drawing move. Of course both Fritz and
@Junior still showed a total winning score. Only Crafty presented the
@winning move after 2'12. Again, all the programs played Rb2 but all had Kh8
@or CSTal Kg7 on their agenda. I mean this is strange. Because if a human
Well, perpetuals present special difficulty to programs.
Take a simple case, Ke8, Pe6, f7, h6, h7. 1. Qc8+, Ke7 2. Qc7+,
Kf8 3. Qb8+, Kg7 4. Qg3+, Kf6 5. Qf4+, Kg6 6. Qg4+, Kf6 7. Qf4+, Ke7
8. Qc7+, Ke8 9. Qc8+ (the original position). A human clearly sees the
draw; but how will a program arrive at a value 0 (at least) for this position
P ... whose initial eval was -4, say? The human saw "P returns to P",
within an overall pattern; the program only sees, 16 ply deep, a repetition
of P... no basis for judging it a draw! It will take 16 more ply to
trigger the 3-fold repetition rule and assign 0 -- just to that node! Uh,
32 ply deep!?... With more and more such 0's emerging, at some stage the
overall fact "perpetual" will be established -- 0's for P and the other
positions appearing. Will it be within our lifetime, though?
The example is naive, but it illustrates that finding a perpetual
is an unnatural activity for eval-based algorithms. And it was a pure
one; consider perpetuals that incorporate mate/stalemate threats, or "quiet"
moves, or combinations to reach perpetuals...
So I'm not surprised if programs, surely more sophisticated than
above, still falter in this area!
@With fca I already discussed the possible similarity of the "hole" in the
@else drawing line with Kasparov's missing of the Qe3 draw. Willie Watson
@had no other choice in our position here. He had to go for Rg6 and then
@hoping for a perpetual. And likewise even an average player like M. Ashley
@saw on the stage that Kasparov's last chance could only be a perpetual. But
@to look at a draw a little bit closer, Garry had to go for it at first. But
@he didn't do that. Folks, Bobby Fischer did never do that. He played till
@the end. If there was the slightest chance to win or save a game. Kasparov
@behaved like a weak amateur who
"...the commentators' peak of creativity was the following
sentence:...
"It is not understandable how (Botvinnik/Tal) did not
notice this move... which little Willie Scherbakov, a
student in the sixth grade, section B, 217th S. R.
School immediately found"
All this is correct. But there really is nothing supernatu-
ral about the move the boy found. Actually, the continua-
tion is not at all complicated, and I think that when I was
in the sixth grade, I could have also found it. It is one
thing to watch the course of the match from the auditorium
or from the press room... and another thing to play on
stage, feeling the eyes of a thousand enthusiasts... to
play with unusual responsibility and great tension."
Mikhail Tal,
"Tal-Botvinnik, 1960"
So Kasparov made a mistake, too!
It was the reverse situation, I think. Watson's only-chance
attack was obvious; the opponent had to find a "hidden" move to foil
it. GK's ...Qe3 would have been suicidal if not followed by the
hidden, quiet move ...Re8 (and also needed a later, quiet ...h5).
It was a move he himself had to find.
Surely GK did think about ...Qe3; but he didn't notice ...Re8.
Nobody else did at that time, either; wishing for a perpetual is fine,
but how to bring it about?! The move was found later.
Mistake, yes; but it's hard to classify it as giving up without
a fight, or shirking one's duty!
Here is a position that should be an interesting test for humans
as well as computers.
Kh2, Qe4, Pd5, g2, h7; Kb7, Qc3, Pa7, b2. White to move.
Ilias
Rolf Tueschen
> Enough for now.
> Cheers,
> Gerry
Thanks very much, Gerry. I'd like to let your post uncorrected here hoping
that someone like Bob Hyatt tells us *his* (proxy) version. You know, he's
our man on usenet always fighting for IBM/DB ("buddies") vs. Kasparov (=
"asshole").
He's also the one I have criticised many times because he tried to neglect
your *science* parameter. The whole PR and interviews with Kasparov always
went around science and *not* simply a big ball shoot out. In short, what
could a machine achieve against one of the best humans. It was *not* about
"Let's see, if we were able to cheat on Kasparov."
It's some kind of strange that our humpty dumpty didn't enter the debate
yet. Probably your good post is a waste of time because people want to
answer and refutate within some hours. But your arguments are simply too
strong.
Let's watch out for some new tricks. Bob will surely try to "concentrate"
on my little post here and leave *yours* behind. Eh, I bet 1 Benjamin
against 3 processors of DB ... ... :)
Rolf Tueschen
I nead some extra time for your deep commentary and all the positions.
> Extraneous inference: there must be "something" going on in
>the ...Rb2 line (otherwise you wouldn't have posted!).
You know me from our Groningen/Lausanne inquiry ...
> Eh, I used my computer to run the newsreader...
:) You're not a typical computerchess freak, Ilias.
> Now it's a different question; not pure analysis, but "how should
>Black think, OTB, and what decision should he reach?".
That's what it was all about. I can't hide my main interests. Psychological
aspects.
> I did analyze ...Rb2, for the sake of interest;
Please, for the sake of my little idea every 6 months, try to analyse it a
little bit deeper. Also against the background of my original post and some
reactions.
> but OTB I would
>have thought as above, and played ...Re8.
> Well, perpetuals present special difficulty to programs.
Yes, but the experts in computerchess seem to think that they have to put
up with the "little" rollback (disadvantage) as sort of very rare exception
but *not* changing their program overall ... *Not* realizing that *then*
the programs simply can't be of a real GM class. I brought up the analogy
of a chain with *one* very weak link. In real contests humans would go
exactly on these "exception" for sure ... Therefore Elo only 2400 maximum.
Another misunderstanding is the implemantation of huge books. This is only
dangerous for amateurs. A GM has the same knowledge. More so he has exactly
analysed the very *rare* exceptions. DOn't you think so?
> So I'm not surprised if programs, surely more sophisticated than
>above, still falter in this area!
Yes. Question is for how long? Or is it a principal boundary?
> Here is a position that should be an interesting test for humans
>as well as computers.
> Kh2, Qe4, Pd5, g2, h7; Kb7, Qc3, Pa7, b2. White to move.
This and also the others. Please give us a little time.
Thanks for your wonderful posting.
Robert Hyatt
:> Close, Rolf, but there was *one* winner: not Deep Blue, but IBM.
:> They got exactly what they wanted i.e. $100 million in free
:> publicity according to the Wall Street Journal (actually it
:> probably cost them around $1million, but still a good return).
:> The more you examine it, the more you will be convinced that this
:> event was 100% business and 0% science.
It depends on whose perspective you look from. From Hsu, Campbell and
group, it was *always* about "science". Just as it has been for me, for
Ken Thompson, Dave Slate, and many others. Just because the IBM marketing
folks noticed what a hell of a P/R opportunity this turned into, doesn't
mean there was no "science"... It only means that the science got over-
ruled by "profit". And that actually seems reasonable, because IBM put
a good bit of hard cash into DB, counting the salaries of the guys doing
the real work, the GM's that were helping them, the hardware cost, etc.
Seems quite reasonable to me...
:> For instance, put this in Alta Vista: "Jeff Kisseloff" AND "Deep
:> Blue". See what lengths IBM went to in order to prevent all
:> discussion (science) while pumping up the PR (business).
And this is somehow a big surprise? I mean they dump hundreds of thousands
of dollars into a project, and should not give any consideration to the
possible profit potential it produced? You ought to attend an IBM
shareholder's meeting, get up on the stage, and suggest that. :)
:> And look at Michael Khodarkovsky's (sp?) book, A New Era: How
I would *not* have given Kasparov *anything*. Because I doubt if Hsu had
asked him after game one, "Garry, would you give me some ideas about which
moves you thought were best, and which moves you spent the most time on?"
Imagine what Kasparov would have thought. And answered. Do you think he
would have responded? *I* don't... And I think the DB team had no
responsibility to do what he obviously wouldn't have done himself.
As far as the printout, you are mistaken. The printout was for the entire
analysis for move 37, from the time where the eval dropped, until the time
DB announced it was using "panic time", until the time it had to stop and
"reconstruct the PV" until it played the move. It was cryptic. But it
looks just like the output from their program that I have seen for about 10
years at events they have attended. No mystery... no cheating... just
chess...
:> The best part of Khodarkovsky's book concerns an incident that took
:> place a few days after the match. Getting into the elevator on his
:> way to an interview with CNN's Larry King, Kasparov encountered
:> movie star Charles Bronson.
:> Bronson: It was your fault Garry.
:> Kasparov: Well, you can only see happy endings in Hollywood. But I
:> will get my revenge in the next match.
:> Bronson: You won't get a rematch from them.
:> Bronson knows show business when he sees it!
If I were IBM I would *never* play him again. He acted like a whining jerk,
rather than simply accepting defeat and saying "wait until next time when I
take this more seriously..." He had to accuse them of cheating, had to whine
about "a big company that only wanted to beat Kasparov".. get a clue.. it was
*all* about beating Kasparov. That has been the Holy Grail of computer chess
since many of us first wrote programs, to write a program that would beat the
world champion.
He was wrong in that regard... it wasn't about beating "Kasparov"... it was
*most definitely* about beating the world champion. Had Anand won, it would
have been Anand in New York, not the whining baby...
:> Kasparov's post-match comments sound like sour grapes, but they
Hmmm.. funny you should mention six game matches... Last time I looked the
FIDE world championship was won by a *blitz game* no less... So stranger
things have certainly happened. And the 6 game format was not IBM's idea..
that was Kasparov's idea... as were all the match constraints... he wanted
his clock.. he got his clock... he wanted this, he got this.. because he held
*all* the cards before match 2 took place. *after* match 2, IBM held all the
cards... which makes perfect sense..
And it doesn't matter *how* he played, it was *who*. Clearly the best player
that is currently playing active chess is Kasparov. Don't forget that his
opponent wasn't a casual wood-pusher either.. it was *very* strong...
:> IBM knew better than to claim any kind of superiority based on the
:> result of the match. Look at their full page ad in July 1997: "The
:> same IBM RS/6000 (tm) technology that, as 'Deep Blue', competed
:> against world champion Garry Kasparov is also conquering deep space."
:> Deep space is conquered, but they don't make the same claim about
:> Kasparov. And when they retired Deep Blue it was "we have climbed
:> Mt. Everest and now it is time to move on". It all reeks of "on our
:> lawyer's advice..." But it all works out, because the media will
:> say it for them. And the public won't know that way, way down in
:> some articles written after the match there were some questions
:> about how good DB was. Didn't GM Alburt say he would bet $10,000
:> on Gulko against DB? And that there were maybe 100 GMs who could
:> beat the machine (I think Seirawan suggested 50 GMs could do it).
:> But IBM got what they wanted, so now it's back to sponsoring golf
:> tournaments, a very low return on their PR dollar (even if they
:> didn't sponsor a golf tournament, everybody would think that they
:> probably did).
You need to attend a local university, and enroll in a "marketing 101"
course... it will all be explained clearly there... it *was* all about
money... Otherwise someone would be run out of town for spending so much
money and getting nothing in return. Large companies don't do such things
for goodwill alone. They expect something in return...
:>
:> I know that some chess programmers will feel some "loyalty" to
:> their "compatriot", Deep Blue, just as chessplayers keep pointing
:> out all of the publicity Kasparov gets for chess. But Kasparov is
:> a black hole publicity-wise: the media only know one chessplayer-
:> Kasparov- and nothing about the whole rest of chess culture. The
:> same goes for Deep Blue and chess programming -- been there, done
:> that, why do you guys keep wasting your time? And you'll never
:> convince anyone otherwise because they think that DB plays perfect
:> chess. So they'll never believe that there are some important things
:> to discover in this field.
I don't think it plays perfect chess. I think it beat Kasparov in a 6 game
match. nothing more. nothing less.
:> Expose Deep Blue or be doomed to the fringes of the geekerati of the
:> chess and programming worlds!
:> Enough for now.
:> Cheers,
:> Gerry
: Thanks very much, Gerry. I'd like to let your post uncorrected here hoping
: that someone like Bob Hyatt tells us *his* (proxy) version. You know, he's
: our man on usenet always fighting for IBM/DB ("buddies") vs. Kasparov (=
: "asshole").
Nice characterization of Kasparov. I think I like it, in fact..
: He's also the one I have criticised many times because he tried to neglect
: your *science* parameter. The whole PR and interviews with Kasparov always
: went around science and *not* simply a big ball shoot out. In short, what
: could a machine achieve against one of the best humans. It was *not* about
: "Let's see, if we were able to cheat on Kasparov."
You keep worrying about the science issue... IBM is worrying about the
money issue. That was what this was about from the beginning... not a science
experiment (that was what Hsu, and group were doing).. but IBM wanted to take
the science and turn it into a profit. They accomplished that...
: It's some kind of strange that our humpty dumpty didn't enter the debate
: yet. Probably your good post is a waste of time because people want to
: answer and refutate within some hours. But your arguments are simply too
: strong.
Humpty was busy debugging a network problem...
: Let's watch out for some new tricks. Bob will surely try to "concentrate"
: on my little post here and leave *yours* behind. Eh, I bet 1 Benjamin
: against 3 processors of DB ... ... :)
You just lost...
Again...
ilias kastanas 08-14-90
Rolf Tueschen <TUESCHEN.MEDIZ...@t-online.de> wrote:
@ika...@sol.uucp (ilias kastanas 08-14-90) wrote:
@
@I nead some extra time for your deep commentary and all the positions.
@
@> Extraneous inference: there must be "something" going on in
@>the ...Rb2 line (otherwise you wouldn't have posted!).
@
@You know me from our Groningen/Lausanne inquiry ...
Yes...
@> Eh, I used my computer to run the newsreader...
@
@ :) You're not a typical computerchess freak, Ilias.
Hey, you said no computer help, didn't you?!
@> Now it's a different question; not pure analysis, but "how should
@>Black think, OTB, and what decision should he reach?".
@
@That's what it was all about. I can't hide my main interests. Psychological
@aspects.
Yes, it's all part of it. (E.g. the "intentional time trouble"
trick). As well as what to analyze and how long. (Story: Someone asked
Miles after a game vs Karpov: "What would you have done against B:h7+
(obscure sac)?" Miles laughed: "Karpov doesn't play such moves!")
@> I did analyze ...Rb2, for the sake of interest;
@
@Please, for the sake of my little idea every 6 months, try to analyse it a
@little bit deeper. Also against the background of my original post and some
@reactions.
OK, next time I'll try!
@> Well, perpetuals present special difficulty to programs.
@
@
@Yes, but the experts in computerchess seem to think that they have to put
@up with the "little" rollback (disadvantage) as sort of very rare exception
@but *not* changing their program overall ... *Not* realizing that *then*
@the programs simply can't be of a real GM class. I brought up the analogy
@of a chain with *one* very weak link. In real contests humans would go
@exactly on these "exception" for sure ... Therefore Elo only 2400 maximum.
Well, it's not easy to fix, being 'alien' to the standard algo-
rithm... And writing special code means a time penalty.
Yes, a well-informed human will capitalize on this, and similar
things... Like the RBB vs RR trick I mentioned in a post about the
Anand games. (I tried it against some programs, angling for positions
where they can give up B+N, say, for R+P or R+2P... They go for it,
and then their ending is long-term lost just about every time).
@Another misunderstanding is the implemantation of huge books. This is only
@dangerous for amateurs. A GM has the same knowledge. More so he has exactly
@analysed the very *rare* exceptions. DOn't you think so?
Story: Portisch to Tal, during post-mortem:
"-- Why didn't you play 14. Rd3 ?"
Tal: "-- It never crossed my mind..."
"-- ...Eh?! You've played that move before!
"-- When? Where?"
"-- Against Benko (Candidates)..."
A huge games-database book might be safer then ECO, say... when
it comes to avoiding GM analyses/"cooks".
@> So I'm not surprised if programs, surely more sophisticated than
@>above, still falter in this area!
@
@Yes. Question is for how long? Or is it a principal boundary?
We'll see... First, it has to start costing points.
@> Here is a position that should be an interesting test for humans
@>as well as computers.
@
@
@> Kh2, Qe4, Pd5, g2, h7; Kb7, Qc3, Pa7, b2. White to move.
@
@
@This and also the others. Please give us a little time.
@
@Thanks for your wonderful posting.
Thanks; I'm glad you liked it.
Ilias
@
@
Rolf Tueschen
>In article <6sf076$foe$3...@news01.btx.dtag.de>,
>Rolf Tueschen <TUESCHEN.MEDIZ...@t-online.de> wrote:
>@ika...@sol.uucp (ilias kastanas 08-14-90) wrote:
>@
>@I nead some extra time for your deep commentary and all the positions.
>@
Ilias, please give me the other part of your position with the pawns and
queens. You forgot the Whites. You remember, the ending, not Mecking-Hort.
ilias kastanas 08-14-90
Rolf Tueschen <TUESCHEN.MEDIZ...@t-online.de> wrote:
@ika...@sol.uucp (ilias kastanas 08-14-90) wrote:
@
@>Rolf Tueschen <TUESCHEN.MEDIZ...@t-online.de> wrote:
@
@Ilias, please give me the other part of your position with the pawns and
@queens. You forgot the Whites. You remember, the ending, not Mecking-Hort.
They must have gotten chopped off somehow. Here is a copy:
@@ Kh2, Qe4, Pd5, g2, h7; Kb7, Qc3, Pa7, b2. White to move.
Ilias
Rolf Tueschen
No, that's not the position. You gave another one. Then first move Qc8 or
such. D'you remember? :)
> Ilias
Gerry Forbes
Because of problems with my news server I did not receive
Robert Hyatt's response to my post in this thread. I did go to dejanews
to read it, however, and I think some of his comments are debatable.
He points out that science was (is) the primary objective of the
programmers in the computer chess field. I agree, and apologise if
I seemed to imply otherwise. He goes on to say that IBM, seeing an
opportunity to get a direct return from their investment, had every right
to ensure that profit overruled science. I wonder if he meant that profit
should be allowed to overrule science terminally. How much profit was
there in dismantling Deep Blue -- does IBM sell used hardware now?
Not that there was much IBM could do to affect the balance once the
match was started *unless* they were up to something underhanded. Why
censor their web page reporter? Kasparov's suspicions were going to
come out anyway -- what did it matter if he looked the paranoid whiner
a couple of days earlier? At least it wouldn't look like IBM was trying
to hide something if they were the first to report it.
Hyatt claims that IBM did comply with Kasparov's request for
printouts by giving the printout for the whole move 37, and not just
the side variation 37.Qb6. My recollection is that only the side
variation was published in the NY Times, but then it is possible that
the Times chose not to publish the complete printout. Even so, this
could hardly be said to satisfy Kasparov's request since he asked for
the printouts for *3* moves. The point was to compare the printouts.
IBM's actions suggest that they provided the only printouts for the
only honest move among the three; notice that they didn't provide the
printout for 36.Bxd6, which Deep Blue spent 15 minutes on.
Hyatt says that he wouldn't give Kasparov a rematch if he
were IBM. Personally, I'm not interested in seeing another rematch
between these two, but if IBM were interested in science at all then
they should *want* to find out what their machine is capable of, not
only by playing Kasparov again, but by playing a variety of other
players as well. The names Anand, Karpov, Seirawan, Andersson, Gulko,
and others have been mentioned. I'd like to see a Suttles-Deep Blue
match. Monty Newborn, in his book "Kasparov versus Deep Blue" says that
if Deep Blue were to win the second match then Kasparov would be
*entitled* to another match.
Then there are the contradictory statements. Beating the
world champion is the Holy Grail of computer chess -- a milestone of
the quality of computing, apparently. Apparently not, since it doesn't
matter how the computer plays, only who it plays. A milestone of quality
where quality doesn't count for anything. So it seems that absolutely
no progress has been made since the Turk. And either Deep Blue was
playing the world champion, *or* Karpov-Anand was a world championship
match, not both. Anyways, there was a lot of doubt about the validity
of a 6-game match being a true test of strength in that case as well.
Hyatt seems to have misread part of my post when he protest that
he doesn't think that Deep Blue plays perfect chess. I said that
*the public* thinks that Deep Blue plays perfect chess and that they
would think that chess programmers are wasting their time on a problem
that is already solved, i.e. that the Deep Blue-Kasparov match was bad
for chess and computer chess (or at least the way the match was mis-
reported was bad for both). When Microsoft bundles SuperChessDude in
Windows 2000 the market will die for chess programs and computer cheating
will be the norm on the chess servers.
Hyatt also seems to think that IBM owes no "goodwill" to the
chess community. Sure, chess has contributed absolutely nothing to
computing. Golf on the other hand has been absolutely indispensable.
But IBM will never sponsor a chess event in North America now because
the media will want to know "Where's Deep Blue? How come it isn't
competing?" and IBM won't want to be as forthright as Hyatt in admitting
"It was all about money."
Robert Hyatt
: Because of problems with my news server I did not receive
: Robert Hyatt's response to my post in this thread. I did go to dejanews
: to read it, however, and I think some of his comments are debatable.
: He points out that science was (is) the primary objective of the
: programmers in the computer chess field. I agree, and apologise if
: I seemed to imply otherwise. He goes on to say that IBM, seeing an
: opportunity to get a direct return from their investment, had every right
: to ensure that profit overruled science. I wonder if he meant that profit
: should be allowed to overrule science terminally. How much profit was
: there in dismantling Deep Blue -- does IBM sell used hardware now?
: Not that there was much IBM could do to affect the balance once the
: match was started *unless* they were up to something underhanded. Why
: censor their web page reporter? Kasparov's suspicions were going to
: come out anyway -- what did it matter if he looked the paranoid whiner
: a couple of days earlier? At least it wouldn't look like IBM was trying
: to hide something if they were the first to report it.
This happens all the time. Trade secrets... industrial espionage...
you name it... vendors never release technical details of something until
it is "done" and the next version is already operational. IE try to find
out some technical details about non-released microprocessors, or non-
released digital signal processors, or ...
: Hyatt claims that IBM did comply with Kasparov's request for
: printouts by giving the printout for the whole move 37, and not just
: the side variation 37.Qb6. My recollection is that only the side
: variation was published in the NY Times, but then it is possible that
: the Times chose not to publish the complete printout. Even so, this
: could hardly be said to satisfy Kasparov's request since he asked for
: the printouts for *3* moves. The point was to compare the printouts.
: IBM's actions suggest that they provided the only printouts for the
: only honest move among the three; notice that they didn't provide the
: printout for 36.Bxd6, which Deep Blue spent 15 minutes on.
The entire printout for the search in question was published... a computer
doesn't produce "side variations" of any kind, except for an old PV that is
replaced by a new PV after a deeper search.
: Hyatt says that he wouldn't give Kasparov a rematch if he
: were IBM. Personally, I'm not interested in seeing another rematch
: between these two, but if IBM were interested in science at all then
: they should *want* to find out what their machine is capable of, not
: only by playing Kasparov again, but by playing a variety of other
: players as well. The names Anand, Karpov, Seirawan, Andersson, Gulko,
: and others have been mentioned. I'd like to see a Suttles-Deep Blue
: match. Monty Newborn, in his book "Kasparov versus Deep Blue" says that
: if Deep Blue were to win the second match then Kasparov would be
: *entitled* to another match.
Again, this is *not* about science to IBM. At the request of "shareholders".
They don't give a squat about how DB would do in the world of chess, they do
worry about maximum dividends on the stocks they own... I'm sure Hsu and
company would love to play DB in every event they could. But that's not the
way for maximum marketing edge... Now once the excitement of the match has
faded to a distant memory, another "outbreak" would probably be profitable
for them, even if they lose, but it will likely be a while...
: Then there are the contradictory statements. Beating the
: world champion is the Holy Grail of computer chess -- a milestone of
: the quality of computing, apparently. Apparently not, since it doesn't
: matter how the computer plays, only who it plays. A milestone of quality
: where quality doesn't count for anything. So it seems that absolutely
: no progress has been made since the Turk. And either Deep Blue was
: playing the world champion, *or* Karpov-Anand was a world championship
: match, not both. Anyways, there was a lot of doubt about the validity
: of a 6-game match being a true test of strength in that case as well.
We only have Kasparov to blame for the 6 game match. As I mentioned, *he*
controlled *everything* about the match. IBM *had* to accept any conditions
he imposed, because they had to play *him* as the recognized best chess player
in the world. When you hold all the cards, and you make a poor agreement on
conditions, you only have yourself to blame...
: Hyatt seems to have misread part of my post when he protest that
: he doesn't think that Deep Blue plays perfect chess. I said that
: *the public* thinks that Deep Blue plays perfect chess and that they
: would think that chess programmers are wasting their time on a problem
: that is already solved, i.e. that the Deep Blue-Kasparov match was bad
: for chess and computer chess (or at least the way the match was mis-
: reported was bad for both). When Microsoft bundles SuperChessDude in
: Windows 2000 the market will die for chess programs and computer cheating
: will be the norm on the chess servers.
: Hyatt also seems to think that IBM owes no "goodwill" to the
: chess community. Sure, chess has contributed absolutely nothing to
: computing. Golf on the other hand has been absolutely indispensable.
: But IBM will never sponsor a chess event in North America now because
: the media will want to know "Where's Deep Blue? How come it isn't
: competing?" and IBM won't want to be as forthright as Hyatt in admitting
: "It was all about money."
I'd answer the same way I have had to answer the "Where is Cray Blitz?" for
a few tournaments... getting a 60 million dollar computer dedicated to play
chess for several days is *non-trivial*... wish it was otherwise, but it
represents resources that can be used better elsewhere, particularly after
such a huge advertising coup as this match..
ilias kastanas 08-14-90
Rolf Tueschen <TUESCHEN.MEDIZ...@t-online.de> wrote:
@ika...@sol.uucp (ilias kastanas 08-14-90) wrote:
@
@>@queens. You forgot the Whites. You remember, the ending, not Mecking-Hort.
@
@> They must have gotten chopped off somehow. Here is a copy:
@
@>@@ Kh2, Qe4, Pd5, g2, h7; Kb7, Qc3, Pa7, b2. White to move.
@
@No, that's not the position. You gave another one. Then first move Qc8 or
@such. D'you remember? :)
Oh, that... It was just a skeleton, right. For a full position
of that type, off the top of my head, try
Kb1, Qc7, Pa2, d2, e3, f2, h2; Ke8, Qa3, Bc1, Pa4, b4, d3, e6, f7.
Wh, to move, has +'s from c8-c7-f4 etc... always a Bishop down
but, as a human sees, with a perpetual (whose cycle, back to c8+, is 16 ply).
Wh also has the "tempting" 1. Q:c1, Q:Q+ 2. K:Q -- equalizing the material,
and with an outside PP to boot! Unfortunately, 2... a3! -+ wins for Black,
as the WK is trapped and can never move to d1.
That's the idea; will a program "see" the perpetual, or will it
accept the Greek gift?
One could refine this... or tinker with the parameters: make the
perpetual shorter, or longer/more involved. Or, the "catch" in the temp-
tation may prove too deep (long-term); one can make it simpler, maybe even
a position not dead lost but just a Pawn down for Wh, say.
Have fun!
Ilias
Ozgur Karabiyik
its mind recommendations to 1.Qc8 1.Qc6 and 1.Qb8 his fourth choice is still
1.Qxc1 with 0.41 points to black.
ilias kastanas 08-14-90 wrote in message <6su6cv$pis$1...@hades.csu.net>...
ilias kastanas 08-14-90
Ozgur Karabiyik <kara...@easynet.co.uk> wrote:
@My program recommend Qxc1 at first but in two seconds of analysis changes
@its mind recommendations to 1.Qc8 1.Qc6 and 1.Qb8 his fourth choice is still
@1.Qxc1 with 0.41 points to black.
@ilias kastanas 08-14-90 wrote in message <6su6cv$pis$1...@hades.csu.net>...
@> Kb1, Qc7, Pa2, d2, e3, f2, h2; Ke8, Qa3, Bc1, Pa4, b4, d3, e6, f7.
@>
@> Wh, to move, has +'s from c8-c7-f4 etc... always a Bishop down
@>but, as a human sees, with a perpetual (whose cycle, back to c8+, is 16 ply).
@>Wh also has the "tempting" 1. Q:c1, Q:Q+ 2. K:Q -- equalizing the material,
@>and with an outside PP to boot! Unfortunately, 2... a3! -+ wins for Black,
@>as the WK is trapped and can never move to d1.
Good choice! Does it evaluate the position as a draw? (what does
0.41 refer to?!). Is its PV the perpetual? If no PV, what does it do
after 1... Ke7 2. Qc7+, Kf6 -- does it go for Qf4+, "giving up" Q:c1?
Ilias
Rolf Tueschen
> Here is a position that should be an interesting test for humans
>as well as computers.
> Kh2, Qe4, Pd5, g2, h7; Kb7, Qc3, Pa7, b2. White to move.
Yes. My programs want to play 1.d5-d6. It's a difficult position. Or is
1.Qe7 better? ;)
Let's wait if some other people have something interesting to say.
Ilias, just as a little tongue in cheek. It's obvious that most interests
want to exchange 'opinions' about God and anything. Me included BTW.
But the moment we should put some pieces on the board or into our
computerchess programs it's hard work ...
Good to know of someone who constantly comes with some own analyses. In the
name of all others I beg you to have some leniency for us average mortals.
:)
> Ilias
ilias kastanas 08-14-90
Rolf Tueschen <TUESCHEN.MEDIZ...@t-online.de> wrote:
@ika...@sol.uucp (ilias kastanas 08-14-90) wrote:
@
@>as well as computers.
@
@Yes. My programs want to play 1.d5-d6. It's a difficult position. Or is
@1.Qe7 better? ;)
@
@Let's wait if some other people have something interesting to say.
To be sure, these two are the only candidates for the 1st move.
We could wait a bit more... Let me know when you want me to answer the
question, or give details.
@Ilias, just as a little tongue in cheek. It's obvious that most interests
@want to exchange 'opinions' about God and anything. Me included BTW.
@But the moment we should put some pieces on the board or into our
@computerchess programs it's hard work ...
People are busy defining God, over in sci.logic, even as we speak...
In the words of Linus Van Pelt, "the theological implications alone are
staggering..."
@Good to know of someone who constantly comes with some own analyses. In the
@name of all others I beg you to have some leniency for us average mortals.
@:)
Actually, mortals are doing just fine... Some of them did get
to the bottom of this position. (Not the two in whose game it came up...
but never mind.) The question is, will some computers also do so?
Will even _one_ of them do so?!
Ilias
Rolf Tueschen
>Again, this is *not* about science to IBM. At the request of "shareholders".
>They don't give a squat about how DB would do in the world of chess, they do
>worry about maximum dividends on the stocks they own... I'm sure Hsu and
>company would love to play DB in every event they could. But that's not the
>way for maximum marketing edge... Now once the excitement of the match has
>faded to a distant memory, another "outbreak" would probably be profitable
>for them, even if they lose, but it will likely be a while...
Here I see a new problem in our adjourned game.
Let's take for granted all that for the next five minutes. Then I still
have a problem. I for one, no, the whole world, look into (what was it)
NEWSWEEK and so on, have read before the match the big reports. With
interviews with Kasparov. Nowhere did I read something about dividends.
Nowhere did I read something about a maximum marketing edge.
Instead I could read about the "match" that should answer the question
who's the best chessplayer? A human being, the PCA Wch for the side of
"mankind" OR the computer system Deep Blue for the side "machine".
Excuse me for the trivial repetition, but one tends to forget about these
trivial things.
Now my problem could be divided into two questions.
Either it's right what you have written above, then all the pre-match PR
was a complete lie. That would also correspond with your statement that for
you the "match" hasn't proven that DB is the better player ...
If that was a lie, then one were allowed to conclude that the
IBM/DB/teamsters did cheat on the whole world, the two of us included.
Or, Bob, this was NOT a lie. It was really about the almost scientifical
question machine or mankind ...?
But then the IBM/DB/teamsters did also cheat on all of us. And on Kasparov
in particular. (That was my theory in the other thread, where we adjourned
the question because you didn't want to discuss about cheats.)
The new situation is full of suspense. How will Bob Hyatt explain that IBM
didn't cheat? Whether before or after the match, sorry, in between?
P.S.
I would be nice if you could also join the debate about the position from
Ilias given here in Misc (1) where White should decide between 1.d6 and
1.Qe7.
My programs want 1.d6. What's about you? Let's concentrate on these
original computerchess questions too. :)
Robert Hyatt
:>Again, this is *not* about science to IBM. At the request of "shareholders".
:>They don't give a squat about how DB would do in the world of chess, they do
:>worry about maximum dividends on the stocks they own... I'm sure Hsu and
:>company would love to play DB in every event they could. But that's not the
:>way for maximum marketing edge... Now once the excitement of the match has
:>faded to a distant memory, another "outbreak" would probably be profitable
:>for them, even if they lose, but it will likely be a while...
: Here I see a new problem in our adjourned game.
: Let's take for granted all that for the next five minutes. Then I still
: have a problem. I for one, no, the whole world, look into (what was it)
: NEWSWEEK and so on, have read before the match the big reports. With
: interviews with Kasparov. Nowhere did I read something about dividends.
: Nowhere did I read something about a maximum marketing edge.
This doesn't matter at all. IBM is a publicly traded company that has to
answer to stockholders. They can't toss millions of dollars down a small
hole with zero hope of ever getting anything back from it. I doubt they
ever anticipated the tremendous P/R boost they would get, until after the
first DB match. At that point, I'd bet my hat that things changes, from Hsu
and group having control, to the accountants/lawyers having control because
suddenly this looks like a giant "cash cow" that can be milked for a lot of
money, *if* done carefully...
So it became an accounting issue to the company, rather than a computer chess
issue to Hsu and group. Because remember that they had played in *many*
events before becoming married to IBM. They didn't miss ACM or WCCC events,
and they played lots of GM games to earn the second-stage Fredkin prize...
: Instead I could read about the "match" that should answer the question
: who's the best chessplayer? A human being, the PCA Wch for the side of
: "mankind" OR the computer system Deep Blue for the side "machine".
That's simply marketing hype. The match was to answer "can a computer beat
the best player in the world in a regular match at normal time controls."
Nothing about fairness, nothing about sharing game scores, nothing about
telling each other what was going on, just that one simple question...
: Excuse me for the trivial repetition, but one tends to forget about these
: trivial things.
: Now my problem could be divided into two questions.
: Either it's right what you have written above, then all the pre-match PR
: was a complete lie. That would also correspond with your statement that for
: you the "match" hasn't proven that DB is the better player ...
Have you *ever* seen ad advertisement that didn't add hyperbole, didn't
stretch the facts just a bit, didn't embellish a tad to make it more news-
worthy? Etc...
Most adds are more fiction than fact... nothing new there at all...
: If that was a lie, then one were allowed to conclude that the
: IBM/DB/teamsters did cheat on the whole world, the two of us included.
: Or, Bob, this was NOT a lie. It was really about the almost scientifical
: question machine or mankind ...?
It was a scientific question: "can the machine beat the best player in the
world in a regular match at normal tournament time controls?" Nothing else.
Not "can the machine beat the best player in the world in a regular match at
normal tournament time controls, after he has had a chance to study lots of
games by the machine to discover its weaknesses?" That's a *different*
question. An interesting one that I enjoy working with, because this is
*exactly* what Bruce (Ferret) and I try to do every day with programs that
play automatically on the chess servers. We simply say "come and play us
and let's see if you can find a weakness you can exploit repeatedly. If you
do, we will fix it." But that's a different issue from DB. It is a reasonable
question. But nobody forced them to ask/answer *that* question. They picked
the question, and then answered it. Just because we don't like the specific
question they asked, doesn't mean it isn't a legitimate question, because 20
years ago no one gave a computer a chance of being able to answer any such
question with "yes".
: But then the IBM/DB/teamsters did also cheat on all of us. And on Kasparov
: in particular. (That was my theory in the other thread, where we adjourned
: the question because you didn't want to discuss about cheats.)
Nope... As I said, they just didn't ask and answer the question *you* wanted
them to. You are free to write a chess program, ask that question, then answer
it with your program. I am able to do this as well. So are they, *if* they
were interested in that question. Because, as I said before, this is certainly
a legitimate question. But it isn't the *only* question...
: The new situation is full of suspense. How will Bob Hyatt explain that IBM
: didn't cheat? Whether before or after the match, sorry, in between?
see above. :)
: P.S.
: I would be nice if you could also join the debate about the position from
: Ilias given here in Misc (1) where White should decide between 1.d6 and
: 1.Qe7.
: My programs want 1.d6. What's about you? Let's concentrate on these
: original computerchess questions too. :)
I tried to stay out of that. Whittington seems to think that too much
"crafty likes this" goes on here anyway...
Rolf Tueschen
>In rec.games.chess.computer Rolf Tueschen <TUESCHEN.MEDIZ...@t-online.de> wrote:
>: Robert Hyatt <hy...@crafty.cis.uab.edu> wrote:
>:>Again, this is *not* about science to IBM. At the request of "shareholders".
>:>They don't give a squat about how DB would do in the world of chess, they do
>:>worry about maximum dividends on the stocks they own... I'm sure Hsu and
>:>company would love to play DB in every event they could. But that's not the
>:>way for maximum marketing edge... Now once the excitement of the match has
>:>faded to a distant memory, another "outbreak" would probably be profitable
>:>for them, even if they lose, but it will likely be a while...
>: Here I see a new problem in our adjourned game.
>: Let's take for granted all that for the next five minutes. Then I still
>: have a problem. I for one, no, the whole world, look into (what was it)
>: NEWSWEEK and so on, have read before the match the big reports. With
>: interviews with Kasparov. Nowhere did I read something about dividends.
>: Nowhere did I read something about a maximum marketing edge.
>This doesn't matter at all. IBM is a publicly traded company that has to
>answer to stockholders. They can't toss millions of dollars down a small
>hole with zero hope of ever getting anything back from it. I doubt they
>ever anticipated the tremendous P/R boost they would get, until after the
>first DB match. At that point, I'd bet my hat that things changes, from Hsu
>and group having control, to the accountants/lawyers having control because
>suddenly this looks like a giant "cash cow" that can be milked for a lot of
>money, *if* done carefully...
Let me try to translate you. Because IBM is a big company, that is mainly
existing to make a lot of money, it's no surprise that DB & team under the
lead and money of IBM had to "cheat" on Kasparov. I, Bob Hyatt would even
say that therefore it wasn't even "cheating", *whatever* it was in the end.
My personal comment would be this. For me, personally, the former
scientists Hsu et al have lost their reputation as scientists. Simply
because they betrayed their mainly scientific duty. Not that I see how
seductive the money could be for anybody, *but* if the scientists let
dictate their behaviour by the business, then they have said good-bye to
serious science. Simple fact.
>So it became an accounting issue to the company, rather than a computer chess
>issue to Hsu and group. Because remember that they had played in *many*
>events before becoming married to IBM. They didn't miss ACM or WCCC events,
>and they played lots of GM games to earn the second-stage Fredkin prize...
Don't ever assume that I misregard the earlier honesty of the scientists.
You know the details better than me.
>: Instead I could read about the "match" that should answer the question
>: who's the best chessplayer? A human being, the PCA Wch for the side of
>: "mankind" OR the computer system Deep Blue for the side "machine".
>That's simply marketing hype. The match was to answer "can a computer beat
>the best player in the world in a regular match at normal time controls."
>Nothing about fairness, nothing about sharing game scores, nothing about
>telling each other what was going on, just that one simple question...
Several new isues. Let me isolate this one for a try. All that might have
been the case and ok but then I ask you why the machine had to be changed
during the match. I mean a machine wanted to beat the human player.
Therefore the machine should have done all that by itself and *not* over
night with the help of many GMs and technicians. Don't you see the
contradiction?
>: Excuse me for the trivial repetition, but one tends to forget about these
>: trivial things.
>: Now my problem could be divided into two questions.
>: Either it's right what you have written above, then all the pre-match PR
>: was a complete lie. That would also correspond with your statement that for
>: you the "match" hasn't proven that DB is the better player ...
>Have you *ever* seen ad advertisement that didn't add hyperbole, didn't
>stretch the facts just a bit, didn't embellish a tad to make it more news-
>worthy? Etc...
I think you misinterprete my wordings. I would see nothing new if a company
cheated about the reality (as you explained very well). But the point you
seem to overlook is the fact that it was Kasparov who distributed the big
scientifivc question. Now here is the point where I see a clear
misinformation/ misleading of Kasparov himself. By whom? Try to find an
answer. Surely Kasparov didn't talk too much with IBM. So I see only one
group who had cheated him already before the match. And that was, NB the
second "crime" of formerly honest scientists, the DB team around Hsu and
Tan.
The sun is shining much less right now, I fear.
>Most adds are more fiction than fact... nothing new there at all...
Now under the impression of my last paragraphe I think DB team is guilty
even more ...
>: If that was a lie, then one were allowed to conclude that the
>: IBM/DB/teamsters did cheat on the whole world, the two of us included.
>: Or, Bob, this was NOT a lie. It was really about the almost scientifical
>: question machine or mankind ...?
>It was a scientific question: "can the machine beat the best player in the
>world in a regular match at normal tournament time controls?" Nothing else.
>Not "can the machine beat the best player in the world in a regular match at
>normal tournament time controls, after he has had a chance to study lots of
>games by the machine to discover its weaknesses?" That's a *different*
>question.
Yup.
> An interesting one that I enjoy working with, because this is
>*exactly* what Bruce (Ferret) and I try to do every day with programs that
>play automatically on the chess servers. We simply say "come and play us
>and let's see if you can find a weakness you can exploit repeatedly. If you
>do, we will fix it." But that's a different issue from DB.
From IBM, excuse my impertinency.
>It is a reasonable
>question. But nobody forced them to ask/answer *that* question. They picked
>the question, and then answered it. Just because we don't like the specific
>question they asked, doesn't mean it isn't a legitimate question, because 20
>years ago no one gave a computer a chance of being able to answer any such
>question with "yes".
The main difference for me is the implementation of all the million games.
>: But then the IBM/DB/teamsters did also cheat on all of us. And on Kasparov
>: in particular. (That was my theory in the other thread, where we adjourned
>: the question because you didn't want to discuss about cheats.)
>Nope... As I said, they just didn't ask and answer the question *you* wanted
>them to. You are free to write a chess program, ask that question, then answer
>it with your program. I am able to do this as well. So are they, *if* they
>were interested in that question. Because, as I said before, this is certainly
>a legitimate question. But it isn't the *only* question...
>: The new situation is full of suspense. How will Bob Hyatt explain that IBM
>: didn't cheat? Whether before or after the match, sorry, in between?
>see above. :)
>: P.S.
>: I would be nice if you could also join the debate about the position from
>: Ilias given here in Misc (1) where White should decide between 1.d6 and
>: 1.Qe7.
>: My programs want 1.d6. What's about you? Let's concentrate on these
>: original computerchess questions too. :)
>I tried to stay out of that. Whittington seems to think that too much
>"crafty likes this" goes on here anyway...
Honestly. How many seconds do you become a non-politician in a day? :)
So you stay away from *my* threads because Chris told something about
crafty ...
Don't let you become too oversensitive. It's really a position where you
programmers could see what your babies could achieve. C'mon.
Robert Hyatt
: Robert Hyatt <hy...@crafty.cis.uab.edu> wrote:
Your conclusion is wrong. You are confusing "Hsu, the scientist" with
"IBM, the publicly held corporation". Hsu works for IBM. He can have
the purest motives in the world, but the company pays his salary, and
pays for his hardware development, and pays for the others in the group,
and pays for the GM's to help them, and pays for the events, and pays for
the P/R before the event.. So "the company" controls this.
You are making an assumption about "Hsu, the scientist" that is *wrong*.
IE, suppose that UAB decided that Crafty has become popular enough that
they'd like to take advantage of this and start selling it. Could they
do this? probably. would they do it? Almost certainly not ever. But
*if* they did, does it suddenly become "Bob, the commercial programmer
when he promised he would never do this?" If so, you are misdirecting
your anger. If not, then there is no reason to make such comments about
Hsu either...
:>So it became an accounting issue to the company, rather than a computer chess
:>issue to Hsu and group. Because remember that they had played in *many*
:>events before becoming married to IBM. They didn't miss ACM or WCCC events,
:>and they played lots of GM games to earn the second-stage Fredkin prize...
: Don't ever assume that I misregard the earlier honesty of the scientists.
: You know the details better than me.
:>: Instead I could read about the "match" that should answer the question
:>: who's the best chessplayer? A human being, the PCA Wch for the side of
:>: "mankind" OR the computer system Deep Blue for the side "machine".
:>That's simply marketing hype. The match was to answer "can a computer beat
:>the best player in the world in a regular match at normal time controls."
:>Nothing about fairness, nothing about sharing game scores, nothing about
:>telling each other what was going on, just that one simple question...
: Several new isues. Let me isolate this one for a try. All that might have
: been the case and ok but then I ask you why the machine had to be changed
: during the match. I mean a machine wanted to beat the human player.
: Therefore the machine should have done all that by itself and *not* over
: night with the help of many GMs and technicians. Don't you see the
: contradiction?
No. I do this *every* last day. They felt the bishop pair bonus had been
set too high. They adjusted this between games one and two. I notice that
Kasparov changed his plan mid-match too, deciding to not play standard opening
theory, and try to take DB out of book as early as possible. Was that fair?
In chess, *anything* goes...
:>: Excuse me for the trivial repetition, but one tends to forget about these
:>: trivial things.
:>: Now my problem could be divided into two questions.
:>: Either it's right what you have written above, then all the pre-match PR
:>: was a complete lie. That would also correspond with your statement that for
:>: you the "match" hasn't proven that DB is the better player ...
:>Have you *ever* seen ad advertisement that didn't add hyperbole, didn't
:>stretch the facts just a bit, didn't embellish a tad to make it more news-
:>worthy? Etc...
: I think you misinterprete my wordings. I would see nothing new if a company
: cheated about the reality (as you explained very well). But the point you
: seem to overlook is the fact that it was Kasparov who distributed the big
: scientifivc question. Now here is the point where I see a clear
: misinformation/ misleading of Kasparov himself. By whom? Try to find an
: answer. Surely Kasparov didn't talk too much with IBM. So I see only one
: group who had cheated him already before the match. And that was, NB the
: second "crime" of formerly honest scientists, the DB team around Hsu and
: Tan.
I don't know Tan. Have no impression of him, nor no personal opinion about
him... But the word "cheat" is *wrong*. Cheat -> doing something contrary
to the rules. This *clearly* wasn't done...
If you would like to propose different "rules" you can do that. But for the
rules used between DB and Kasparov, everything was quite above-board...
: The sun is shining much less right now, I fear.
:>Most adds are more fiction than fact... nothing new there at all...
: Now under the impression of my last paragraphe I think DB team is guilty
: even more ...
Define team. I call this Hsu, Campbell, Hoan, etc... And I disagree. If
you call the team Hsu, Campbel, + the IBM lawyers and marketing types, then
things change. But not the "team" that most everyone would refer to.
:>: If that was a lie, then one were allowed to conclude that the
:>: IBM/DB/teamsters did cheat on the whole world, the two of us included.
:>: Or, Bob, this was NOT a lie. It was really about the almost scientifical
:>: question machine or mankind ...?
:>It was a scientific question: "can the machine beat the best player in the
:>world in a regular match at normal tournament time controls?" Nothing else.
:>Not "can the machine beat the best player in the world in a regular match at
:>normal tournament time controls, after he has had a chance to study lots of
:>games by the machine to discover its weaknesses?" That's a *different*
:>question.
: Yup.
:> An interesting one that I enjoy working with, because this is
:>*exactly* what Bruce (Ferret) and I try to do every day with programs that
:>play automatically on the chess servers. We simply say "come and play us
:>and let's see if you can find a weakness you can exploit repeatedly. If you
:>do, we will fix it." But that's a different issue from DB.
: From IBM, excuse my impertinency.
:>It is a reasonable
:>question. But nobody forced them to ask/answer *that* question. They picked
:>the question, and then answered it. Just because we don't like the specific
:>question they asked, doesn't mean it isn't a legitimate question, because 20
:>years ago no one gave a computer a chance of being able to answer any such
:>question with "yes".
: The main difference for me is the implementation of all the million games.
Don't understand that statement...
:>: But then the IBM/DB/teamsters did also cheat on all of us. And on Kasparov
:>: in particular. (That was my theory in the other thread, where we adjourned
:>: the question because you didn't want to discuss about cheats.)
:>Nope... As I said, they just didn't ask and answer the question *you* wanted
:>them to. You are free to write a chess program, ask that question, then answer
:>it with your program. I am able to do this as well. So are they, *if* they
:>were interested in that question. Because, as I said before, this is certainly
:>a legitimate question. But it isn't the *only* question...
:>: The new situation is full of suspense. How will Bob Hyatt explain that IBM
:>: didn't cheat? Whether before or after the match, sorry, in between?
:>see above. :)
:>: P.S.
:>: I would be nice if you could also join the debate about the position from
:>: Ilias given here in Misc (1) where White should decide between 1.d6 and
:>: 1.Qe7.
:>: My programs want 1.d6. What's about you? Let's concentrate on these
:>: original computerchess questions too. :)
:>I tried to stay out of that. Whittington seems to think that too much
:>"crafty likes this" goes on here anyway...
: Honestly. How many seconds do you become a non-politician in a day? :)
: So you stay away from *my* threads because Chris told something about
: crafty ...
: Don't let you become too oversensitive. It's really a position where you
: programmers could see what your babies could achieve. C'mon.
Not just *your* threads. Chris seems to think that all I can talk about is
how "crafty" does something. It offers me a good way of illustrating something,
but apparently some think it is "too much"... So I react...
Rolf Tueschen
>Your conclusion is wrong.
I'm sorry, but I think that I'm not.
>You are confusing "Hsu, the scientist" with
>"IBM, the publicly held corporation".
How could I do that? When I myself wrote that he (&team) betrayed the
duties of "science"? You gave me more and more insight for my conclusion.
BTW it's one of the most thrilling topics for all modern
scientists/science. Because it's the question of how far or "how" whatever
the i n f l u e n c e of the business/ (the same with) the
"government" will go (be accepted by the scientists) referring to the
standards of science itself.
To be precisesly on the spot. A scientist does *never* cheat. Period.
Exception: A scientist does cheat. Which is very human because a scientist
is a human being *after all* ...
> Hsu works for IBM. He can have
>the purest motives in the world, but the company pays his salary, and
>pays for his hardware development, and pays for the others in the group,
>and pays for the GM's to help them, and pays for the events, and pays for
>the P/R before the event.. So "the company" controls this.
That will go into the record of Dejas, Bob.
Let me still inform you that you're badly wrong.
You could have continued. Even if they paid for Hsu's children (school,
university, flight to Jupiter etc.), his wife and dog (horses), car
grandmother, the whole East-Asia, Hsu's "duties" as scientist would still
remain the same. Period.
(That's the OPPENHEIMER dilemma. Never thought about that?)
>You are making an assumption about "Hsu, the scientist" that is *wrong*.
In the case of Hsu and Co. I was badly wrong, that's true ... ;-(
>IE, suppose that UAB decided that Crafty has become popular enough that
>they'd like to take advantage of this and start selling it. Could they
>do this? probably. would they do it? Almost certainly not ever. But
>*if* they did, does it suddenly become "Bob, the commercial programmer
>when he promised he would never do this?" If so, you are misdirecting
>your anger. If not, then there is no reason to make such comments about
>Hsu either...
I told you before that I already had understood that you were on a proxy
mission for "them" buddies. But this time you screw up the whole job.
You even tried to sac "yourself" in your wish to defend 'em.
I wouldn't say anything the like in your direction only because you had
always confirmed us that you never wanted to go commercial ... Nonsense.
You completely missed the point. It's *not* that Hsu &Co. went commercial.
It was that he betrayed his scientific duties. With going commercial
*alone* that would be touched at all. And to keep the tension at the peak.
It's *not* the point that he allowed a behaviour that was (as you said) in
concordance to the written contract. It was the point --- now what do you
think?
It's the cheat that the "match" as it was held could *not* answer the
question tooted before the match, mankind vs or "or" machine. Simply
because it was *not* the machine but "machine+human factor X". And this
doesn't mean at all the often made tongue in cheek, "that of course the
machine must be made somehow by human beings, no?" No, the point is, that
between rounds the human gang (sorry, I meant "team") tweaked and twisted,
**NOT** to input something like the value of the BB-pair.
I would even accept the like, **IF** they were so stupid prior to the match
forgetting about such a factor. This is NOT the point.
The point is this.
And this is the reason why I speak of cheat both as a matter of science and
of sports.
a) science
==========
I hope you will understand my "newbie" excursion into the jungle of
computerchess science.
It was May 1997 last year in New York, right? We weren't interested in the
the possibilities of the "team" if they imputed this or that as add-ons.
That could have been interesting for the next matches. We wanted to see,
what DB (the machine) could do in May 97.
Under the assumption (a trivial chessic one) that a human player had to
find sort of match strategy, because a human won't play always the same
strategies against player A or B, it's completely absurd (under the
precondition that you want to test the "strength" of your machine against a
*human* player!) if you tweak and twist so that the human player never can
become adapted to his opponent (in a series of only six games). What you
could at best test under these new conditions, is the capability of the
human to adapt to these special conditions. But you might agree with me
that *these* are imputed alone by the human helpers of the team and *not*
at all by the machine itself.
And the factor of the confusing of the human player, excuse me I want to
reserve the notion "cheating" for that one, was *not* the original
question, no?
Another example of a factor among others that could be found.
You said that it was about "winning against the Wch". Fine with me. But
winning on the basis of the strength of the machine (as it was in May 97)
and *not* on the basis of some confusingly twistings here and there ...
subpoint 123a)
You might argue, as you already did BTW, that this "cheating" should be
allowed and no problem for Kasparov, because "he himself also had the
possibility to do the same cheating". It's more a point for the next
paragraphe "sports". But also in science this is relevant.
If you want to test the strength of a machine against a human champ. You
cannot argue that "because the human champ has typically human advantages
of flexibility for example, that it therefore should be allowed that the
machine could also be twisted (by its human operators of course). Because
then you're testing a totally different question. As I already explained it
it's the question if it's possible to cheat/confuse a human player in a
short match of 6 games with the help of in-between-round twistings of the
machine. If the press would have explained *that* question before the
match, I think both Kasparov and the audience worldwide would have
understood the task for the human champ.
So because of this cheat the scientifical reputation of Hsu and his team is
gone into the Hudson (?) river.
b) sports
========
Do I really need to explain whjat the cheat meant in relation to fairness
in sports? I tell you. It means nothing.
In a one-game match?
~~~~~~~~~~~~~~~~~
It's a common knowledge that a total newbie could beat a master in a
competition where the master cannot exploit his superior knowledge.
It's the same in backgammon or pokerstrip. Or in a Las Vegas gambling hall.
You might win a fortune with a 10 dollars entry. But this is called
gambling. If you take that into a serious sport like chess you get what you
imputed. Stupid nonsense.
Some time ago I saw the 17 y. old Michael Chang play the then number one
Lendl in Paris. Chang had succeeded in fooling Lendl with a faked complete
weakness in his legs. Suddenly he tried to go even further. He faked that
his arms too were weak and it was his serve. So he hit the ball from below
like a beginner. And Lendl was so surprised, he had never seen that ball in
all his training sessions, that he was completely out of his concentration.
Exactly his best quality, the ability to concentrate on the spot, was
working against him, because the laughter of the audience was disturbing
him too. An average hobby player would have given Michael Chang a return
that he surely wouldn't returned again.
What did this prove? That Chang was "better" than Lendl? No. Only that you
can exploit even the best strengths of a player if you are allowed to use
all tricks. I for one did never like the sportsman Chang. Because I never
believed him anything. What I still accepted was his strength as a
comparably "little" player among the giants ... Also his will-power and
concentration. But honesty was no longer something I saw in him.
What did IBM/DB do?
~~~~~~~~~~~~~~~~~~
They achieved to confuse Kasparov after the first game. Period.
For me it's still not proven that there wasn't something fishy. I can only
rely on Kasparov's Oxford lecture where he stated that 9 months after the
event he hadn't still the output. Whatever that meant. And I think we could
assume that Kasparov knows about the complications in a endless check
situation. We agreed to give it a rest for a later re-entry.
Also they affected his mental stability. Because they acted unfriendly
against him.
I saw again many tennis matches that were mainly decided through the
interventions of the referee!! Because some players are so sensitive that
they loose control if something goes wrong.
And you should understand that a human champ against a machine could well
be confused by the behaviour of the operators of the machine.
I saw this man (Hsu?) in the press conference after the match. And Bob,
excuse me, I had the impression to see an older clone of Michael Chang. I
simply could see the cheat in his eyes. I must apologize to Chang. Even he
looks more honest that this man on the press conference. Excuse me Bob, I'm
only telling you what I saw. I have it on video and I have always the same
bad feelings when I see it ...
>:>So it became an accounting issue to the company, rather than a computer chess
>:>issue to Hsu and group. Because remember that they had played in *many*
>:>events before becoming married to IBM. They didn't miss ACM or WCCC events,
>:>and they played lots of GM games to earn the second-stage Fredkin prize...
>: Don't ever assume that I misregard the earlier honesty of the scientists.
>: You know the details better than me.
>:>: Instead I could read about the "match" that should answer the question
>:>: who's the best chessplayer? A human being, the PCA Wch for the side of
>:>: "mankind" OR the computer system Deep Blue for the side "machine".
>:>That's simply marketing hype. The match was to answer "can a computer beat
>:>the best player in the world in a regular match at normal time controls."
>:>Nothing about fairness, nothing about sharing game scores, nothing about
>:>telling each other what was going on, just that one simple question...
>: Several new isues. Let me isolate this one for a try. All that might have
>: been the case and ok but then I ask you why the machine had to be changed
>: during the match. I mean a machine wanted to beat the human player.
>: Therefore the machine should have done all that by itself and *not* over
>: night with the help of many GMs and technicians. Don't you see the
>: contradiction?
>No. I do this *every* last day. They felt the bishop pair bonus had been
>set too high. They adjusted this between games one and two. I notice that
>Kasparov changed his plan mid-match too, deciding to not play standard opening
>theory, and try to take DB out of book as early as possible. Was that fair?
For all it's not true.
How would you call the first game? Regular opening choice? :)
>In chess, *anything* goes...
In a gambling hall too ...
>:>: Excuse me for the trivial repetition, but one tends to forget about these
>:>: trivial things.
>:>: Now my problem could be divided into two questions.
>:>: Either it's right what you have written above, then all the pre-match PR
>:>: was a complete lie. That would also correspond with your statement that for
>:>: you the "match" hasn't proven that DB is the better player ...
>:>Have you *ever* seen ad advertisement that didn't add hyperbole, didn't
>:>stretch the facts just a bit, didn't embellish a tad to make it more news-
>:>worthy? Etc...
>: I think you misinterprete my wordings. I would see nothing new if a company
>: cheated about the reality (as you explained very well). But the point you
>: seem to overlook is the fact that it was Kasparov who distributed the big
>: scientifivc question. Now here is the point where I see a clear
>: misinformation/ misleading of Kasparov himself. By whom? Try to find an
>: answer. Surely Kasparov didn't talk too much with IBM. So I see only one
>: group who had cheated him already before the match. And that was, NB the
>: second "crime" of formerly honest scientists, the DB team around Hsu and
>: Tan.
>I don't know Tan. Have no impression of him, nor no personal opinion about
>him... But the word "cheat" is *wrong*. Cheat -> doing something contrary
>to the rules. This *clearly* wasn't done...
Excuse me for my dickhead. It *clearly* was against all standards of
science and fair sports.
>
>If you would like to propose different "rules" you can do that. But for the
>rules used between DB and Kasparov, everything was quite above-board...
Dickhead cannot agree here. :)
>: The sun is shining much less right now, I fear.
>:>Most adds are more fiction than fact... nothing new there at all...
>: Now under the impression of my last paragraphe I think DB team is guilty
>: even more ...
>Define team. I call this Hsu, Campbell, Hoan, etc... And I disagree. If
>you call the team Hsu, Campbel, + the IBM lawyers and marketing types, then
>things change. But not the "team" that most everyone would refer to.
I know it must hurt you. But it's the truth. Sorry. They are guilty.
>: Yup.
Eh, I was reminding you of the out-of-engine situated opening department.
The rest is a "bit" faster but still not on GM level. Understood this way:
regarding the weaknesses. Not the many strengths already.
I must come back on the cheating, excuse me. It's cheating if you twist the
moment you (the human operator) observe that the human opponent of your
machine has "gone on" a particular weakness. Especially in a 6 game match
where you only have three games with the same colour ...
It's as if you said to the human player, go ahead but you won't understand
in the three games against what you're playing when.
I don't know of any sport where such a secrecy of real class is uses
systematically. What we already know is the fact that a newcomer always has
this certain advantage due to his status. So in the beginning of his career
he will win solely based on that factor.
(It should be repeated that a human newbie or expert could change his mind
overnight, *but* he could never change his overall personality. Under
stress he will remain the same in a certain range of course. After some
time the others know how to deal with him either to their advantage or in
some cases to their worst. Because it's a typical fact in sports that some
opponents no matter how strong they are objectively will be too strong for
someone (who could also rank higher in the hierarchy).)
And in chess in the case of Kasparov?
How should he change his chessic style overnight? I mean he knows all about
it. So he can't just say, now I will finally exploit the strength od my
bishop pair. I really can't understand what you meant when you therefore
claimed that kasparov too should be locked up in quarantaine. Does a
machine talk?
>:>see above. :)
>:>: P.S.
It's always good to react or to prove one's flexibility but here you went
too far. Ilias had a nice position for the question of "depth". You should
react on-topic.
But in the other question. It's often said that it's this what seperates
men from women. Sort of stability and abstraction of the fallacies of
personal involvement. I for one would like to hear what you had to argue
against Chris. What I don't appreciate is the sort of sissy-like
demonstrating of "feeling-insulted". I think we could demonstrate the
beauty and elegance of science if we could show to the group that we could
talk in a civilized mode even if we are not the closest friends. But the
level must be there of course.
What you're doing here is similar to Ed Schroders reaction. As if the
critic against you would aim against your personal dignity. I mean you
created crafty, so about what should you talk about? Sure, I also claimed
that you as expert had even *more* to see the different approaches and to
mediate in-between. Perhaps this will happen with your becoming older. :)
We know that many here in rgcc want to hear you about crafty and all. So
please don't take all of them ias hostages because you found that Chris has
made some attacks. Rather try to answer those attacks. As you have done so
many times before.
Let me appeal to you and all. The actual stigmatizing of Chris (who's your
collegue-programmer) is also a sort of cheating for scientists at least.
You shouldn't follow this trend of defamation. And you surely shouldn't be
the leader. But then I agree, you can't be scientist and politician at the
same time. BTW, did you therefore stop talking about crafty?
Robert Hyatt
: Robert Hyatt <hy...@crafty.cis.uab.edu> wrote:
:>In rec.games.chess.computer Rolf Tueschen <TUESCHEN.MEDIZ...@t-online.de> wrote:
:>: Robert Hyatt <hy...@crafty.cis.uab.edu> wrote:
:>Your conclusion is wrong.
Easy. Because "Hsu and team" don't control the "system" there. They
developed DB on IBM's "nickel" and while under contract with IBM, which
almost certainly grants IBM intellectual rights to anything they develop.
This means they have *no* control over how/when it is used, higher-up
"bean counters" decide how the DB project can be "spun" to generate the
highest possible revenue.
That's the point. Hsu can't agree to another DB match. It's an IBM
decision. Hsu can't agree to play DB in FIDE tournaments, it is an
IBM decision. IF, for some reason, IBM becomes convinced that having DB
in FIDE events will further their corporate "image" then this might
happen. At present they are riding the crest of a huge PR wave created
by the match... and they won't do anything until they begin to slide
down the backside of that wave and decide that it is to their advantage
to roll the beast back out again because everyone has "forgotten" about
it.
: To be precisesly on the spot. A scientist does *never* cheat. Period.
: Exception: A scientist does cheat. Which is very human because a scientist
: is a human being *after all* ...
They didn't "cheat". From Webster's: Cheat, verb, to break a rule. They
did *not* do this. Or, to frame it differently, you provide the "rule" that
they broke in the match.
:> Hsu works for IBM. He can have
:>the purest motives in the world, but the company pays his salary, and
:>pays for his hardware development, and pays for the others in the group,
:>and pays for the GM's to help them, and pays for the events, and pays for
:>the P/R before the event.. So "the company" controls this.
: That will go into the record of Dejas, Bob.
: Let me still inform you that you're badly wrong.
: You could have continued. Even if they paid for Hsu's children (school,
: university, flight to Jupiter etc.), his wife and dog (horses), car
: grandmother, the whole East-Asia, Hsu's "duties" as scientist would still
: remain the same. Period.
You simply don't know what you are talking about. When you sign on at IBM,
you sign a "contract". They retain all intellectual property rights for
anything *you* create while working for them. You can leave, but if you try
to take the technology, or the ideas, and use them, you end up broke, and
alone, because of the contract. So Hsu doesn't have the "power" you'd like
to attribute to him. He has definitely "reported" on what he did, which
is the purpose of any good scientist. But he can't use the thing in any
way he sees fit, when it is so valuable to the company. Since you've never
worked for such a company, you don't understand just how much control they
have. But if IBM so chooses, DB will *never* be seen again, and there's
nothing Hsu can do about it...
: (That's the OPPENHEIMER dilemma. Never thought about that?)
I met the man at Los Alamos, in fact. And I don't see any "dilemma" (or
delimma) either. He's published what he did. But Los Alamos would not
have let him build a bomb on their nickel, and then take it off and use
it wherever he wanted. Had he left, he had a contract that was *very*
specific. It is on exhibit at the museum and the Los Alamos lab, in
fact... And I can guarantee you that had he left Los Alamos before the
project was finished, he'd have been followed for the rest of his life
to be certain he didn't reveal anything he had promised not to reveal.
:>You are making an assumption about "Hsu, the scientist" that is *wrong*.
: In the case of Hsu and Co. I was badly wrong, that's true ... ;-(
But not for the reason you try to imply.
:>IE, suppose that UAB decided that Crafty has become popular enough that
:>they'd like to take advantage of this and start selling it. Could they
:>do this? probably. would they do it? Almost certainly not ever. But
:>*if* they did, does it suddenly become "Bob, the commercial programmer
:>when he promised he would never do this?" If so, you are misdirecting
:>your anger. If not, then there is no reason to make such comments about
:>Hsu either...
: I told you before that I already had understood that you were on a proxy
: mission for "them" buddies. But this time you screw up the whole job.
: You even tried to sac "yourself" in your wish to defend 'em.
I do *not* "sac" myself. I point out the very tight straightjacket they
are in, working for a very large company with a huge legal department..
: I wouldn't say anything the like in your direction only because you had
: always confirmed us that you never wanted to go commercial ... Nonsense.
And I've sit across the table from Hsu, or in a relaxed setting with other
programmers and Hsu, or talked with him via email... and he is every bit
the same as the rest of us. He is *less* secretive than the commercial
programmers, because he has revealed every detail about the hardware,
about the software, and so forth. He just can't take the machine and compete
wherever he wants, because the IBM guys discovered that there was a lot of
money in this project, via P/R benefits...
: You completely missed the point. It's *not* that Hsu &Co. went commercial.
: It was that he betrayed his scientific duties. With going commercial
: *alone* that would be touched at all. And to keep the tension at the peak.
: It's *not* the point that he allowed a behaviour that was (as you said) in
: concordance to the written contract. It was the point --- now what do you
: think?
His scientific duty is to develop technology and provide the details of that
technology to others. He has done this. I know how his hardware operates.
So do many others. The "product" which is owned by IBM is something else.
And they control that to use it as they see fit to maximize company profits.
But he has *definitely* published the internal details of DB.
: It's the cheat that the "match" as it was held could *not* answer the
: question tooted before the match, mankind vs or "or" machine. Simply
: because it was *not* the machine but "machine+human factor X". And this
: doesn't mean at all the often made tongue in cheek, "that of course the
: machine must be made somehow by human beings, no?" No, the point is, that
: between rounds the human gang (sorry, I meant "team") tweaked and twisted,
: **NOT** to input something like the value of the BB-pair.
That's stupid. Of course it is man vs machine+human factor. Who wrote
the software and built the hardware? A room full of monkeys? Then it would
be man vs machine+monkey_factor. There's *no* way that the "human factor"
won't be involved when humans write the software and design and build the
hardware, and there's nothing wrong with humans modifying the software between
rounds, since that has been an allowable rule in computer chess since the
first tournament in 1970.
: I would even accept the like, **IF** they were so stupid prior to the match
: forgetting about such a factor. This is NOT the point.
: The point is this.
: And this is the reason why I speak of cheat both as a matter of science and
: of sports.
: a) science
: ==========
: I hope you will understand my "newbie" excursion into the jungle of
: computerchess science.
: It was May 1997 last year in New York, right? We weren't interested in the
: the possibilities of the "team" if they imputed this or that as add-ons.
: That could have been interesting for the next matches. We wanted to see,
: what DB (the machine) could do in May 97.
Yep...
: Under the assumption (a trivial chessic one) that a human player had to
: find sort of match strategy, because a human won't play always the same
: strategies against player A or B, it's completely absurd (under the
: precondition that you want to test the "strength" of your machine against a
: *human* player!) if you tweak and twist so that the human player never can
: become adapted to his opponent (in a series of only six games). What you
: could at best test under these new conditions, is the capability of the
: human to adapt to these special conditions. But you might agree with me
: that *these* are imputed alone by the human helpers of the team and *not*
: at all by the machine itself.
Not insane at all. I do this every day. Bruce does this every day. If,
after game one, they noticed a bug or problem, nothing in the rules nor
morals of the match said "no changee". I'd expect that if game 6 were
played again, do you not think Kasparov would play the correct move order
this time? Isn't that the same sort of "cheating" that you are disdaining?
Of course..
: And the factor of the confusing of the human player, excuse me I want to
: reserve the notion "cheating" for that one, was *not* the original
: question, no?
Had *nothing* to do with the original "question" so it was, and still is
an irrelevant issue...
: Another example of a factor among others that could be found.
: You said that it was about "winning against the Wch". Fine with me. But
: winning on the basis of the strength of the machine (as it was in May 97)
: and *not* on the basis of some confusingly twistings here and there ...
That is *your* definition of the question. It is *not* the definition of
the question that has been posed for so many years.
: subpoint 123a)
: You might argue, as you already did BTW, that this "cheating" should be
: allowed and no problem for Kasparov, because "he himself also had the
: possibility to do the same cheating". It's more a point for the next
: paragraphe "sports". But also in science this is relevant.
: If you want to test the strength of a machine against a human champ. You
: cannot argue that "because the human champ has typically human advantages
: of flexibility for example, that it therefore should be allowed that the
: machine could also be twisted (by its human operators of course). Because
: then you're testing a totally different question. As I already explained it
: it's the question if it's possible to cheat/confuse a human player in a
: short match of 6 games with the help of in-between-round twistings of the
: machine. If the press would have explained *that* question before the
: match, I think both Kasparov and the audience worldwide would have
: understood the task for the human champ.
This is stupid reasoning. Suppose the program hits something "new" in the
code and crashes. It would lose that game on time, as the rules forbid
changes *during* a game. But should it have to play the match with that
bug? And lose game after game? Should Kasparov have to play the game 6
blunder every time or can he run to the opening books and look to see what
he did wrong after the game?
Computers don't learn like humans. Computers don't "think" like humans.
Computers don't correct their own bugs like humans. Get the gist here:
*computers* *are* *not* *like* *humans* *at* *all*. So stop trying to cast
the match as man vs machine-in-a-box. That wasn't the question. That wasn't
the way the match was planned. That wasn't in the rules of the match. If you
don't like this, go set up your own match, put up your own money, and make any
rules you like. But for *that* match this wasn't in the rules, and it makes
no sense to continue the "banter" about it.
: So because of this cheat the scientifical reputation of Hsu and his team is
: gone into the Hudson (?) river.
Not at all. Again, that is *your* opinion. Held by very few. Most are
quite impressed with what they accomplished..
: b) sports
: ========
: Do I really need to explain whjat the cheat meant in relation to fairness
: in sports? I tell you. It means nothing.
: In a one-game match?
: ~~~~~~~~~~~~~~~~~
: It's a common knowledge that a total newbie could beat a master in a
: competition where the master cannot exploit his superior knowledge.
: It's the same in backgammon or pokerstrip. Or in a Las Vegas gambling hall.
: You might win a fortune with a 10 dollars entry. But this is called
: gambling. If you take that into a serious sport like chess you get what you
: imputed. Stupid nonsense.
A newbie *can* beat a master. But the probability of this happening is
so low as to be called *zero*.
: Some time ago I saw the 17 y. old Michael Chang play the then number one
: Lendl in Paris. Chang had succeeded in fooling Lendl with a faked complete
: weakness in his legs. Suddenly he tried to go even further. He faked that
: his arms too were weak and it was his serve. So he hit the ball from below
: like a beginner. And Lendl was so surprised, he had never seen that ball in
: all his training sessions, that he was completely out of his concentration.
: Exactly his best quality, the ability to concentrate on the spot, was
: working against him, because the laughter of the audience was disturbing
: him too. An average hobby player would have given Michael Chang a return
: that he surely wouldn't returned again.
: What did this prove? That Chang was "better" than Lendl? No. Only that you
: can exploit even the best strengths of a player if you are allowed to use
: all tricks. I for one did never like the sportsman Chang. Because I never
: believed him anything. What I still accepted was his strength as a
: comparably "little" player among the giants ... Also his will-power and
: concentration. But honesty was no longer something I saw in him.
And the point would be? This wasn't a "chang" issue, it was idiocy on the
part of Lendl...
: What did IBM/DB do?
: ~~~~~~~~~~~~~~~~~~
: They achieved to confuse Kasparov after the first game. Period.
: For me it's still not proven that there wasn't something fishy. I can only
: rely on Kasparov's Oxford lecture where he stated that 9 months after the
: event he hadn't still the output. Whatever that meant. And I think we could
: assume that Kasparov knows about the complications in a endless check
: situation. We agreed to give it a rest for a later re-entry.
This is wrong. He clearly stated that he did get the output for the three
moves he questioned. He did *not* get the output for all the games. He
should *never* get this, IMHO. Yasser wrote an article in the JICCA that
discusses this at length. And he saw the printouts according to the article.
: Also they affected his mental stability. Because they acted unfriendly
: against him.
*tough* *stuff*. He accused *them* of cheating. And he then says they
are unfriendly? Har-de-har-har...
: I saw again many tennis matches that were mainly decided through the
: interventions of the referee!! Because some players are so sensitive that
: they loose control if something goes wrong.
: And you should understand that a human champ against a machine could well
: be confused by the behaviour of the operators of the machine.
Again, tough. He's the best in the world.
: I saw this man (Hsu?) in the press conference after the match. And Bob,
: excuse me, I had the impression to see an older clone of Michael Chang. I
: simply could see the cheat in his eyes. I must apologize to Chang. Even he
: looks more honest that this man on the press conference. Excuse me Bob, I'm
: only telling you what I saw. I have it on video and I have always the same
: bad feelings when I see it ...
Bullshit... If you want to resort to personal insults, I'm going to terminate
this discussion from my end. I know Hsu. Your remarks are insulting to *all*
of us that know him. Stop, or talk to yourself from here on out...
:>No. I do this *every* last day. They felt the bishop pair bonus had been
:>set too high. They adjusted this between games one and two. I notice that
:>Kasparov changed his plan mid-match too, deciding to not play standard opening
:>theory, and try to take DB out of book as early as possible. Was that fair?
: For all it's not true.
: How would you call the first game? Regular opening choice? :)
yes...
:>I don't know Tan. Have no impression of him, nor no personal opinion about
:>him... But the word "cheat" is *wrong*. Cheat -> doing something contrary
:>to the rules. This *clearly* wasn't done...
: Excuse me for my dickhead. It *clearly* was against all standards of
: science and fair sports.
"standards" has nothing to do with it. "cheating" is about *rules*...
:>
:>If you would like to propose different "rules" you can do that. But for the
:>rules used between DB and Kasparov, everything was quite above-board...
: Dickhead cannot agree here. :)
You didn't have to agree, only Kasparov. He agreed to the rules...
:>Define team. I call this Hsu, Campbell, Hoan, etc... And I disagree. If
:>you call the team Hsu, Campbel, + the IBM lawyers and marketing types, then
:>things change. But not the "team" that most everyone would refer to.
: I know it must hurt you. But it's the truth. Sorry. They are guilty.
Only in *your* mind, which doesn't hold a lot of weight with *me* nor
anyone else working in the field of computer science...
:>:>question. But nobody forced them to ask/answer *that* question. They picked
:>:>the question, and then answered it. Just because we don't like the specific
:>:>question they asked, doesn't mean it isn't a legitimate question, because 20
:>:>years ago no one gave a computer a chance of being able to answer any such
:>:>question with "yes".
:>: The main difference for me is the implementation of all the million games.
:>Don't understand that statement...
: Eh, I was reminding you of the out-of-engine situated opening department.
: The rest is a "bit" faster but still not on GM level. Understood this way:
: regarding the weaknesses. Not the many strengths already.
: I must come back on the cheating, excuse me. It's cheating if you twist the
: moment you (the human operator) observe that the human opponent of your
: machine has "gone on" a particular weakness. Especially in a 6 game match
: where you only have three games with the same colour ...
Again, this is *not* cheating unless it is written into the rules of the
game, *or* the rules of the match. It was *not*.
: It's as if you said to the human player, go ahead but you won't understand
: in the three games against what you're playing when.
: I don't know of any sport where such a secrecy of real class is uses
: systematically. What we already know is the fact that a newcomer always has
: this certain advantage due to his status. So in the beginning of his career
: he will win solely based on that factor.
: (It should be repeated that a human newbie or expert could change his mind
: overnight, *but* he could never change his overall personality. Under
: stress he will remain the same in a certain range of course. After some
: time the others know how to deal with him either to their advantage or in
: some cases to their worst. Because it's a typical fact in sports that some
: opponents no matter how strong they are objectively will be too strong for
: someone (who could also rank higher in the hierarchy).)
: And in chess in the case of Kasparov?
: How should he change his chessic style overnight? I mean he knows all about
: it. So he can't just say, now I will finally exploit the strength od my
: bishop pair. I really can't understand what you meant when you therefore
: claimed that kasparov too should be locked up in quarantaine. Does a
: machine talk?
Deep Blue didn't change its "style" overnight. They adjusted the value for
the bishop pair. To kasparov, that affected how it behaved. I have adjusted
that particular score a few dozen times myself and it does change things. But
it *doesn't* change the deep tactical accuracy of the program, nor its overall
"personality". Just that Kasparov decided to play on the bishop pair theme,
and it suddenly behaved differently. Just like Kasparov decided midway thru
the match to play oddball openings to get it out of book quickly. That sounds
like cheating according to your very loose and informal definition...
: It's always good to react or to prove one's flexibility but here you went
: too far. Ilias had a nice position for the question of "depth". You should
: react on-topic.
: But in the other question. It's often said that it's this what seperates
: men from women. Sort of stability and abstraction of the fallacies of
: personal involvement. I for one would like to hear what you had to argue
: against Chris. What I don't appreciate is the sort of sissy-like
: demonstrating of "feeling-insulted". I think we could demonstrate the
: beauty and elegance of science if we could show to the group that we could
: talk in a civilized mode even if we are not the closest friends. But the
: level must be there of course.
: What you're doing here is similar to Ed Schroders reaction. As if the
: critic against you would aim against your personal dignity. I mean you
: created crafty, so about what should you talk about? Sure, I also claimed
: that you as expert had even *more* to see the different approaches and to
: mediate in-between. Perhaps this will happen with your becoming older. :)
Not the same thing at all. I just choose to stay away from the "crafty
does it like this" a little more..
: We know that many here in rgcc want to hear you about crafty and all. So
: please don't take all of them ias hostages because you found that Chris has
: made some attacks. Rather try to answer those attacks. As you have done so
: many times before.
: Let me appeal to you and all. The actual stigmatizing of Chris (who's your
: collegue-programmer) is also a sort of cheating for scientists at least.
: You shouldn't follow this trend of defamation. And you surely shouldn't be
: the leader. But then I agree, you can't be scientist and politician at the
: same time. BTW, did you therefore stop talking about crafty?
I have responded to technical questions about it. I have ignored many
"what do other programs play here?" posts...
Phil Innes
each other's. I only have something small to say, and do not want to be
a distraction to your conversation - so I make a comment at the end of
this segment of your conversation:
>Your conclusion is wrong.
(Phil>) My wife's family are from Princeton, and they have
many an Einstein anecdote - some of which are rather funny -
but, more soberly, is there a clearer example in this century
of a scientist torn between his work and the social
implications of it? Its understanding and deployment?
I know Bob/Rolf are discussing something slightly different,
to wit: "to what degree is this science from IBM still
'honest,'" but I only mean to say that the situation is far
from unique.
In a technical way Einstein even had the gravest reservations
about other scientists of his time who fully engaged quantum
theory. Until the end he thought that this was a false path, a
series of 'commercial' aggregates, and its disciplines "knew
not what they did."
Chris Whittington
Phil Innes wrote in message <36054E32...@sover.net>...
>Phil Innes writes; please (Bob/Rolf) ignore my post in preference to
>each other's. I only have something small to say, and do not want to be
>a distraction to your conversation - so I make a comment at the end of
>this segment of your conversation:
>
> many an Einstein anecdote - some of which are rather funny -
> but, more soberly, is there a clearer example in this century
> of a scientist torn between his work and the social
> implications of it? Its understanding and deployment?
>
> I know Bob/Rolf are discussing something slightly different,
> to wit: "to what degree is this science from IBM still
> 'honest,'" but I only mean to say that the situation is far
> from unique.
>
> In a technical way Einstein even had the gravest reservations
> about other scientists of his time who fully engaged quantum
> theory. Until the end he thought that this was a false path, a
> series of 'commercial' aggregates, and its disciplines "knew
> not what they did."
>
Is the dilemma entirely reasonable in this circumstance ?
Deep Blue v. Kasparov for Champ of the World (chess compartment) is not
exactly dropping atom bombs. Or is it ?
Isn't there a point at which a scientist makes a choice (career choice) to
suspend the dictates of 'pure' science; and start making products, or PR or
whatever ?
Is this not acceptable ?
And can he not return to 'science' aftwerwards ?
Chris Whittington
Chris Whittington
Rolf Tueschen
>Phil Innes wrote in message <36054E32...@sover.net>...
I wrote:
-->>> (That's the OPPENHEIMER dilemma. Never thought about that?)
>>
>> (Phil>) My wife's family are from Princeton, and they have
>> many an Einstein anecdote - some of which are rather funny -
>> but, more soberly, is there a clearer example in this century
>> of a scientist torn between his work and the social
>> implications of it? Its understanding and deployment?
>>
>> I know Bob/Rolf are discussing something slightly different,
>> to wit: "to what degree is this science from IBM still
>> 'honest,'" but I only mean to say that the situation is far
>> from unique.
>>
>> In a technical way Einstein even had the gravest reservations
>> about other scientists of his time who fully engaged quantum
>> theory. Until the end he thought that this was a false path, a
>> series of 'commercial' aggregates, and its disciplines "knew
>> not what they did."
>>
>Is the dilemma entirely reasonable in this circumstance ?
>Deep Blue v. Kasparov for Champ of the World (chess compartment) is not
>exactly dropping atom bombs. Or is it ?
You're right. Absolutely.
>Isn't there a point at which a scientist makes a choice (career choice) to
>suspend the dictates of 'pure' science; and start making products, or PR or
>whatever ?
>Is this not acceptable ?
Absolutely.
But the question between Bob and me was if (as Phil also understood us)
they then could stil remain "scientists" during their commercial period.
Bob seems to thinkyes, but tells me at the same time that IBM surely had
the rights of telling the scientists what to do. My very trivial conclusion
was that *then* (following the IBM agenda) the scientists lost their
reputation. Another question is also if the match between Kasparob and DB
*then* could be about science? Bob said "of course not". But then I accused
them of cheating because they had announced the match with exactly the
scientific relevance.
>And can he not return to 'science' aftwerwards ?
Absolutely.
>Chris Whittington
>Chris Whittington
Rolf Tueschen
>Phil Innes writes; please (Bob/Rolf) ignore my post in preference to
>each other's. I only have something small to say, and do not want to be
>a distraction to your conversation - so I make a comment at the end of
>this segment of your conversation:
>>: My personal comment would be this. For me, personally, the former
>>Your conclusion is wrong.
> (Phil>) My wife's family are from Princeton, and they have
> many an Einstein anecdote - some of which are rather funny -
> but, more soberly, is there a clearer example in this century
> of a scientist torn between his work and the social
> implications of it? Its understanding and deployment?
> I know Bob/Rolf are discussing something slightly different,
> to wit: "to what degree is this science from IBM still
> 'honest,'" but I only mean to say that the situation is far
> from unique.
> In a technical way Einstein even had the gravest reservations
> about other scientists of his time who fully engaged quantum
> theory. Until the end he thought that this was a false path, a
> series of 'commercial' aggregates, and its disciplines "knew
> not what they did."
That's not the whole story. As we know Einstein was probably the one who
convinced Roosevelt (was it I think?) to finally built the bomb against
Germany. But later when it should be used against Japan Einstein saw
himself in a great dilemma. And he did oppose that the bomb should be
thrown ...
I'm still pride that I've chosen that topic for the final examins in
gymnasium.
After all Einstein saw that his engagement was misused. Suddenly he had no
army. Like the Pope. And the politicians did what they did.
The other question if the scientists were aware of what they were doing,
the most interesting question, was well described in your little article.
Thanks Phil.
Chris Whittington
Rolf Tueschen wrote in message <6u3jl1$2hg$1...@news02.btx.dtag.de>...
>"Chris Whittington" <chr...@cpsoft.demon.co.uk> wrote:
>
>>Phil Innes wrote in message <36054E32...@sover.net>...
>
>I wrote:
>
>
>
>>>
>>> (Phil>) My wife's family are from Princeton, and they have
>>> many an Einstein anecdote - some of which are rather funny -
>>> but, more soberly, is there a clearer example in this century
>>> of a scientist torn between his work and the social
>>> implications of it? Its understanding and deployment?
>>>
>>> I know Bob/Rolf are discussing something slightly different,
>>> to wit: "to what degree is this science from IBM still
>>> 'honest,'" but I only mean to say that the situation is far
>>> from unique.
>>>
>>> In a technical way Einstein even had the gravest reservations
>>> about other scientists of his time who fully engaged quantum
>>> theory. Until the end he thought that this was a false path, a
>>> series of 'commercial' aggregates, and its disciplines "knew
>>> not what they did."
>>>
>
>
>>Deep Blue v. Kasparov for Champ of the World (chess compartment) is not
>>exactly dropping atom bombs. Or is it ?
>
>You're right. Absolutely.
>
>>Isn't there a point at which a scientist makes a choice (career choice) to
>>suspend the dictates of 'pure' science; and start making products, or PR
or
>>whatever ?
>
>>Is this not acceptable ?
>
>Absolutely.
>
>But the question between Bob and me was if (as Phil also understood us)
>they then could stil remain "scientists" during their commercial period.
>Bob seems to thinkyes, but tells me at the same time that IBM surely had
>the rights of telling the scientists what to do. My very trivial conclusion
>was that *then* (following the IBM agenda) the scientists lost their
>reputation.
IBM could tell them to do science. Then they'ld still be scientists.
Or IBM could tell them to *beat* Kasparov. They must then scientifically
engineer their machine, accoring to their principles. They must use GM
knowledge, also according to their scintific principles.
But once they are beyond that 'science' then other 'beating' activities are
that. Other.
If they put the package together and call it 'science' then .....
Chris Whittington
Rolf Tueschen
>IBM could tell them to do science. Then they'ld still be scientists.
Yes.
>Or IBM could tell them to *beat* Kasparov. They must then scientifically
>engineer their machine, accoring to their principles. They must use GM
>knowledge, also according to their scintific principles.
You mean IBMsters or the former DBteam?
>But once they are beyond that 'science' then other 'beating' activities are
>that. Other.
I think Bob exactly explained what happend -- this way.
>If they put the package together and call it 'science' then .....
Yes, they called or let it be called "science".
I never read a different interpretation in the pre-match ballyhoo.
And exactly on that grounds Kasparov was 100% sure that he would beat DB
similarily easy as in 96. I remember a report, in the German SPIEGEL I
think, where he agreed that beyond all difficulties DB still had serious
weaknesses. And a whole bunch of them.
He never gave the impression that he had to be careful because IBM(?) or
the former DBteam should treaet him differently to the way they treated him
in 96. I'm not sure, but *then* they gave him a post-game data sample so
that he could try to see what was going on in reality.
I already mentioned to Bob, that for me Kasparov didn't really need the
output to be a better player but it was surely for him a certain
information about the thought process of the machine. Ok, one could argue
with Bob, that this was an unfair advantage for the human player who
himself didn't explain his own thought process. But I think we must be
fair.
To whom Kasparov should explain if he wanted?? To the machine or to
Benjamin? So that *he* then could inform the operators where they should
twist and tweak between rounds??
I for one am convinced that the former deal/handling with a friendly
treatment of the human opponent who was 90% guarantee of the success of the
whole event was justified. *Not* because it gave Kasparov an unfair
advantage. No, it was simply the typical way scientists treated their
experimental clients all over the world.
In New York the cheated on him in the way social scientists often behaved
when they told the client a completely false description of the
experimental setting. They do that to obtain important data. If they told
the client what was really going on most of them wouldn't have cooperated.
But NB the important difference to the event in 97. Here Kasparov was
mistreated in all public. He was humiliated like a fool who was shown too
"emotionally affectable" or choose your own notions.
Chris, believe me one thing. I would never have entered such a debate if it
was "only" about my personal bad feelings. The way we discuss a football
match where we always were the better trainer and also the better
goalgetter. I did NOT enter the debate with such a motivation. No. I tell
you why I did it. Because I almost immediately saw on that press-conference
that this official from the former DBteam had lost all respect for
Kasparov. I could feel it. And not enough. I felt that Kasparov knew it
too. That all together gave the crazy mixture of that event. Kasparov was
treated, let me give you an example from my own professional experience,
like the average patient with that illness, when the professor enters his
room with the tail of subordinates in his follow-up, and the chief god in
white speaks the golden words "How are *we* doing today?" And if you try to
ask a question. You *know* that you don't get a complete/ true/ sufficient/
understandable answer. But you play the game because you want to be cheated
because you fear the moment the medical expert tells you the truth. As long
as he didn't do that you can still hope that it will not be the case ...
That's anxiety, but all of those who got the right information will tell
you that it wasn't bad after all. and that they now could plan for their
future. How short it would ever be ...
In the other case, if the answer is already there but not yet told, you
simply lose precious lifetime! Stolen by your doctor.
The interesting topic for me is also that Bob thought that a sportive fair
Kasparov only had to admit that he screwed it and then all would have been
fine.
This however insinuates that Kasparov could have seperated his still sane
and oh so strong inner self plus chessic genius from the emotional fallacy
he was sucked into. From the outside I saw all that. And I support Kasparov
that he still doesn't want to admit that he failed himself. Also for me it
was a crazy situation with unbelievably high tension for Garry. I think
from the outside we can only try to understand the amount of that tension.
And the humiliating aspect is because he knows it, we know it, that DB is
not at all that strong GM. In such a short show with all the human
interferences of course this must have been like a burning hell for
Kasparov. Therefore he spoke of the only 10 (?) others who could compete
with the machine. But I know that Garry knows that this isn't true. And he
told us different before the match. But then he was unaware of the
incredible confusion /tension he should meet because of the human factor on
the other side.
I think it was Bob who made the point that in his view Kasparov simply
didn't prepare good enough for the match. As a psychologist I can only
agree. And the always friendly FRITZ surely gave him a wrong feeling.
Perhaps the next time I will be invited to put some training sessions
together. There're plenty of ideas how you could simulate that crazy ghoast
feeling. How you could handle your own paranoia. All that could be done.
But not by Frederic Friedel. It's more a task for someone working hard in
the background.
What do you think? :)
Phil Innes
final examins in
gymnasium.
After all Einstein saw that his engagement was misused. Suddenly he had
no
army. Like the Pope. And the politicians did what they did.
The other question if the scientists were aware of what they were doing,
the most interesting question, was well described in your little
article.
Thanks Phil.
Well, thank you Rolf. But after I wrote it I wondered if it
was fair, or just something clever to say in the news cafe.
Some months ago I mentioned an aftermath of this dilemma which
formed more in England than anywhere else, after the world war
and following the bomb, and was the "Two Cultures" debate
between the mandarin C.P. Snow and the literary F.R.Leavis. It
may seem obscure to mention this here - as if we stray very
far from our subject of computer chess, yet this argument
which Leavis won in debate, was a lost issue, and government
proceeded after Snow, which still influences us greatly here
at the end of the century.
The result was a splitting away of science and technology from
the "shackles" of the humanities (which Leavis viewed with
great apprehension.)
In the past few months I have seen an increasing number of
articles which ask for an academic reunion of many branches of
study - and even a call for a "unified field theorum" not just
from the sciences, but like some grand integer from the
Renaissance.
In many fields we now see research being undertaken by hybrid
departments, and I have mentioned (too much) the U. Sussex
neuroscience and computing department, currently collaborating
with scads of biologists on patterning and dissemination
issues.
Results from these projects are still scant, and there may not
be much more to say for some time, however more results and
breakthroughs are coming from these strange zyzygious
couplings than from established 'pure' disciplines.
Of course, this only interests those of us who are interested
in huge amounts of funding - which is apparently forthcoming -
and the above-mentioned Millenial Renaissance. Phil
Phil Innes
view Kasparov simply
didn't prepare good enough for the match. As a psychologist I can only
agree. And the always friendly FRITZ surely gave him a wrong feeling.
Perhaps the next time I will be invited to put some training sessions
together. There're plenty of ideas how you could simulate that crazy
ghoast
feeling. How you could handle your own paranoia. All that could be done.
But not by Frederic Friedel. It's more a task for someone working hard
in
the background.
What do you think? :)
This is slightly funnier than it may appear. What Rolf doesn't
know (I haven't mentioned it to him privately) is the
possibility of Kasparov playing another computer match.
Details are still being knocked around and being considered,
but this part is not secret Rolf, both Bob and Chris have said
that they are game (means willing to back their respective
programs, and perhaps running on some super-cooled hardware)
so in your question you ask the devils themselves!
Crafty and CSTal, versus K (and Rolf.)
But I would like to join it to yet another issue that has
emerged, but this is getting a little too complicated to
manage theoretical ideas versus actual (or potential) events
in a public group discussion, and its late at night!
Anyway - elsewhere Rolf, the difficulty is discussed of
implementing the special knowledge of first-tier GM knowledge
into chess programs, and the need for some intermediary to do
this. A sub-set of questions has arisen about the (somewhat
hilarious! I'm sorry, don't mean to insult anyone, my own
knowledge is meagre enough) views that many programmers have
on what this intelligence is that plays chess.
See. I have compounded too many ideas and it needs a moment or
three to make sense of it. Phil.
ilias kastanas 08-14-90
@Rolf Tueschen wrote:I think it was Bob who made the point that in his
@view Kasparov simply
@didn't prepare good enough for the match. As a psychologist I can only
@agree. And the always friendly FRITZ surely gave him a wrong feeling.
@Perhaps the next time I will be invited to put some training sessions
@together. There're plenty of ideas how you could simulate that crazy
@ghoast
@feeling. How you could handle your own paranoia. All that could be done.
The ghostly feeling is largely due to the absence of a standard
staple in human chess: moves drastically affecting mood and attitude...
sensing the opponent is comfortable here, uneasy there... fears this,
hopes for that... Such factors are _used_; they are integrated into a
human's decisions. They are part of the play. And a human needs to unlearn
them when facing computers.
An amusing instance of "humanity": Johner - Stallda, 1938, saw
Kh2, Qc4, Rf3, Nh5, Pa2, b3, c3, e4, f2, g2, h3; Kh6, Qc1, Rd1, Ne7, Pb6,
c7, e5, f6, g5.
White has a winning position... e.g. 1. R:f6+. But he played
1. Qf7?
and Black replied
1... Rh1+ 2. Kg3, R:h3+!
White was horrified: both 3. K:R and 3. g:R get him mated! He
went into sacrificial shock, and resigned on the spot.
Now, a computer would surely have noticed that 3. Kg4! was avai-
lable... and Black has only a draw.
A "higher-level" example: Smyslov - Petrosian, Zuerich 1953
Kg4, Qd5, Be7, Pc3, d6, f3, g2, h2; Kh7, Qe1, Nd3, Pa7, b6, c4, g7, h6.
Smyslov had to decide how to convert his big advantage to an actual
win. He chose
46. d7
A crusher, to be sure; 46... Q:e7 47. d8=Q is hopeless, and
on 46... Ne5+ 47. Kh3 threatens Qe4+ too. Black can only resign.
But young Petrosian didn't resign:
46... Qe5!!
Miraculous! 47. Q:e5 loses d7; 47. d8=Q?? gets mated; on 47. Qd4,
Q:h2 Black has perpetual check! Smyslov, amazed, conceded a draw. Every-
one was struck and impressed.
Most of that time's strongest players were competing at Zuerich. All
of them, as well as all annotators and the entire chess world were entranced
by ...Qe5. It was _months_ later that an untitled player pointed out the
truth:
47. Qd6!! wins!
Now h2 is protected, and there is no mate either, only a couple
of checks! But everybody had been too "hypnotized" to notice!
Ilias
Rolf Tueschen
>:>Your conclusion is wrong.
A little correction if you don't mind.
It's not about what you descibed.
I was talking about the influence on their work as such. That hasn't
changed in my eyes whether they did it in a universitary lab or "in" IBM. I
will give an example later in this post.
>That's the point. Hsu can't agree to another DB match. It's an IBM
>decision. Hsu can't agree to play DB in FIDE tournaments, it is an
>IBM decision. IF, for some reason, IBM becomes convinced that having DB
>in FIDE events will further their corporate "image" then this might
>happen.
Ok, but if they participate does Hsu *then* have the right to follow his
scientific standards (moral)?
> At present they are riding the crest of a huge PR wave created
>by the match... and they won't do anything until they begin to slide
>down the backside of that wave and decide that it is to their advantage
>to roll the beast back out again because everyone has "forgotten" about
>it.
I see. So we are just right in the middle of preventing exactly that ... :)
>: To be precisesly on the spot. A scientist does *never* cheat. Period.
>: Exception: A scientist does cheat. Which is very human because a scientist
>: is a human being *after all* ...
>They didn't "cheat". From Webster's: Cheat, verb, to break a rule. They
>did *not* do this. Or, to frame it differently, you provide the "rule" that
>they broke in the match.
Good to know that you didn't yet inform Kenneth Starr of my crimes.
But I want to answer your question. Not as an expert of justice. Also I
wouldn't say that scientific standards could be made juridically relevant.
So, no new court trials, Bob ...
Another side remark. We can say that we haven't understood the other but we
shouldn't say that the other doesn't know what he's talking about. We
shouldn't make "CCC of Nov97" to our reference.
Here#s my simple answer. The DBteam left their former scientific interests.
What they then tested was how far Kasparov could be affected by unfriendly
non-cooperative behaviour of the *operating team*. NB that the team didn't
play the games. Just in case you completely mixed it up. That's as if the
captain of a certain team went crazy in a team match. Of course the players
get affected ...
>:> Hsu works for IBM. He can have
>:>the purest motives in the world, but the company pays his salary, and
>:>pays for his hardware development, and pays for the others in the group,
>:>and pays for the GM's to help them, and pays for the events, and pays for
>:>the P/R before the event.. So "the company" controls this.
>: That will go into the record of Dejas, Bob.
>: Let me still inform you that you're badly wrong.
>: You could have continued. Even if they paid for Hsu's children (school,
>: university, flight to Jupiter etc.), his wife and dog (horses), car
>: grandmother, the whole East-Asia, Hsu's "duties" as scientist would still
>: remain the same. Period.
>You simply don't know what you are talking about.
(!)
>When you sign on at IBM,
>you sign a "contract". They retain all intellectual property rights for
>anything *you* create while working for them. You can leave, but if you try
>to take the technology, or the ideas, and use them, you end up broke, and
>alone, because of the contract. So Hsu doesn't have the "power" you'd like
>to attribute to him. He has definitely "reported" on what he did, which
>is the purpose of any good scientist. But he can't use the thing in any
>way he sees fit, when it is so valuable to the company. Since you've never
>worked for such a company, you don't understand just how much control they
>have. But if IBM so chooses, DB will *never* be seen again, and there's
>nothing Hsu can do about it...
I can't hold back the tears. But that's not at all the problem I was
talking about. I exclusively talked about the behaviour of the team during
the match.
>: (That's the OPPENHEIMER dilemma. Never thought about that?)
>I met the man at Los Alamos, in fact. And I don't see any "dilemma" (or
>delimma) either. He's published what he did. But Los Alamos would not
>have let him build a bomb on their nickel, and then take it off and use
>it wherever he wanted. Had he left, he had a contract that was *very*
>specific. It is on exhibit at the museum and the Los Alamos lab, in
>fact... And I can guarantee you that had he left Los Alamos before the
>project was finished, he'd have been followed for the rest of his life
>to be certain he didn't reveal anything he had promised not to reveal.
But that was NOT the dilemma! I meant the question of the ammount of the
right to a say for the top scientists.
Scientific standards and moral are "higher-situated" than the power/law of
the government.
>:>You are making an assumption about "Hsu, the scientist" that is *wrong*.
>: In the case of Hsu and Co. I was badly wrong, that's true ... ;-(
>But not for the reason you try to imply.
For 100%.
Kasparov has well described the sitiuation after the second game.
>:>IE, suppose that UAB decided that Crafty has become popular enough that
>:>they'd like to take advantage of this and start selling it. Could they
>:>do this? probably. would they do it? Almost certainly not ever. But
>:>*if* they did, does it suddenly become "Bob, the commercial programmer
>:>when he promised he would never do this?" If so, you are misdirecting
>:>your anger. If not, then there is no reason to make such comments about
>:>Hsu either...
>: I told you before that I already had understood that you were on a proxy
>: mission for "them" buddies. But this time you screw up the whole job.
>: You even tried to sac "yourself" in your wish to defend 'em.
>I do *not* "sac" myself. I point out the very tight straightjacket they
>are in, working for a very large company with a huge legal department..
But you seem to start a fog-machine. Of course they had their loyalities.
But the former mentioned standards could not been altered.
>: I wouldn't say anything the like in your direction only because you had
>: always confirmed us that you never wanted to go commercial ... Nonsense.
>And I've sit across the table from Hsu, or in a relaxed setting with other
>programmers and Hsu, or talked with him via email... and he is every bit
>the same as the rest of us. He is *less* secretive than the commercial
>programmers, because he has revealed every detail about the hardware,
>about the software, and so forth. He just can't take the machine and compete
>wherever he wants, because the IBM guys discovered that there was a lot of
>money in this project, via P/R benefits...
The next time you exchange emails with him please tell him that I don't
talk about his commercial deal, the details of which I don't know anyway.
Tell him that I'm deeply concerend with the artifical weakening of Garry by
his unfriendly, non-cooperative behaviour after the second game. Hsu will
probably understand me. If he should read this I beg him to show at least
the class to admit it now, after the event. He simply srewed up his
experimental setting. We were not informed and didn't know that it was all
about the influence of superstition and causing nightmares to a human
player but about the strength of Deep Blue. Period.
>: You completely missed the point. It's *not* that Hsu &Co. went commercial.
>: It was that he betrayed his scientific duties. With going commercial
>: *alone* that would be touched at all. And to keep the tension at the peak.
>: It's *not* the point that he allowed a behaviour that was (as you said) in
>: concordance to the written contract. It was the point --- now what do you
>: think?
>His scientific duty is to develop technology and provide the details of that
>technology to others. He has done this. I know how his hardware operates.
>So do many others.
Excuse me. Hsu was happy that he could get into contact with Kasparov. It's
simply not smart to let him then enter through the bathroom window and to
treat him like an unwanted person. Period.
(The point is that Kasparov couldn't tell all that even less because he
really would have been judged as a poor loser. But I can make that point.)
>The "product" which is owned by IBM is something else.
>And they control that to use it as they see fit to maximize company profits.
>But he has *definitely* published the internal details of DB.
>: It's the cheat that the "match" as it was held could *not* answer the
>: question tooted before the match, mankind vs or "or" machine. Simply
>: because it was *not* the machine but "machine+human factor X". And this
>: doesn't mean at all the often made tongue in cheek, "that of course the
>: machine must be made somehow by human beings, no?" No, the point is, that
>: between rounds the human gang (sorry, I meant "team") tweaked and twisted,
>: **NOT** to input something like the value of the BB-pair.
>That's stupid.
That's the way you try to understand my arguments.
>Of course it is man vs machine+human factor. Who wrote
>the software and built the hardware? A room full of monkeys?
Judging from the horrible press conference one could well come into
conflict about that question.
But to keep you stable I confirm you that I also took that for granted. :)
>Then it would
>be man vs machine+monkey_factor. There's *no* way that the "human factor"
>won't be involved when humans write the software and design and build the
>hardware,
Until here it's ok.
>and there's nothing wrong with humans modifying the software between
>rounds, since that has been an allowable rule in computer chess since the
>first tournament in 1970.
You must be kidding. You were operating and plaaying and twisting all on
the same objects. The computer hardware and chess software. You did it and
your opponents too. Fair deal.
But did *you* ever before play the match "mankind vs machine"? I must have
missed that one.
Can't you see the difference?
I'm not saying that in all computer vs human matches that factor should be
treated in the same way. But my strong opinion about the match Kasparov vs
Deep Blue in 1997 is that here the machine had to have played on his own.
Formerly well instructed by the human "factor".
It's true I never had the chance to work for IBM or any company of that
big, it's true I did never write my own chess program, it's also true that
I'm not known in the world of science due to my hundreds of specific
publications. But I know one thing. The importance of the isolation of the
relevant factors/variables carefully in an experimental setting. And here
the teamsters, no matter how smart and experienced they might be, have
badly failed. The result simply didn't mean much anymore. Period.
>: The point is this.
>: a) science
>: ==========
>Yep...
It's funny how much you're fixed on rather juridical questions. We're
talking about politeness and scientifical decency. If Bruce and you
cooperate in the way you described so vividly, then I can only congratulate
you both.
But I can't understand what that had to do with our questioning the
performance of the scientists around DB in New York, May 97.
>I'd expect that if game 6 were
>played again, do you not think Kasparov would play the correct move order
>this time? Isn't that the same sort of "cheating" that you are disdaining?
>Of course..
>: And the factor of the confusing of the human player, excuse me I want to
>: reserve the notion "cheating" for that one, was *not* the original
>: question, no?
>Had *nothing* to do with the original "question" so it was, and still is
>an irrelevant issue...
For you. Because you have lost contact to human chess. As it seems to me.
>: Another example of a factor among others that could be found.
>: You said that it was about "winning against the Wch". Fine with me. But
>: winning on the basis of the strength of the machine (as it was in May 97)
>: and *not* on the basis of some confusingly twistings here and there ...
>That is *your* definition of the question. It is *not* the definition of
>the question that has been posed for so many years.
Excuse the impetuosity of a rather young student. :)
>: subpoint 123a)
>: You might argue, as you already did BTW, that this "cheating" should be
>: allowed and no problem for Kasparov, because "he himself also had the
>: possibility to do the same cheating". It's more a point for the next
>: paragraphe "sports". But also in science this is relevant.
>: If you want to test the strength of a machine against a human champ. You
>: cannot argue that "because the human champ has typically human advantages
>: of flexibility for example, that it therefore should be allowed that the
>: machine could also be twisted (by its human operators of course). Because
>: then you're testing a totally different question. As I already explained it
>: it's the question if it's possible to cheat/confuse a human player in a
>: short match of 6 games with the help of in-between-round twistings of the
>: machine. If the press would have explained *that* question before the
>: match, I think both Kasparov and the audience worldwide would have
>: understood the task for the human champ.
>This is stupid reasoning. Suppose the program hits something "new" in the
>code and crashes. It would lose that game on time, as the rules forbid
>changes *during* a game. But should it have to play the match with that
>bug? And lose game after game? Should Kasparov have to play the game 6
>blunder every time or can he run to the opening books and look to see what
>he did wrong after the game?
Only if running is allowed ...
>Computers don't learn like humans. Computers don't "think" like humans.
>Computers don't correct their own bugs like humans. Get the gist here:
I would allow a technical hardware support but nothing else. After the
prominent Tueschen Rules this should be allowed, yes, I should think so.
>*computers* *are* *not* *like* *humans* *at* *all*. So stop trying to cast
>the match as man vs machine-in-a-box. That wasn't the question. That wasn't
>the way the match was planned. That wasn't in the rules of the match. If you
>don't like this, go set up your own match, put up your own money, and make any
>rules you like. But for *that* match this wasn't in the rules, and it makes
>no sense to continue the "banter" about it.
Hey, I just had this idea. Why IBM shouldn't buy Kasparov to let him play
the next games for DB ...?
>: So because of this cheat the scientifical reputation of Hsu and his team is
>: gone into the Hudson (?) river.
>Not at all. Again, that is *your* opinion. Held by very few. Most are
>quite impressed with what they accomplished..
The difficulty of our discussion could well be described with the irony I
had to agree with to your last statement. And yes, we should talk in a
relaxed atmosphere. And no, this is not boxing.
>: b) sports
>: ========
>: Do I really need to explain whjat the cheat meant in relation to fairness
>: in sports? I tell you. It means nothing.
>: In a one-game match?
>: ~~~~~~~~~~~~~~~~~
>: It's a common knowledge that a total newbie could beat a master in a
>: competition where the master cannot exploit his superior knowledge.
>: It's the same in backgammon or pokerstrip. Or in a Las Vegas gambling hall.
>: You might win a fortune with a 10 dollars entry. But this is called
>: gambling. If you take that into a serious sport like chess you get what you
>: imputed. Stupid nonsense.
>A newbie *can* beat a master. But the probability of this happening is
>so low as to be called *zero*.
ROTFL.
Never saw that in tennis e.g.?
>: Some time ago I saw the 17 y. old Michael Chang play the then number one
>: Lendl in Paris. Chang had succeeded in fooling Lendl with a faked complete
>: weakness in his legs. Suddenly he tried to go even further. He faked that
>: his arms too were weak and it was his serve. So he hit the ball from below
>: like a beginner. And Lendl was so surprised, he had never seen that ball in
>: all his training sessions, that he was completely out of his concentration.
>: Exactly his best quality, the ability to concentrate on the spot, was
>: working against him, because the laughter of the audience was disturbing
>: him too. An average hobby player would have given Michael Chang a return
>: that he surely wouldn't returned again.
>: What did this prove? That Chang was "better" than Lendl? No. Only that you
>: can exploit even the best strengths of a player if you are allowed to use
>: all tricks. I for one did never like the sportsman Chang. Because I never
>: believed him anything. What I still accepted was his strength as a
>: comparably "little" player among the giants ... Also his will-power and
>: concentration. But honesty was no longer something I saw in him.
>And the point would be? This wasn't a "chang" issue, it was idiocy on the
>part of Lendl...
Yes, it's always the same blind spot in your reasoning. But your blind spot
is my whole profession. :)
>: What did IBM/DB do?
>: ~~~~~~~~~~~~~~~~~~
>: They achieved to confuse Kasparov after the first game. Period.
>: For me it's still not proven that there wasn't something fishy. I can only
>: rely on Kasparov's Oxford lecture where he stated that 9 months after the
>: event he hadn't still the output. Whatever that meant. And I think we could
>: assume that Kasparov knows about the complications in a endless check
>: situation. We agreed to give it a rest for a later re-entry.
>This is wrong. He clearly stated that he did get the output for the three
>moves he questioned. He did *not* get the output for all the games. He
>should *never* get this, IMHO. Yasser wrote an article in the JICCA that
>discusses this at length. And he saw the printouts according to the article.
Please send me a copy, please.
I still am on the actuality of 1996 ... (JICCA)
>: Also they affected his mental stability. Because they acted unfriendly
>: against him.
>*tough* *stuff*. He accused *them* of cheating. And he then says they
>are unfriendly? Har-de-har-har...
Rah.. ed.. rah .. rah.
The other way round. At first they were unfriendly. *Then* he accused them.
"Show me one single proof where he *started* the like." (Guess who once has
written such a similar statement?)
>: I saw again many tennis matches that were mainly decided through the
>: interventions of the referee!! Because some players are so sensitive that
>: they loose control if something goes wrong.
>: And you should understand that a human champ against a machine could well
>: be confused by the behaviour of the operators of the machine.
>Again, tough. He's the best in the world.
Dr. Frankenstein again. :)
>: I saw this man (Hsu?) in the press conference after the match. And Bob,
>: excuse me, I had the impression to see an older clone of Michael Chang. I
>: simply could see the cheat in his eyes. I must apologize to Chang. Even he
>: looks more honest that this man on the press conference. Excuse me Bob, I'm
>: only telling you what I saw. I have it on video and I have always the same
>: bad feelings when I see it ...
>Bullshit... If you want to resort to personal insults, I'm going to terminate
>this discussion from my end. I know Hsu. Your remarks are insulting to *all*
>of us that know him. Stop, or talk to yourself from here on out...
I already asked you to adjourn. But aren't you the one who's famous for his
superficiality what concerning the data as such? How could that what I
personally saw on video be an insult? Check the data before you start
another of your campaigns. I can only tell you that it was NOT my intention
to insult someone. The same in my personal descriptions about Michael
Chang. It must be allowed to publish one's personal opinions. That's a
democratic right. Whatever that means to you.
But believe me I had the same strong feelings when you talked about the
"jackass" Kasparov ...
>:>No. I do this *every* last day. They felt the bishop pair bonus had been
>:>set too high. They adjusted this between games one and two. I notice that
>:>Kasparov changed his plan mid-match too, deciding to not play standard opening
>:>theory, and try to take DB out of book as early as possible. Was that fair?
>: For all it's not true.
>: How would you call the first game? Regular opening choice? :)
>yes...
>:>I don't know Tan. Have no impression of him, nor no personal opinion about
>:>him... But the word "cheat" is *wrong*. Cheat -> doing something contrary
>:>to the rules. This *clearly* wasn't done...
>: Excuse me for my dickhead. It *clearly* was against all standards of
>: science and fair sports.
>"standards" has nothing to do with it. "cheating" is about *rules*...
Of decency in science. New for you?
>:>
>:>If you would like to propose different "rules" you can do that. But for the
>:>rules used between DB and Kasparov, everything was quite above-board...
>: Dickhead cannot agree here. :)
>You didn't have to agree, only Kasparov. He agreed to the rules...
>:>Define team. I call this Hsu, Campbell, Hoan, etc... And I disagree. If
>:>you call the team Hsu, Campbel, + the IBM lawyers and marketing types, then
>:>things change. But not the "team" that most everyone would refer to.
>: I know it must hurt you. But it's the truth. Sorry. They are guilty.
>Only in *your* mind, which doesn't hold a lot of weight with *me* nor
>anyone else working in the field of computer science...
That's why you showed so much class here in denying the point of decency
and standards of science ...
>:>:>question. But nobody forced them to ask/answer *that* question. They picked
>:>:>the question, and then answered it. Just because we don't like the specific
>:>:>question they asked, doesn't mean it isn't a legitimate question, because 20
>:>:>years ago no one gave a computer a chance of being able to answer any such
>:>:>question with "yes".
>:>: The main difference for me is the implementation of all the million games.
>:>Don't understand that statement...
>: Eh, I was reminding you of the out-of-engine situated opening department.
>: The rest is a "bit" faster but still not on GM level. Understood this way:
>: regarding the weaknesses. Not the many strengths already.
>: I must come back on the cheating, excuse me. It's cheating if you twist the
>: moment you (the human operator) observe that the human opponent of your
>: machine has "gone on" a particular weakness. Especially in a 6 game match
>: where you only have three games with the same colour ...
>Again, this is *not* cheating unless it is written into the rules of the
>game, *or* the rules of the match. It was *not*.
Some rules are eternal rules. They don't have to be written on paper.
>: It's as if you said to the human player, go ahead but you won't understand
>: in the three games against what you're playing when.
>: I don't know of any sport where such a secrecy of real class is uses
>: systematically. What we already know is the fact that a newcomer always has
>: this certain advantage due to his status. So in the beginning of his career
>: he will win solely based on that factor.
>: (It should be repeated that a human newbie or expert could change his mind
>: overnight, *but* he could never change his overall personality. Under
>: stress he will remain the same in a certain range of course. After some
>: time the others know how to deal with him either to their advantage or in
>: some cases to their worst. Because it's a typical fact in sports that some
>: opponents no matter how strong they are objectively will be too strong for
>: someone (who could also rank higher in the hierarchy).)
>: And in chess in the case of Kasparov?
>: How should he change his chessic style overnight? I mean he knows all about
>: it. So he can't just say, now I will finally exploit the strength od my
>: bishop pair. I really can't understand what you meant when you therefore
>: claimed that kasparov too should be locked up in quarantaine. Does a
>: machine talk?
>Deep Blue didn't change its "style" overnight. They adjusted the value for
>the bishop pair. To kasparov, that affected how it behaved. I have adjusted
>that particular score a few dozen times myself and it does change things.
Yes, and it should be forbidden after the moment the match had begun.
>But
>it *doesn't* change the deep tactical accuracy of the program, nor its overall
>"personality". Just that Kasparov decided to play on the bishop pair theme,
>and it suddenly behaved differently. Just like Kasparov decided midway thru
>the match to play oddball openings to get it out of book quickly. That sounds
>like cheating according to your very loose and informal definition...
>: It's always good to react or to prove one's flexibility but here you went
>: too far. Ilias had a nice position for the question of "depth". You should
>: react on-topic.
>: But in the other question. It's often said that it's this what seperates
>: men from women. Sort of stability and abstraction of the fallacies of
>: personal involvement. I for one would like to hear what you had to argue
>: against Chris. What I don't appreciate is the sort of sissy-like
>: demonstrating of "feeling-insulted". I think we could demonstrate the
>: beauty and elegance of science if we could show to the group that we could
>: talk in a civilized mode even if we are not the closest friends. But the
>: level must be there of course.
>: What you're doing here is similar to Ed Schroders reaction. As if the
>: critic against you would aim against your personal dignity. I mean you
>: created crafty, so about what should you talk about? Sure, I also claimed
>: that you as expert had even *more* to see the different approaches and to
>: mediate in-between. Perhaps this will happen with your becoming older. :)
>Not the same thing at all. I just choose to stay away from the "crafty
>does it like this" a little more..
A pity.
>: We know that many here in rgcc want to hear you about crafty and all. So
>: please don't take all of them ias hostages because you found that Chris has
>: made some attacks. Rather try to answer those attacks. As you have done so
>: many times before.
>: Let me appeal to you and all. The actual stigmatizing of Chris (who's your
>: collegue-programmer) is also a sort of cheating for scientists at least.
>: You shouldn't follow this trend of defamation. And you surely shouldn't be
>: the leader. But then I agree, you can't be scientist and politician at the
>: same time. BTW, did you therefore stop talking about crafty?
>I have responded to technical questions about it. I have ignored many
>"what do other programs play here?" posts...
Do what you think is good for you. But perhaps you jump in later again.
Robert Hyatt
: A little correction if you don't mind.
: It's not about what you descibed.
: I was talking about the influence on their work as such. That hasn't
: changed in my eyes whether they did it in a universitary lab or "in" IBM. I
: will give an example later in this post.
:>That's the point. Hsu can't agree to another DB match. It's an IBM
:>decision. Hsu can't agree to play DB in FIDE tournaments, it is an
:>IBM decision. IF, for some reason, IBM becomes convinced that having DB
:>in FIDE events will further their corporate "image" then this might
:>happen.
: Ok, but if they participate does Hsu *then* have the right to follow his
: scientific standards (moral)?
Hard to say, but probably *no*. IE when you work under a contract with
a large company, you are in a box. You can leave (and leave the
intellectual property you developed behind since they own it) or you can
jump when they say "FROG!"
:)
: I see. So we are just right in the middle of preventing exactly that ... :)
I doubt *we* will prevent anything. We could both yell and shout to the
top of our ability, and not even .0001% of the people in the world that matter
would know about it. The world is a big place. We are very small in that
scale..
: Good to know that you didn't yet inform Kenneth Starr of my crimes.
: But I want to answer your question. Not as an expert of justice. Also I
: wouldn't say that scientific standards could be made juridically relevant.
: So, no new court trials, Bob ...
: Another side remark. We can say that we haven't understood the other but we
: shouldn't say that the other doesn't know what he's talking about. We
: shouldn't make "CCC of Nov97" to our reference.
: Here#s my simple answer. The DBteam left their former scientific interests.
: What they then tested was how far Kasparov could be affected by unfriendly
: non-cooperative behaviour of the *operating team*. NB that the team didn't
: play the games. Just in case you completely mixed it up. That's as if the
: captain of a certain team went crazy in a team match. Of course the players
: get affected ...
IBM wasn't "unfriendly" to Kasparov at all. *until* he started the game-2
garbage, any way. Before that they had, after all, put up over one million
bucks for him, and a year ago put another big wad of cash in his pocket.
So it wasn't exactly like they were "out to ruin him."
: I can't hold back the tears. But that's not at all the problem I was
: talking about. I exclusively talked about the behaviour of the team during
: the match.
What did or didn't they do during the match? Give him program output?
He wouldn't do the same for them had they asked him after game one "why
did you do this and not that?" I miss the "point"...
: But that was NOT the dilemma! I meant the question of the ammount of the
: right to a say for the top scientists.
How much do you think *he* reported during the Manhattan project's lifetime?
Can you spell *treason*? One mention and he'd have been dead, buried and
replaced before the word was out of his mouth... There was *maximum*
secrecy during that project. That's why Los Alamos was built up in the
mountains as far away from anywhere as possible, without being in the
desert where it was obvious...
: But you seem to start a fog-machine. Of course they had their loyalities.
: But the former mentioned standards could not been altered.
I don't get your point then... They had to do what they were told to do,
assuming that they wanted to keep getting a paycheck. Most of us do that.
: The next time you exchange emails with him please tell him that I don't
: talk about his commercial deal, the details of which I don't know anyway.
: Tell him that I'm deeply concerend with the artifical weakening of Garry by
: his unfriendly, non-cooperative behaviour after the second game. Hsu will
: probably understand me. If he should read this I beg him to show at least
: the class to admit it now, after the event. He simply srewed up his
: experimental setting. We were not informed and didn't know that it was all
: about the influence of superstition and causing nightmares to a human
: player but about the strength of Deep Blue. Period.
He screwed up the setting *you* wanted. Again, they wanted to beat
Kasparov at the game of chess. Nothing more, nothing less. They did
accomplish that. Maybe we now want to see the goal "become the world
champion via the FIDE tournament trail." Good next goal. Would eliminate
your main argument about secrecy. But that wasn't the goal for NY.
: Excuse me. Hsu was happy that he could get into contact with Kasparov. It's
: simply not smart to let him then enter through the bathroom window and to
: treat him like an unwanted person. Period.
: (The point is that Kasparov couldn't tell all that even less because he
: really would have been judged as a poor loser. But I can make that point.)
Too late. He is a poor loser IMHO.
: That's the way you try to understand my arguments.
No, but this gets repetitious. You have your question, they had theirs.
There's no way you can force your question into their mouth and then
complain because they didn't answer it... You question is legitimate.
Theirs was too... IE even drug tests are often done "blind", for
different reasons, but blind nonetheless.
: Judging from the horrible press conference one could well come into
: conflict about that question.
I thought we were talking about the DB hardware and team of programmers,
not the Kasparov camp? :)
: You must be kidding. You were operating and plaaying and twisting all on
: the same objects. The computer hardware and chess software. You did it and
: your opponents too. Fair deal.
: But did *you* ever before play the match "mankind vs machine"? I must have
: missed that one.
: Can't you see the difference?
Nope...
: I'm not saying that in all computer vs human matches that factor should be
: treated in the same way. But my strong opinion about the match Kasparov vs
: Deep Blue in 1997 is that here the machine had to have played on his own.
: Formerly well instructed by the human "factor".
Again, *you* want to make rules. That's ok... but only when *you* make
the match happen. *they* designed the hardware, spent the time testing
everything, paid GM's for advice, put up the prize fund, and *they* should
have consulted you (or someone else) for a modified set of rules?
: It's true I never had the chance to work for IBM or any company of that
: big, it's true I did never write my own chess program, it's also true that
: I'm not known in the world of science due to my hundreds of specific
: publications. But I know one thing. The importance of the isolation of the
: relevant factors/variables carefully in an experimental setting. And here
: the teamsters, no matter how smart and experienced they might be, have
: badly failed. The result simply didn't mean much anymore. Period.
Except in this case something was done that had never been done before.
Next time, maybe they should play in FIDE events first. But for *this*
time, the result was most impressive, regardless of whether we'd like to
see other constraints or not...
: It's funny how much you're fixed on rather juridical questions. We're
: talking about politeness and scientifical decency. If Bruce and you
: cooperate in the way you described so vividly, then I can only congratulate
: you both.
I'm not hung on "decency"... I am hung on rules that I have been playing by
(in computer chess) for nearly 30 years. Rules adopted by the USCF chess
federation concerning computers in human events, etc... Nothing has changed
at all.
: For you. Because you have lost contact to human chess. As it seems to me.
possibly. But this wasn't *human* chess. And that is the point here...
: Only if running is allowed ...
:>Computers don't learn like humans. Computers don't "think" like humans.
:>Computers don't correct their own bugs like humans. Get the gist here:
: I would allow a technical hardware support but nothing else. After the
: prominent Tueschen Rules this should be allowed, yes, I should think so.
Again, you can make rules as you wish. But getting them into a K vs DB
match is something else...
:>A newbie *can* beat a master. But the probability of this happening is
:>so low as to be called *zero*.
: ROTFL.
: Never saw that in tennis e.g.?
Nope. Not enough to count. How often does an unseeded player beat #1?
rarely enough that it makes the world-wide news...
Chris Whittington
Robert Hyatt wrote in message <6u6t80$kd6$1...@juniper.cis.uab.edu>...
[ with reference to Oppenheimer dilemma ]
>
>
>: But that was NOT the dilemma! I meant the question of the ammount of the
>: right to a say for the top scientists.
>
>How much do you think *he* reported during the Manhattan project's
lifetime?
>Can you spell *treason*? One mention and he'd have been dead, buried and
>replaced before the word was out of his mouth... There was *maximum*
>secrecy during that project. That's why Los Alamos was built up in the
>mountains as far away from anywhere as possible, without being in the
>desert where it was obvious...
>
I can't believe this.
Hyatt, you are this illiterate in your knowledge ?
You really don't understand ?
Incredible.
Chris Whittington
Rolf Tueschen
later.
I think I should express my feelings after another big exchange. I want to
thank you for the time you invested. Also for the relative fairness you
adopted for this serie at least. Although I saw that in the last post you
was losing control again. It's good to see now that you apparently
understood that I tried to differentiate strong opinions about public
events and persons of modern history if you want, and solely destructive
intentions that could be seen BTW very good actually in the House of
Representatives where they sell out all decency politics normally still has
in the population. And that world-wide. I think I did never do that in the
case of Hsu and team. It's also nice to see that you restricted yourself a
little bit in the often tried ad hominem of reproaching incompetence,
because following that line almost nobody in the world had the "rights" to
criticise anything about the famous and successful experts of that Hsu
team.
For me most of such exchanges had even a deeper relevance for me
personally. But this is an open forum so I want to take care of not going
too deep into details right now.
Beyond all differences I want to thank you for your inspiring competition.
As far as the many details are concerned you will surely agree that the
context of the paragraphes was now screwed up by your snippets. But I agree
at the same time that we had reached another peak record as far as quotes
are concerned. Technically we had to stop that but the logic was no longer
there afterwards. No offense of course because Usenet is not an academic
colloquium. For those who know to read the different standpoints must have
shown through anyway.
I think that you in your role should even go further in the polite
acceptance of such sorts of opinions because not meaning that you *agree*
this would promote the attraction of scientific smartness on the interested
young readers. I hope that the idea alone could clarify itself against the
background of my primitive use of the English language. For all when it was
not at all my own idea, but I'm only the *reporter* ...
The value of c a t a l y s t s so to speak should be ranked much
higher in general. We tend too often to take someone one-to-one for what he
tried to *tell* us. It's true, even a catalyst needs an aerea of freedom
and creativity to express himself. And often certain expressions need sort
of preliminary heat. Moderation understood as a bureaucratic job would
screw up the emergence of the main topics.
But now enough. Thanks for your patience again.
Robert Hyatt <hy...@crafty.cis.uab.edu> wrote:
[..]
Robert Hyatt
: later.
: I think I should express my feelings after another big exchange. I want to
: thank you for the time you invested. Also for the relative fairness you
: adopted for this serie at least. Although I saw that in the last post you
: was losing control again. It's good to see now that you apparently
: understood that I tried to differentiate strong opinions about public
: events and persons of modern history if you want, and solely destructive
: intentions that could be seen BTW very good actually in the House of
: Representatives where they sell out all decency politics normally still has
: in the population. And that world-wide. I think I did never do that in the
: case of Hsu and team. It's also nice to see that you restricted yourself a
: little bit in the often tried ad hominem of reproaching incompetence,
: because following that line almost nobody in the world had the "rights" to
: criticise anything about the famous and successful experts of that Hsu
: team.
I didn't say that at all. I said "without evidence" this shouldn't happen.
I also said that you want to ask a different question, an admittedly valid
question, but still a *different* question than the DB match was established
to answer.
I, too, am interested in how DB or any program would do in steady competition
against top GM players. Sadly, we may never know about DB *or* other programs,
because such competition is very infrequent at best...
And I hope I didn't "come close to losing control." I only point out,
repeatedly, that what you wanted to learn from the match is *not* necessarily
what everyone else wanted to learn from the match, particularly those setting
the match up.
: For me most of such exchanges had even a deeper relevance for me
: personally. But this is an open forum so I want to take care of not going
: too deep into details right now.
: Beyond all differences I want to thank you for your inspiring competition.
: As far as the many details are concerned you will surely agree that the
: context of the paragraphes was now screwed up by your snippets. But I agree
: at the same time that we had reached another peak record as far as quotes
: are concerned. Technically we had to stop that but the logic was no longer
: there afterwards. No offense of course because Usenet is not an academic
: colloquium. For those who know to read the different standpoints must have
: shown through anyway.
: I think that you in your role should even go further in the polite
: acceptance of such sorts of opinions because not meaning that you *agree*
: this would promote the attraction of scientific smartness on the interested
: young readers. I hope that the idea alone could clarify itself against the
: background of my primitive use of the English language. For all when it was
: not at all my own idea, but I'm only the *reporter* ...
I *never* disagreed with your idea. I did disagree with your dismissal of
the DB team as "cheats" simply because the match they created didn't answer
the specific question you wanted answered. IE no modifications between
rounds, GM has access to all old games and printouts from each round during
the match, etc...
: The value of c a t a l y s t s so to speak should be ranked much
: higher in general. We tend too often to take someone one-to-one for what he
: tried to *tell* us. It's true, even a catalyst needs an aerea of freedom
: and creativity to express himself. And often certain expressions need sort
: of preliminary heat. Moderation understood as a bureaucratic job would
: screw up the emergence of the main topics.
: But now enough. Thanks for your patience again.
: Robert Hyatt <hy...@crafty.cis.uab.edu> wrote:
: [..]
Gerry Forbes
I'm surprised that this thread is still going. If my news server
behaves I'll add a couple of comments.
Robert Hyatt still believes that DB achieved the Holy Grail of
computer chess because it followed the rules, the quality of chess
being irrelevant. Does this mean that a computer will have achieved
AI when it can hire a lawyer who successfully sues somebody for
saying it isn't intelligent? But maybe it will sue for being called
artificial! I could only accept Hyatt's argument if he could
convince me that Kasparov's arrogance is Deep Blue's ability.
It is unnecessary for Rolf to imply that Hsu's integrity is
compromised. It's much easier to believe that he was unlucky:
not only did he have to seek his opportunity with IBM, he had
to have Kasparov as world champion. Things only go downhill fast
when Kasparov loses the first game in 1996. Actually it was not
that unlikely that K would lose one game in six, and that if he
did lose a game it would most likely be game one, especially if
DB was white. Unfortunately game one is when the media, especially
the TV crews, would be most numerous and maybe a little eager to
rub it in on some smartass chessplayer who's just bored them for
the last few hours with reasons why they shouldn't expect the
computer to win. If Deep Blue gets the same score but wins in
game five or six instead of game one maybe the suits don't step
in with their unscientific motives.
But Hyatt is deluded if he thinks Deep Blue will ever make a
comeback. He says that computer chess is only a tiny part of a
great big world, but he doesn't seem to realise that that great
big world was only shocked when DB won game one in 1996 because
they were told they were supposed to be shocked--it was already
common knowledge that computers always beat humans at chess. The
public will be really pissed off if they find out they have to
be undeceived *again* and IBM won't feel like answering all the
questions that come up. If Deep Blue returns it will be playing
go, not chess, since all it has to do there is stomp the best
go program to get publicity. IBM can easily promote it as the next
stage in their development at a small risk, there's the $1.6
million challenge and it helps their PR in Asia. Much better
return than trying to explain K-DB this long after the fact--
just ask Clinton.
Robert Hyatt
: I'm surprised that this thread is still going. If my news server
: behaves I'll add a couple of comments.
: Robert Hyatt still believes that DB achieved the Holy Grail of
: computer chess because it followed the rules, the quality of chess
: being irrelevant. Does this mean that a computer will have achieved
: AI when it can hire a lawyer who successfully sues somebody for
: saying it isn't intelligent? But maybe it will sue for being called
: artificial! I could only accept Hyatt's argument if he could
: convince me that Kasparov's arrogance is Deep Blue's ability.
The "quality" of chess wasn't irrelevant. IE every GM I know of agrees that
game 2 of the match was as well-played as *any* game between two GM's in the
world, excepting the problem with Kasparov resigning. But nothing says that
neither player made no bad or inferior moves. *Both* did. Remember that
Kasparov actually resigned in a drawn position... So it wasn't perfect, but
it was absolutely amazing to watch the machine play chess with Kasparov and
not feel like it was always on the verge of getting blown out. Rather, it
was always "what magical wizardry is the machine going to weave in *this*
game?"
: It is unnecessary for Rolf to imply that Hsu's integrity is
: compromised. It's much easier to believe that he was unlucky:
: not only did he have to seek his opportunity with IBM, he had
: to have Kasparov as world champion. Things only go downhill fast
: when Kasparov loses the first game in 1996. Actually it was not
: that unlikely that K would lose one game in six, and that if he
: did lose a game it would most likely be game one, especially if
: DB was white. Unfortunately game one is when the media, especially
: the TV crews, would be most numerous and maybe a little eager to
: rub it in on some smartass chessplayer who's just bored them for
: the last few hours with reasons why they shouldn't expect the
: computer to win. If Deep Blue gets the same score but wins in
: game five or six instead of game one maybe the suits don't step
: in with their unscientific motives.
: But Hyatt is deluded if he thinks Deep Blue will ever make a
: comeback. He says that computer chess is only a tiny part of a
: great big world, but he doesn't seem to realise that that great
: big world was only shocked when DB won game one in 1996 because
: they were told they were supposed to be shocked--it was already
: common knowledge that computers always beat humans at chess. The
: public will be really pissed off if they find out they have to
: be undeceived *again* and IBM won't feel like answering all the
: questions that come up. If Deep Blue returns it will be playing
: go, not chess, since all it has to do there is stomp the best
: go program to get publicity. IBM can easily promote it as the next
: stage in their development at a small risk, there's the $1.6
: million challenge and it helps their PR in Asia. Much better
: return than trying to explain K-DB this long after the fact--
: just ask Clinton.
I disagree. This is a marketing issue, not a scientific issue. IBM found
a huge PR flash in Deep Blue, and estimates about the worth of this project
from that perspective suggest staggering sums of money. And if they become
convinced that rolling DB back out again will be worthwhile financially, from
a P/R perspective, out it'll come again...