486 beat me at Doom shock

G Cooper

unread,

May 16, 1997, 3:00:00 AM5/16/97

to

Hmm,

I don't want to send any shockwaves into the Deep Blue debate, but I
bring you the bad news that I was unsuccessful in beating my 486
computer (nicknamed "Old Socks") at a game of 'Doom' last night. I have
been able to beat it in previous contests on the "easy" difficulty level
but this time I tried out the "hard" level, where there are twenty seven
times as many monsters and that.

Luckily it is the best of three games (or perhaps five), so there is
hope for us all yet.

If I fail you, I will try another contest, where I try to write the
numbers 1 to 1000000 with a pen and paper before "Old Socks" can print
them all on it's screen. It's going to be a close one, it's such a
clever machine!

Gareth Cooper

Journalism UK
http://www.octopod.demon.co.uk/journ_uk.htm

David Longley

unread,

May 16, 1997, 3:00:00 AM5/16/97

to

In article <1d$86EAGy...@ffaltd.demon.co.uk>
oct...@octopod.demon.co.uk "G Cooper" writes:

"Many a true word......."
--
David Longley (check end reply line #)

Jim Balter

unread,

May 21, 1997, 3:00:00 AM5/21/97

to

G Cooper wrote:
>
> Hmm,
>
> I don't want to send any shockwaves into the Deep Blue debate, but I
> bring you the bad news that I was unsuccessful in beating my 486
> computer (nicknamed "Old Socks") at a game of 'Doom' last night.

Perhaps, as with Kasparov, you simply played badly (7 ... h6).

--
<J Q B>

al c

unread,

May 21, 1997, 3:00:00 AM5/21/97

to

In article <1d$86EAGy...@ffaltd.demon.co.uk>
G Cooper <gar...@ffaltd.demon.co.uk> writes:

> If I fail you, I will try another contest, where I try to write the
> numbers 1 to 1000000 with a pen and paper before "Old Socks" can print
> them all on it's screen. It's going to be a close one, it's such a
> clever machine!

You could always cheat and write them down beforehand. The 486 would
never know!

-AL

ignorance is not bliss, but bliss is ignorance

Ralph Silverman

unread,

May 23, 1997, 3:00:00 AM5/23/97

to

David Longley (Da...@longley.demon.co.uk) wrote:
: In article <1d$86EAGy...@ffaltd.demon.co.uk>
: oct...@octopod.demon.co.uk "G Cooper" writes:

: > Hmm,

: >
: > I don't want to send any shockwaves into the Deep Blue debate, but I
: > bring you the bad news that I was unsuccessful in beating my 486

: > computer (nicknamed "Old Socks") at a game of 'Doom' last night. I have

: > been able to beat it in previous contests on the "easy" difficulty level
: > but this time I tried out the "hard" level, where there are twenty seven
: > times as many monsters and that.
: >
: > Luckily it is the best of three games (or perhaps five), so there is
: > hope for us all yet.

: >
: > If I fail you, I will try another contest, where I try to write the

: > numbers 1 to 1000000 with a pen and paper before "Old Socks" can print
: > them all on it's screen. It's going to be a close one, it's such a
: > clever machine!

: >
: > Gareth Cooper
: >
: > Journalism UK
: > http://www.octopod.demon.co.uk/journ_uk.htm
: >

: "Many a true word......."
: --
: David Longley (check end reply line #)

meaning of 'deep blue' win

a) Metaphysics tells us that
computers can do
only what they are programmed
to do !
we may infer that
a computer can not play chess
better than its best programmer ...

computers are not 'intelligent' and can not
'outsmart' human opponents ...

the confusion felt on this by
common folk must be resisted by Metaphysicians ...

( does 'deep blue' cry at weddings ? )

b) changing paradigms ...
of the 'post modern' era open new opportunities
in the field of
Metaphysics
generally and
in
the philosophy of artificial intelligence
as well as
the philosophy of science ...
particularly !
now we can 'let in' some of you to
'behind the scenes' developments !

c) Philosophy of Aviation Engineering ...
reportedly, the fastest airplane
is
lockheed sr-71
( sustained level flight )
...
no human exceeds 60mph ( sustained foot travel )
( @120kph )
therefore
no lockheed engineer exceeds 60mph ( sustained foot travel )
( @ 120kph )
Metaphysics tells us
no lockheed aircraft exceeds 60mph
( @120kph )
however puzzling this may be to
common folks !

since some lockheed aircraft is fastest;
Metaphysics tells us
no aircraft exceeds 60mph
( @120kph )
and none will ...
until aviation engineers can run faster !

d) practical significance of Metaphysics
...
how much money has been mis-spent
researching 'artificial intelligence'
... ( has anyone studied this ? ) ...
how much money has been mis-spent
researching aircraft exceeding
60mph ( @120kph ) ( sustained level flight ) ?
... ( has anyone studied this ? ) ...
and what other tragic mis-allocations of
planetary resources remain undetected ?

e) show me the money !
tragically ...
budgets for Metaphysicians
who might research
such matters are inadequate ...

--

Ralph Silverman
z007...@bcfreenet.seflin.lib.fl.us

Peter Hayes

unread,

May 25, 1997, 3:00:00 AM5/25/97

to

In article <5m46vv$n...@nntp.seflin.org>, Ralph Silverman <z007400b@bcfre
enet.seflin.org> writes

> meaning of 'deep blue' win
>
> a) Metaphysics tells us that
> computers can do
> only what they are programmed
> to do !

> e) show me the money !
> tragically ...
> budgets for Metaphysicians
> who might research
> such matters are inadequate ...
>

>Ralph Silverman
>z007...@bcfreenet.seflin.lib.fl.us
>

What a pile of dingo's kidneys!
--
Peter Hayes

Anyone who disagrees with me is pushing his sad religion.

Alicia Carla Longstreet

unread,

May 25, 1997, 3:00:00 AM5/25/97

to

When IBM comes up with a computer that can beat me at Bridge, I will be
impressed. I am NOT a great Bridge player, merely a decent one.
Actually, if the people at IBM can produce a computer that is even
barely competent at Bridge, I will be impressed.

As for 'Big Blue' beating Kasperov at Chess, this is nothing more that
using brute force to solve a problem, it is most UNimpressive.
Actually, I AM imressed that Kasperov was able to beat the machine,
considering that he could not begin to approach the computational speed
of the computer.

To me, the 'Big Blue' vs Kasperov match is comparable to racing a human
on foot, to a remote computer controlled 100 horse power race car, I
will be impressed if the human manages to beat the moterized computer
controlled car even once.

--
********************************************
* Alicia Carla Longstreet ca...@ici.net
********************************************
Be Good,
If you can't be good, be careful,
If you can't be careful, be lucky,
and if you're not lucky...
You damn well better be good!

H. M. Hubey

unread,

May 25, 1997, 3:00:00 AM5/25/97

to

Alicia Carla Longstreet <ca...@ici.net> writes:

>When IBM comes up with a computer that can beat me at Bridge, I will be
>impressed. I am NOT a great Bridge player, merely a decent one.

Most people probably won't be.

That's why probably nobody will spend millions making
bridge-playing computers. The reason for picking chess is
obvious; it has been shown everywhere as the prime example
of an intelligence-requiring and an intellectual game.

>As for 'Big Blue' beating Kasperov at Chess, this is nothing more that
>using brute force to solve a problem, it is most UNimpressive.

It also happens to be the case that Kasparov plays chess by
using brute force. But because it is done via the human
brain, we are told that it is intelligence.

There are many problems in which there is a time-space
trade-off. It just so happens that the natural brains have
their brute force in spatial parallelism. In fact they
eventually get hard-wired for certain tasks. We call that
expertise.

Here is a simple example of a task which can be done in
various ways. The fastest in-place sorting algorithm works
in O(n*ln(n)). That is because there is a space restriction.

Here is an algorithm that will work in linear time (given
unlimited space). For concreteness assume that the data
consists of integers. Simply run thorugh the data once
and determine the min and max numbers. Then make a 2-D
array, indexed from min to max. Then run through the data
and put each data in its own bucket. That's linear. IF we
had multiple processors we could divide it up into M buckets
and then we'd need O(N/M) time. If we had as many processors
as the number of data (i..e N pieces) and hard perfect
parallelism could sort it in place in constant time.

That's what the neurons do. There is really no difference
between electronic brawn and electronic brain. So it is
with natural brains too. If it weren't so, we'd probablyu
have monkeys more intelligent than human beings. But they
don't exist, because in the brain-size department (after
accounting for size of animal etc) we are the result
of evolution that ranks highest. And we are the most
intelligent.

>Actually, I AM imressed that Kasperov was able to beat the machine,
>considering that he could not begin to approach the computational speed
>of the computer.

I am impressed that the machine beat kasparov since it is
obvious that it is not working in parallel like Kasparov's
brain :-)

>To me, the 'Big Blue' vs Kasperov match is comparable to racing a human
>on foot, to a remote computer controlled 100 horse power race car, I
>will be impressed if the human manages to beat the moterized computer
>controlled car even once.

Yep.

From now on, we will be impressed if humans can match machines.

And this is the beginning. :-)

--
Mark Hubey ---------------------------------------------------------
http://www.csam.montclair.edu/Faculty/Hubey.html
hu...@pegasus.montclair.edu hub...@alpha.montclair.edu
--------------------------------------------------------------------------

G.V. Morgon

unread,

May 26, 1997, 3:00:00 AM5/26/97

to

z007...@bcfreenet.seflin.org (Ralph Silverman) wrote:

ą
ą meaning of 'deep blue' win
ą
ą a) Metaphysics tells us that
ą computers can do
ą only what they are programmed
ą to do !
ą we may infer that
ą a computer can not play chess
ą better than its best programmer ...
ą

Oh gawd... Your giving me a relapse into a Philosophy class (most useless class
I ever took, even after sociology). So since computers can beat their
programmers at chess, sounds to me that disproves metaphysics.

ą d) practical significance of Metaphysics

none

ą e) show me the money !
ą tragically ...
ą budgets for Metaphysicians
ą who might research
ą such matters are inadequate ...

I don't call that tragic. They shouldn't even be given a budget to begin with.

We may infer that: Metaphysicals ammounts to no more then philosophical drivel

Since we've digressed send resp. via email
Glenn

A. Sinan Unur

unread,

May 26, 1997, 3:00:00 AM5/26/97

to

H. M. Hubey wrote:
>
> From now on, we will be impressed if humans can match machines.
>
> And this is the beginning. :-)
>

i find it more accurate to think of this as a competition between a
group of highly skilled programmers, hardware designers, analysts etc
and kasparov. those people, through a lot of work and investment, came
up with a machine and a program that could analyze a particular
situation at the same level as kasparov does. hence, they beat him (this
time.) no the "machine". the machine was doing just what it had been
told to do. it is the people who have decided exactly what to tell it
and how to implement that who have succeded, that's all.

--
Sinan

*******************************************************************
A. Sinan Unur WWWWWW
|--O+O
mailto:sinan...@cornell.edu C ^
http://www.people.cornell.edu/pages/asu1/ \ ~/

Unsolicited e-mail is _not_ welcome, and will be billed for.
*******************************************************************

Paul Hsieh

unread,

May 26, 1997, 3:00:00 AM5/26/97

to

On 25 May 1997, H. M. Hubey (hu...@pegasus.montclair.edu) said:
> Alicia Carla Longstreet <ca...@ici.net> writes:
>
> >When IBM comes up with a computer that can beat me at Bridge, I will be
> >impressed. I am NOT a great Bridge player, merely a decent one.
>
> Most people probably won't be.
>
> That's why probably nobody will spend millions making
> bridge-playing computers. The reason for picking chess is
> obvious; it has been shown everywhere as the prime example
> of an intelligence-requiring and an intellectual game.

Actually, I've seen some very strong bridge programs (usually combined
with a tutorial) out there. But as you suggest, nobody really cares, so
these programs fell into relative oblivion.

> >As for 'Big Blue' beating Kasperov at Chess, this is nothing more that
> >using brute force to solve a problem, it is most UNimpressive.

Your knowledge about the internal working of Deep Blue is what is most
unimpressive. The Deep Blue team has in the past published several
papers which have been considered great contributions to the field of AI.
As far as I know they have continued to use some of their more successful
techniques in the Deep Blue program that helped it be more than just a
brute force search engine.

> It also happens to be the case that Kasparov plays chess by
> using brute force. But because it is done via the human
> brain, we are told that it is intelligence.

:o) An interesting way of looking at it. Many people believe that the
pursuit of artificial intelligence must be to model the way the human
mind works. But it also encompasses using the computer to think about
problems in ways unfeasible to humans, such as was done with Deep Blue.

> >Actually, I AM imressed that Kasperov was able to beat the machine,
> >considering that he could not begin to approach the computational speed
> >of the computer.
>
> I am impressed that the machine beat kasparov since it is
> obvious that it is not working in parallel like Kasparov's
> brain :-)

Actually, Deep Blue *is* a highly parallelized machine. Though, perhaps
not in the sense you mean.

> >To me, the 'Big Blue' vs Kasperov match is comparable to racing a human
> >on foot, to a remote computer controlled 100 horse power race car,

Your analogy belittles what chess is, and how hard it was for the Deep
Blue team to create such a device with such capability.

> > I
> >will be impressed if the human manages to beat the moterized computer
> >controlled car even once.

> Yep.

> From now on, we will be impressed if humans can match machines.
> And this is the beginning. :-)

:o) You have that little faith in man kind? Computers ability to do
some things better than humans will consequentially make humans better at
doing other things. For example the computer's ability to calculate
spread sheets faster than a human can do by hand, intrinsically made
humans (at least some) better at designing spread sheets to tell them
more, and give them more information.

Even in Chess, as these programs have been released as commercial
products, even the top grandmasters have been using them to check their
analysis, keep a database of openings, special opening repetoires as well
as being used as a way of studying the games of their opponents. It has
also been used to further the state of opening theory in some cases. But
they computer doesn't win the cash prize, the human who used it does.

--
Paul Hsieh
http://www.geocities.com/SiliconValley/9498/mailme.html

Richard Pitre

unread,

May 27, 1997, 3:00:00 AM5/27/97

to

Alicia Carla Longstreet wrote:
>
> When IBM comes up with a computer that can beat me at Bridge, I will be
> impressed. I am NOT a great Bridge player, merely a decent one.

> Actually, if the people at IBM can produce a computer that is even

> barely competent at Bridge, I will be impressed.

>
> As for 'Big Blue' beating Kasperov at Chess, this is nothing more that
> using brute force to solve a problem, it is most UNimpressive.

> Actually, I AM imressed that Kasperov was able to beat the machine,
> considering that he could not begin to approach the computational speed
> of the computer.
>

> To me, the 'Big Blue' vs Kasperov match is comparable to racing a human

> on foot, to a remote computer controlled 100 horse power race car, I

> will be impressed if the human manages to beat the moterized computer
> controlled car even once.
>

> --

How will we see Deep Blue 20 years from now. A very weak machine?
Then it will look like a human chess player can barely keep up with a
very weak machine of 20 years ago in a game of chess.

I believe that machines will surpass humans in every defineable
measure of reasoning ability.

The use of reason in the medium of language
to solve problems is often a successful approach to meeting some basic
human needs. On the other hand, the use of language and
rationality is nothing more than a particular type of behavior.
It obviously does not encompass all of useful
human behavior and it is silly egotism to expect that it can, in any
sense,
encompass the universe or a *description* of the universe i.e.
rationalism is
a religious belief, a deification of a tool.

Machines will extend our facility with language
behavior so that we can deal with more complex constructs more
efficiently.
We may find that at some degree of complexity language behavior loses
its
utility as a mechanism for dealing efficiently with our environment.
In some areas of need we may have long ago passed that point.
We just haven't stopped the mental masturbation long enough to notice
that
it doesn't even feel good anymore.

richard

Paul Campbell

unread,

May 27, 1997, 3:00:00 AM5/27/97

to

> Alicia Carla Longstreet wrote:
> >
> > When IBM comes up with a computer that can beat me at Bridge, I will be
> > impressed. I am NOT a great Bridge player, merely a decent one.
> > Actually, if the people at IBM can produce a computer that is even
> > barely competent at Bridge, I will be impressed.
> >
> > As for 'Big Blue' beating Kasperov at Chess, this is nothing more that
> > using brute force to solve a problem, it is most UNimpressive.

Agreed, no more impressive than the fact that my pocket calculator is
better than me (and presumably its programmers) at arithmetic.

Paul C.
UK.

Walter

unread,

May 27, 1997, 3:00:00 AM5/27/97

to

(Just my opinion..)
I think it is pretty impressive. Playing chess is something else than
calculating 1+1. Kasparov seemed very surprised by Deep Blue's
performance and he was very angry that he lost... :)

However, I also think Kasparov could beat Deep Blue in a next game
just the same. Deep Blue may not have won the war yet.

--
*** When you build it, they will come ***

wal...@ronix.ptf.hro.nl Go to BrintaBBS: telnet://bbs.hro.nl
wal...@and.nl where I am known as 'Walter'
http://www.ptf.hro.nl/~walter/

Dale Pennington

unread,

May 27, 1997, 3:00:00 AM5/27/97

to

Walter <wal...@and.nl> wrote in article <338AC...@and.nl>...

>
> (Just my opinion..)
> I think it is pretty impressive. Playing chess is something else than
> calculating 1+1. Kasparov seemed very surprised by Deep Blue's
> performance and he was very angry that he lost... :)
>
> However, I also think Kasparov could beat Deep Blue in a next game
> just the same. Deep Blue may not have won the war yet.
>
> --
> *** When you build it, they will come ***
>
> wal...@ronix.ptf.hro.nl Go to BrintaBBS: telnet://bbs.hro.nl
> wal...@and.nl where I am known as 'Walter'
> http://www.ptf.hro.nl/~walter/
>

Actually, as a friend of mine commented, Big Blue had a major unfair
advantage. It (actually, it's programming team including at least one chess
grand master) had plenty of time to analyze Kasparov's playing style and
determine the weaknesses in his game. It could look at his past history of
playing, as well as the books he had written. How much do you think
Kasparov knew of Big Blue's style prior to the match (can you say 0). So
Big Blue started with a major strategic advantage that no human player in
that class would ever have.

Szu-Wen Huang

unread,

May 27, 1997, 3:00:00 AM5/27/97

to

[wow, what a spam. can't quite trim it anymore.]

Alicia Carla Longstreet (ca...@ici.net) wrote:
: When IBM comes up with a computer that can beat me at Bridge, I will be
: impressed. I am NOT a great Bridge player, merely a decent one.
: Actually, if the people at IBM can produce a computer that is even
: barely competent at Bridge, I will be impressed.

People like you will, when computers finally beat everybody at Bridge
(or Game X), retreat and say, "yeah, but can it write poetry?"

Years ago, people just like you determined that chess-playing as
an activity that required intelligence. People like you drove the
development of Deep Blue and all its little brothers - and now you
want it to play Bridge.

And if Bridge is solved again in some years with "Brute Force"?
Where will you retreat to? What will you redefine as the lower
limit of intelligence at that time?

Now, Deep Blue is *not* a revolution. It is at most a milestone -
proving that the human brain, end result of evolution in earth's
environment, is not the best way to approach the problem of chess.
Kasparov has billions of little neurons and is capable of complex
thought, yet I don't hear anybody asking him to carry out chess
calculations like Deep Blue. Deep Blue has 32 processors capable
of high speed calculations - why the hell should Deep Blue try to
think like humans to beat humans? Is it really so hard to accept
that our brain, while good at hunting and gathering and killing
animals, is lousy at chess?

: As for 'Big Blue' beating Kasperov at Chess, this is nothing more that

: using brute force to solve a problem, it is most UNimpressive.

I don't expect Kasparov to calculate chess like Deep Blue - why do you
expect Deep Blue to calculate chess like Kasparov? Is it also "most
unimpressive" that your desk calculator adds faster than you can?
Why do people like you seek to deprive what a machine does best (stupid
but patient calculations) and ask it to think like a human?

Why limit "AI" (whether chess is considered AI or not is beside the
point) to the way humans think? Why not just define it as any set of
functions that *act* intelligently? If chess requires intelligence,
Deep Blue certainly acts intelligently. I don't see any other primates
play chess - do you?

Garry Kasparov has challenged Deep Blue (IBM, naturally) to a "real"
match, on what he considers neutral grounds. To me that sounds like
he is treating Deep Blue like a fellow human grandmaster. He no longer
is willing to stay in a room too cold for his comfort *because* he is
playing a machine. He no longer wants to accept IBM sponsorship
*because* he is playing a machine. He wants to play the machine in
a "fair game". That sounds like RESPECT to me.

[...]
: To me, the 'Big Blue' vs Kasperov match is comparable to racing a human

: on foot, to a remote computer controlled 100 horse power race car, I
: will be impressed if the human manages to beat the moterized computer
: controlled car even once.

It *is*. The hoopla is only because chess is traditionally regarded as
something requiring intelligence - something in the domain of humanity.
We have long accepted that we cannot run faster than many machines and
most animals, but we are *smarter* than them. Now that computers beat
everybody at chess, your kind conveniently redefines chess as not
requiring intelligence or solvable by brute force. In fact, you redefine
it to be Bridge.

How will you face yourself when you find that computers beat you
regularly at Bridge and you still don't feel comfortable acknowledging
that computers are smart? Redefine intelligence yet again?

Monument

unread,

May 27, 1997, 3:00:00 AM5/27/97

to

In comp.ai.alife Dale Pennington <dpen...@ingr.com> wrote:
: Actually, as a friend of mine commented, Big Blue had a major unfair

: advantage. It (actually, it's programming team including at least one chess
: grand master) had plenty of time to analyze Kasparov's playing style and
: determine the weaknesses in his game. It could look at his past history of
: playing, as well as the books he had written. How much do you think
: Kasparov knew of Big Blue's style prior to the match (can you say 0). So
: Big Blue started with a major strategic advantage that no human player in
: that class would ever have.

Point #1: It's DEEP BLUE, not Big Blue.

Point #2: Kasparov knew who the grand master(s) were associated with
the programming team, and was(or should have been) familiar with their
styles and taken that into consideration for Deep Blue's playing
style.

Point #3: Deep Blue would have adapted to Kasparov's playing style
even if it hadn't been fed his books and previous games in a database.
It was, to my knowledge, designed with a NNet/GA combination where
they rebuilt the NN after each game if possible.

Point #4: Kasparov DID know how Deep Blue played: they have played
each other once before.

You were pissing and moaning about a strategic advantage somewhere, I
just can't seem to find it... ;)

--
----------------------------------------------------------------------------
| Jeff Goslin - Monument | "Oh Bentson, you are so |
| jggo...@vela.acs.oakland.edu | mercifully free from the |
| | ravages of intellect." |
| http://www.acs.oakland.edu/~jggoslin | --Evil, The Time Bandits |
----------------------------------------------------------------------------
| how come everyone elses religion is a cult but your cult is a religion |
----------------------------------------------------------------------------

Mario Klebsch DG1AM

unread,

May 27, 1997, 3:00:00 AM5/27/97

to

Richard Pitre <rpi...@ziplink.net> writes:

>I believe that machines will surpass humans in every defineable
>measure of reasoning ability.

Oh, not at all. Playing (is the machine really playing?) chess is only
one point, where the machine is getting better. But what about some
really "simple" taskes, compared to chess.

Almost everyone old enough in the western world can drive a car,
although only a few people driving a car only have the chance to beat
Kasperov in chess. So it must be much easier to do for a computer. But
it isn't. In fact, I naver sa a computer driving a car.

73, Mario
--
Mario Klebsch, DG1AM, M.Kl...@tu-bs.de +49 531 / 391 - 7457
Institut fuer Robotik und Prozessinformatik der TU Braunschweig
Hamburger Strasse 267, 38114 Braunschweig, Germany

Kaz Kylheku

unread,

May 27, 1997, 3:00:00 AM5/27/97

to

In article <5mesma$s...@news1.mnsinc.com>,
Szu-Wen Huang <hu...@mnsinc.com> wrote:

>Kasparov has billions of little neurons and is capable of complex
>thought, yet I don't hear anybody asking him to carry out chess
>calculations like Deep Blue. Deep Blue has 32 processors capable
>of high speed calculations - why the hell should Deep Blue try to

Correction: 256 processors. :) Each of the 32 nodes is coupled to 8 VLSI
``accelerator'' chips which are wired to perform chess related calculations. I
think that a bare bones 32 node RS/6000 box would have a very tough time
against Kasparov. :)

Dale Pennington

unread,

May 27, 1997, 3:00:00 AM5/27/97

to

Monument <jggo...@oakland.edu> wrote in article
<5mf21b$66$1...@news2.acs.oakland.edu>...

> In comp.ai.alife Dale Pennington <dpen...@ingr.com> wrote:
> : Actually, as a friend of mine commented, Big Blue had a major unfair
> : advantage. It (actually, it's programming team including at least one
chess
> : grand master) had plenty of time to analyze Kasparov's playing style
and
> : determine the weaknesses in his game. It could look at his past history
of
> : playing, as well as the books he had written. How much do you think
> : Kasparov knew of Big Blue's style prior to the match (can you say 0).
So
> : Big Blue started with a major strategic advantage that no human player
in
> : that class would ever have.
>
> Point #1: It's DEEP BLUE, not Big Blue.
>

Noted

> Point #2: Kasparov knew who the grand master(s) were associated with
> the programming team, and was(or should have been) familiar with their
> styles and taken that into consideration for Deep Blue's playing
> style.

By that thesis, we should be able to determine every possible grand
master's style in the future from the fact they will train from the current
set of grand masters. I am not sure this is guaranteed.

>
> Point #3: Deep Blue would have adapted to Kasparov's playing style
> even if it hadn't been fed his books and previous games in a database.
> It was, to my knowledge, designed with a NNet/GA combination where
> they rebuilt the NN after each game if possible.
>

True, but it got a large amount of adaption to Kasparov's style before the
first game they played. If it had started with no Kasparov knowledge, then
I think Kasparov would have won this contest hands down, while the computer
developed the database to be a competitor later on.

> Point #4: Kasparov DID know how Deep Blue played: they have played
> each other once before.
>

Kasparov has a few games experience with Deep Blue before the current
reprogramming. What 10 games before a radical change vs 100's of recorded
games.

> You were pissing and moaning about a strategic advantage somewhere, I
> just can't seem to find it... ;)

I still see the strategic advantage. I am saying that Deep Blue winning is
not such a surprising thing, given the advantages it had in knowing its
opponent, while its opponent had very little knowledge on it.

Sort of like the advantage we had in Desert Storm. (only maybe not quite so
extreme).

Richard Ottolini

unread,

May 27, 1997, 3:00:00 AM5/27/97

to

A human has about a trillion neurons, each branching about a thousand,
at about 10 Hz, or one the order 10 quadrillion state transitions per second.
Deep blue has 286 processors at about 10 million transitors apiece (including the
memory) and 300 Mhz or about a trillion state transitions per second,
or a ten thousand times less. Of course perhaps only a percent of the human
circuitry may be working on chess at the time.

Michael LaLena

unread,

May 27, 1997, 3:00:00 AM5/27/97

to

Alicia Carla Longstreet wrote:
>
> When IBM comes up with a computer that can beat me at Bridge, I will be
> impressed. I am NOT a great Bridge player, merely a decent one.
> Actually, if the people at IBM can produce a computer that is even
> barely competent at Bridge, I will be impressed.

Looking at my 1000 page Gorens (sp?) bridge book, 90% of the game is
bidding.

> As for 'Big Blue' beating Kasperov at Chess, this is nothing more that
> using brute force to solve a problem, it is most UNimpressive.

> Actually, I AM imressed that Kasperov was able to beat the machine,
> considering that he could not begin to approach the computational speed
> of the computer.

Actually, Kasperov was upset because Deep Blue did not use strictly
brute
force to win. Last year IBM tried that and failed. Deep Blue made
mistakes
that brute force didn't see. The new version was faster, but also knew
how
to play chess, and had been training with a Grand Master for a year.
The question still remains, will brute force, without any knowledge of
the problem, beat a human?
Spending years writing an intelligent chess program is fine, but
when you can develop a computer fast enough that it can win at chess
with a 100 line program that only contains the rules of chess and was
written in 1 day, then you have accomplished something.
At that point, without the help of humans, the computers are smarter.

> To me, the 'Big Blue' vs Kasperov match is comparable to racing a human
> on foot, to a remote computer controlled 100 horse power race car, I
> will be impressed if the human manages to beat the moterized computer
> controlled car even once.

PS: Robinson still beat Deep Blue in basketball.
(IBM Commerical)
--
Mike LaLena
Email: mailto:mla...@geocities.com
WWW: http://www.geocities.com/SiliconValley/Park/4456/

Sajid Ahmed the peaceman

unread,

May 27, 1997, 3:00:00 AM5/27/97

to

H. M. Hubey wrote:
> Here is a simple example of a task which can be done in
> various ways. The fastest in-place sorting algorithm works
> in O(n*ln(n)). That is because there is a space restriction.
>
> Here is an algorithm that will work in linear time (given
> unlimited space). For concreteness assume that the data
> consists of integers. Simply run thorugh the data once
> and determine the min and max numbers. Then make a 2-D
> array, indexed from min to max. Then run through the data
> and put each data in its own bucket. That's linear. IF we
> had multiple processors we could divide it up into M buckets
> and then we'd need O(N/M) time. If we had as many processors
> as the number of data (i..e N pieces) and hard perfect
> parallelism could sort it in place in constant time.
>

Actually, that's O(1) time.

Peaceman

Lawrence Kirby

unread,

May 27, 1997, 3:00:00 AM5/27/97

to

In article <5mf21b$66$1...@news2.acs.oakland.edu>
jggo...@oakland.edu "Monument" writes:

>Point #4: Kasparov DID know how Deep Blue played: they have played
>each other once before.

Kasparov played a different (although related) program before, or was he
given examples of games played by the current program before the match?

--
-----------------------------------------
Lawrence Kirby | fr...@genesis.demon.co.uk
Wilts, England | 7073...@compuserve.com
-----------------------------------------

Monument

unread,

May 27, 1997, 3:00:00 AM5/27/97

to

In comp.ai.alife Dale Pennington <dpen...@ingr.com> wrote:

: Monument <jggo...@oakland.edu> wrote in article
: > Point #1: It's DEEP BLUE, not Big Blue.
: Noted

AH THE SUPERIORITY OF MY VAST INTELLECT IS ASTOUNDING BWAHAHAH!!!!

....ahem... uh sorry. ;)

: > Point #2: Kasparov knew who the grand master(s) were associated with

: > the programming team, and was(or should have been) familiar with their
: > styles and taken that into consideration for Deep Blue's playing
: > style.
: By that thesis, we should be able to determine every possible grand
: master's style in the future from the fact they will train from the current
: set of grand masters. I am not sure this is guaranteed.

No, but Kasparov, being the intelligent person that he assumedly is,
should have been able to discern that Deep Blue would most likely
follow a style of play similar to the people who taught it.

For example, if one is taught that the best defence is a good offense,
and one's enemy knows that, the enemy should expect that the
opposition will, MOST LIKELY, go on the offensive.

: > Point #3: Deep Blue would have adapted to Kasparov's playing style

: > even if it hadn't been fed his books and previous games in a database.
: > It was, to my knowledge, designed with a NNet/GA combination where
: > they rebuilt the NN after each game if possible.
: True, but it got a large amount of adaption to Kasparov's style before the
: first game they played. If it had started with no Kasparov knowledge, then
: I think Kasparov would have won this contest hands down, while the computer
: developed the database to be a competitor later on.

Kasparov is the reigning world champion at chess, is he not? Let us
assume for the moment that I am some chess child prodigy that has
never played a tournament game in my life, and yet somehow I manage to
finangle a game with the champ. He knows nothing about my playing
style, yet, I have undoubtedly heard of and studied him. Does that
give me an unfair advantage simply because I know my opponent?

: > Point #4: Kasparov DID know how Deep Blue played: they have played
: > each other once before.
: Kasparov has a few games experience with Deep Blue before the current

: reprogramming. What 10 games before a radical change vs 100's of recorded
: games.

It's just the same as if someone does a little bit of training for a
boxing match, gets their ass beat on severely, and then actually
trains for the followup match.

: I still see the strategic advantage. I am saying that Deep Blue winning is

: not such a surprising thing, given the advantages it had in knowing its
: opponent, while its opponent had very little knowledge on it.
: Sort of like the advantage we had in Desert Storm. (only maybe not quite so
: extreme).

Hopefully Deep Blue won't carpet bomb the tournament site into the
stone age. ;)

Sajid Ahmed the peaceman

unread,

May 27, 1997, 3:00:00 AM5/27/97

to

Actually, that O(1) time.

Peaceman

Richard Pitre

unread,

May 27, 1997, 3:00:00 AM5/27/97

to

Mario Klebsch DG1AM wrote:
>
> Richard Pitre <rpi...@ziplink.net> writes:
>
> >I believe that machines will surpass humans in every defineable
> >measure of reasoning ability.
>
> Oh, not at all. Playing (is the machine really playing?) chess is only
> one point, where the machine is getting better. But what about some
> really "simple" taskes, compared to chess.
>
> Almost everyone old enough in the western world can drive a car,
> although only a few people driving a car only have the chance to beat
> Kasperov in chess. So it must be much easier to do for a computer. But
> it isn't. In fact, I naver sa a computer driving a car.
>

Please note that I used the word *will* not *can*.
I also used the word *believe* because nothing but the
passage of time will prove the point. While the chess thing really
is a minor milestone it has certainly stimulated some healthy
speculation about what the future holds.

There is nothing in the past to support the notion that computers will
never
be substantially more capable than they are today so I don't understand
the
point about how people can drive cars today. Computers will eventually
be able to drive cars but by that time we might not care if they can or
cannot
since, by that time, we might do most of our traveling in a virtual
reality suit.

The rest of my post was an attempt to put the significance of
computer *intelligence* in some context that wasn't so threatening
to people who feel that being *intelligent* is a large
and important part of being human. In time we will be able to buy
intelligence. Things like motive and intent will
always be free and unavoidable.

richard

ccr...@pacific.net

unread,

May 27, 1997, 3:00:00 AM5/27/97

to

In <5mesma$s...@news1.mnsinc.com>, on 05/27/97
at 02:56 PM, hu...@mnsinc.com (Szu-Wen Huang) said:

:How will you face yourself when you find that computers beat you

:regularly at Bridge and you still don't feel comfortable acknowledging
:that computers are smart? Redefine intelligence yet again?

Actually, I think that I could write a bridge playing program that would
beat Alicia regularly, given that her self assessment is correct. I played
tournament level duplicate bridge for a while, and I know that I can write
a progam which can beat me -- based primarily on the fact that the
computer has perfect recall of all cards played in each hand, including
who played them., and in what order. Top bridge players can do this, of
course, but most of us have to settle for just keeping track of what cards
are still unplayed.

The primary reason why chess has has long been considered the measure of
intelligence, rather than bridge, is that chess is a game of perfect
information, and bridge is not. In addition, bridge is a partnership game.
Who would the computer's partner be in such a challenge? Who would
Alicia's partner be?

Best,
-- Chuck Crayne
-----------------------------------------------------------
ccr...@pacific.net
-----------------------------------------------------------

Ralph Silverman

unread,

May 27, 1997, 3:00:00 AM5/27/97

to

Walter (wal...@and.nl) wrote:
: Paul Campbell wrote:
: >

: > > Alicia Carla Longstreet wrote:
: > > >
: > > > When IBM comes up with a computer that can beat me at Bridge, I will be
: > > > impressed. I am NOT a great Bridge player, merely a decent one.
: > > > Actually, if the people at IBM can produce a computer that is even
: > > > barely competent at Bridge, I will be impressed.

: > > >
: > > > As for 'Big Blue' beating Kasperov at Chess, this is nothing more that

: > > > using brute force to solve a problem, it is most UNimpressive.

: >
: > Agreed, no more impressive than the fact that my pocket calculator is

: > better than me (and presumably its programmers) at arithmetic.
: >
: > Paul C.
: > UK.

: (Just my opinion..)

: I think it is pretty impressive. Playing chess is something else than
: calculating 1+1. Kasparov seemed very surprised by Deep Blue's
: performance and he was very angry that he lost... :)

: However, I also think Kasparov could beat Deep Blue in a next game
: just the same. Deep Blue may not have won the war yet.

: --
: *** When you build it, they will come ***

: wal...@ronix.ptf.hro.nl Go to BrintaBBS: telnet://bbs.hro.nl
: wal...@and.nl where I am known as 'Walter'
: http://www.ptf.hro.nl/~walter/

deep blue is 32 node parallel system
...
similar systems >100 node are documented
...
if necessary ...
technology could, evidently scale up !

<HTML>
<HEAD>
<TITLE>How Deep Blue Works</TITLE>
</HEAD>

<IMG SRC=/meet/pics/tabbar.gif BORDER=0 WIDTH=600 HEIGHT=24 ALIGN=TOP>

To date, Deep Blue is the most powerful chess-playing
computer ever developed. But what makes Deep Blue so great at playing chess?
How does it so accurately "choose" its next move from a list of thousands of
possible options? And why is it so much better than other computer chess
'players'? The answer lies in its unique combination of innovate software
engineering and massive parallel processing power.

At the heart of Deep Blue's ability to play chess is its evaluation function.
The evaluation function is an algorithm that measures the "goodness" of a
given chess position. Positions with positive values are good for White, and
conversely, positions with negative values are good for Black. If the overall
score is negative, for example, this means that Black has the advantage.

Deep Blue's evaluation function looks at four basic chess
values: material, position, King safety and tempo. Material is based on
the "worth" of particular chess pieces. For example, if a pawn is valued at
1, then the rook is worth 5 and the Queen is valued at 9. The King, of
course, is beyond value because his loss means the loss of the game.

The simplest way to understand position is by looking at your pieces and
counting the number of safe squares they can attack. King safety is a
defensive aspect of position. It is determined by assigning a value to the
safety of the King's position in order to know how to make a purely defensive
move. Tempo is related to position but focuses on the race to develop control
of the board. A player is said to "lose a tempo" if he dillydallies while the
opponent is making more productive advances.

Deep Blue is not only the finest chess-playing computer in the world, it is
also the fastest. This makes perfect sense, because history has proven that
the fastest computers conduct the most extensive searches into possible
positions. More searches gives the computer a wider array of moves to choose
from and therefore a greater chance of choosing the optimum move.

Deep Blue employs a system called selective extensions to examine chessboard
positions. Selective extensions allow the computer to more efficiently search
deeply into critical board arrangements. Instead of attempting to conduct an
exhaustive "brute force" search into every possible position, Deep Blue
selectively chooses distinct paths to follow, eliminating irrelevant searches
in the process.

Deep Blue uses "live" software that can actually generate up to 200,000,000
positions per second when searching for the optimum move. The software
begins this process by taking a strategic look at the board. It then computes
everything it knows about the current position, integrates the chess
information pre-programmed by the development team, and then generates a
multitude of new possible arrangements. From these, it then chooses its best
possible next move.

Deep Blue's extensive searches make full use of the computer's massively
parallel design. "At the search level, you're saying 'OK, here's the
position. I need to search all the moves," says Joe Hoane, the Deep Blue
development team member in charge of software. "And you go search all the
moves, all at the same time, preferably on a bunch of different computers."

The software inside of Deep Blue is one all-inclusive program written in C,
running under the AIX operating system. Deep Blue utilizes the IBM SP
Parallel System called MPI. "It's a message-passing system," says Hoane. "So
the search is just all control logic. You're passing control messages back
and forth that say, well, what am I doing? Did you finish this? OK, here's
your next job. That kind of thing at the SP level."

The latest iteration of the Deep Blue computer is a 32-node <A
HREF=http://www.rs6000.ibm.com> IBM RS/6000</A> SP high-performance computer,
which utilizes the new Power Two Super Chip processors (P2SC). Each node of
the SP employs a single microchannel card containing 8 dedicated VLSI chess
processors, for a total of 256 processors working in tandem. The net result
is a scalable, highly parallel system capable of calculating 60 billion
moves within three minutes, which is the time allotted to each player's move
in classical chess.


</TD>

<A HREF=d.3.3.html><IMG WIDTH=6 HEIGHT=9 BORDER=0 ALT=next SRC=../../pics/b.arrow_right.gif> Deep Blue FAQ</A> 
The answers to the questions about this powerful chess-playing computer 

<A HREF=d.3.1.html><IMG WIDTH=6 HEIGHT=9 BORDER=0 ALT=next SRC=../../pics/b.arrow_right.gif> The making of Deep Blue</A> 
A timeline of Deep Blue's development 

How Deep Blue works 
Under the hood of this powerful parallel processor 

<A HREF=d.3.4.html><IMG WIDTH=6 HEIGHT=9 BORDER=0 ALT=next SRC=../../pics/b.arrow_right.gif> All this power just for chess?</A> 
How Deep Blue technology is affecting the way we live 

<A HREF=d.4.html><IMG BORDER=0 ALT="The Team" VSPACE=5 WIDTH=190 HEIGHT=13 SRC=../pics/b_team.gif></A> 
<A HREF=d.2.html><IMG BORDER=0 ALT="The Comparison" VSPACE=5 WIDTH=190 HEIGHT=13 SRC=../pics/b_comp.gif></A>



"You have to be on full guard every move of the game, which means it is more exhausting. I think Deep Blue is stronger than all but a handful of top human players." - Garry Kasparov
 
<A HREF=d.1.html><IMG WIDTH=6 HEIGHT=9 BORDER=0 ALT=next SRC=../../pics/b.arrow_right.gif> meet the players</A>

<A HREF=/reference/html/i.3.a.html>
<A HREF=/reference/html/i.3.a.html>
<IMG WIDTH=190 HEIGHT=17 BORDER=0 VSPACE=5 HSPACE=-3 ALT=chess pieces SRC=../../pics/heading_chess.gif></A></A> 
no. 58 

The first movie about chess was CHESS FEVER, made in Moscow in 1925 and starring Jose Capablanca.

<IMG WIDTH=8 HEIGHT=8 SRC=../../pics/bullet_chess.gif>

</TD>

</TR>

<TR>
<TD WIDTH=380 COLSPAN=2> 
 </TD>
</TR>

</TABLE>

<A HREF=/ssi/4bnav.map>
<IMG SRC="/pics/bnavbar.gif" BORDER=0 WIDTH=600 HEIGHT=49 ALIGN=LEFT ISMAP USEMAP="#bnavbar">
</A>
<MAP NAME=bnavbar>
<AREA SHAPE=RECT COORDS="6,0,94,18" HREF="/press/html/g.html" ALT="Press room" onMouseOver="self.status='Press room' ;return true">
<AREA SHAPE=RECT COORDS="98,0,208,18" HREF="/reference/html/i.html" ALT="Reference" onMouseOver="self.status='Reference' ;return true">
<AREA SHAPE=RECT COORDS="215,0,292,17" HREF="/feedback/html/k.html" ALT="feedback" onMouseOver="self.status='Feedback' ;return true">
<AREA SHAPE=RECT COORDS="294,0,378,19" HREF="/guide/html/j.1.html" ALT="Site guide" onMouseOver="self.status='Site guide' ;return true">
<AREA SHAPE=RECT COORDS="506,9,591,41" HREF="http://www.ibm.com/" ALT="IBM home" onMouseOver="self.status='IBM home' ;return true">
</MAP>
 
 
 
<TABLE ALIGN=LEFT WIDTH=600>
<TR>
<TD>
<center>IBM creates and hosts this web site, which is powered by an IBM <A HREF="http://www.rs6000.IBM.com/">RS/6000 SP</A> - the same technology that's behind Deep Blue. 
Kasparov vs. Deep Blue: the rematch is
under the auspices of <A HREF="http://www.acm.org/">ACM</a>
 
*Java is a registered trademark of Sun Microsystems, Inc.
 
<A HREF="http://www.IBM.com/Legal/">© Copyright IBM Corporation 1997. All rights reserved.</A>
</center>
</TD>
</TR>
</TABLE>

</BODY>
</HTML>

<IMG SRC="/pics/infobar_may11.gif" BORDER=0 WIDTH=600 HEIGHT=25 ALIGN=TOP > 


<A HREF=/ssi/tnavbar.map><IMG WIDTH=600 HEIGHT=67 SRC=/learn/pics/top.gif BORDER=0 ALIGN=BOTTOM USEMAP=#tnavbar ISMAP></A> 
<MAP NAME=tnavbar>
<AREA SHAPE=RECT COORDS=4,0,58,16 HREF=/home/html/b.html ALT=Home onMouseOver="self.status='Kasparov home' ;return true">
<AREA SHAPE=RECT COORDS=64,0,141,16 HREF=/watch/html/c.html ALT=Match onMouseOver="self.status='Watch the match';return true">
<AREA SHAPE=RECT COORDS=146,0,230,16 HREF=/meet/html/d.html ALT=Players onMouseOver="self.status='Meet the players';return true">
<AREA SHAPE=RECT COORDS=235,0,341,17 HREF=/learn/html/e.html ALT=Learn onMouseOver="self.status='Learn about the technology';return true">
<AREA SHAPE=RECT COORDS=347,0,425,16 HREF=http://chess.minds.com/ ALT=Later onMouseOver="self.status='Community';return true">
</MAP>

<TD WIDTH=22 ROWSPAN=1>     

</TD>


<TD VALIGN=TOP WIDTH=358><IMG WIDTH=358 HEIGHT=129 ALT="CERN" SRC=../pics/cnasa.gif>

The SPs at NASA research labs help tackle Grand Challenge 

The Numerical Aerodynamics Simulation facility at NASA's Ames Research Center in Moffett Field, Calif., has a world-class supercomputing capability, accessible to the nation's aeronautical researchers in government, industry and academia.

The facility seeks to create a highly parallel computing environment for solving critical path problems, ones for which a supercomputer will yield an order-of-magnitude improvement in performance -- thereby reducing and eliminating obstacles for aircraft
designers and manufacturers. Its objectives include acting as a pathfinder in advanced, large-scale computing and providing a national computational capability to help ensure continued U.S. leadership in computational fluid dynamics and related aerospace
disciplines.

Led by the Numerical Aerodynamics Simulation facility, NASA's research centers established the Computational AeroSciences (CAS) program to address NASA's Grand Challenge in Aeronautics -- to create an environment in which a complete aerospace vehicle syst
em can be simulated within a computing time ranging from one to several hours.

To bring the agency a step closer to meeting that challenge -- and to exploit the benefits of scalable parallel computing for competitive advantage in a global economy -- NASA awarded a three-year, $22 billion CAS cooperative research agreement to an IBM-
led consortium. As part of the agreement, IBM has installed three Scalable POWERparallel Systems (SP2) as the hardware test beds for the research -- the first of them the 160-node SP2 at the NASA Ames Research Center.

Cooperative research 
"We were interested in obtaining a parallel computer that had the best performance available for the dollar," says Thomas Lasinski, chief for the data analysis branch at the Numerical Aerodynamics Simulation facility. "NASA released a request for proposal
s that was broadly advertised, and there was a lot of interest in it. We received ten proposals, which a team of eight evaluated against a set of performance benchmarks. The decision, made on the basis of those numbers, was that the SP2 offered us the bes
t price/performance."

Agreements like the one NASA signed with the IBM-led consortium are designed to foster cooperative research between industry and government labs by offering private firms advantageous rights to patents and other intellectual property from the joint resear
ch. The consortium includes Boeing Computer Services, Rensselaer Polytechnic Institute, Lockheed Missiles and Space Research, Centric Engineering Systems, Intelligent Aerodynamics and Rice University. Members will support NASA's investment in the agreemen
t with equipment, research and facilities that will significantly increase the resources available.

"We found other features interesting as well," Lasinski continues. "The SP2 is based on workstation technology, for example, and a lot of our customers in the aerospace industry have large clusters of workstations. We're impressed with the RS/6000 Model 5
90 architecture, which has very good memory access. It was also important that the machine chosen be expandable to a reasonably large number of processors. And the SP2 can be updated with new RISC processors in a very timely and economical manner. But mai
nly we saw a machine that was competitive with the largest supercomputer NASA had -- the Cray C90 -- for a third of the price."

Benchmark winner 
The SP2 at NASA Ames has 160 RS/6000 590 nodes, with 128MB of memory on most of them, and 512MB on several. The configuration is equipped with IBM's High Performance Switch, along with High Performance Parallel Interface (HIPPI) connections and FDDI fiber
optic capability.

Parallel Systems Manager Toby Harness: "Important to us -- and increasingly to the parallel computing community -- are the benchmarks developed here in an attempt to come up with some objective measures of performance. Those benchmarks consist of five ker
nels and three synthetic computational fluid dynamics applications, which are the ones of most interest to us. In addition, we developed for this acquisition three additional I/O benchmarks -- network, disk and peak I/O. The SP2's performance in the I/O s
tudies was extremely strong, and it came in better than our estimates on almost all the benchmarks."

IBM's T.J. Watson Research Center -- responsible for many of the SP2's hardware and system software technologies -- provided key contributions to the benchmarking efforts and will participate as a partner in the consortium.

"It took us about a month to get through the initial software installation and acceptance tests, and we had users taking advantage of the system while we were still figuring out how to tune it," says Charles Niggley, senior computer scientist of the paral
lel systems support group. "I was surprised at how well the High Performance Switch performed right from the start. We figured there'd be several, possibly major, hardware problems -- but we didn't encounter any. Compared to massively parallel systems pre
viously installed, we're still amazed at how fast the SP2 came up, the reliability it's had to date and how quickly we're getting productive use out of the system."

Major payoffs 
Numerical Aerodynamics Simulation technology offers major payoffs to the aeronautics industry in reducing cycle time and design cost and enabling designs otherwise not achievable. Using computational fluid dynamics, for example – the foundation for the wo
rk done at the facility -- led to a design change that resulted in millions of dollars of fuel saved by a commercial airline.

"Solving the coupled, partial differential equations that govern how fluid flows over an arbitrary body -- like a wing or a rotor blade -- is very computer intensive and very memory intensive. We use every byte of memory available on the SP2," points out
Timothy Barth of the advanced algorithms and applications branch.

"It's the first machine that's really put us in the zone where we can think about understanding flow physics and how to design better high lift geometries. It's really changed the scale of the problems we can work on -- algorithms that parallelize quite w
ell on a machine such as the SP2."

This page illustrates how one customer uses IBM products. Many factors have contributed to the results and benefits described. IBM does not guarantee comparable results. All information contained herein was provided by the featured customer
and IBM Business Partners. IBM does not attest to its accuracy.


</TD>

<A HREF=e.1.html><IMG WIDTH=6 HEIGHT=9 BORDER=0 ALT=next SRC=../../pics/b.arrow_right.gif> Cern</A> 

<A HREF=e.2.html><IMG WIDTH=6 HEIGHT=9 BORDER=0 ALT=next SRC=../../pics/b.arrow_right.gif> 96 Summer Olympics</A> 

<A HREF=e.3.html><IMG WIDTH=6 HEIGHT=9 BORDER=0 ALT=next SRC=../../pics/b.arrow_right.gif> MHPCC</A> 

   NASA 

<A HREF=e.5.html><IMG WIDTH=6 HEIGHT=9 BORDER=0 ALT=next SRC=../../pics/b.arrow_right.gif> Charles Schwab</A>

<A HREF=e.html><IMG BORDER=0 ALT="Guest essays" VSPACE=5 WIDTH=190 HEIGHT=13
SRC=../pics/guestbuta.gif></A> 

<A HREF=e.6.html><IMG BORDER=0 ALT="Next Horizons" VSPACE=5 WIDTH=190 HEIGHT=13
SRC=../pics/nextbuta.gif></A>


"It's the first machine that's really put us in the zone where we can think about understanding flow physics and how to design better high lift geometries." - NASA's Timothy Barth on the technology behind Deep Blue 
<A HREF=../../meet/html/d.3.html><IMG WIDTH=6 HEIGHT=9 BORDER=0 SRC=../../pics/b.arrow_right.gif> meet the players</A>




<IMG VSPACE=5 ALT="chess pieces" WIDTH=190 HEIGHT=17 SRC=../../pics/heading_chess.gif>
no. 8 

Although far from an expert chess player, the exiled Lenin was so preoccupied with correspondence chess that he often rattled on about it in his sleep.
<IMG WIDTH=8 HEIGHT=8 SRC=../../pics/bullet_chess.gif >

</TD>

</TR>

<TR>
<TD WIDTH=380 COLSPAN=2> 

</TD>
</TR>

</TABLE>

<A HREF=/ssi/4bnav.map>
<IMG SRC="/pics/bnavbar.gif" BORDER=0 WIDTH=600 HEIGHT=49 ALIGN=LEFT ISMAP USEMAP="#bnavbar">
</A>
<MAP NAME=bnavbar>
<AREA SHAPE=RECT COORDS="6,0,94,18" HREF="/press/html/g.html" ALT="Press room" onMouseOver="self.status='Press room' ;return true">
<AREA SHAPE=RECT COORDS="98,0,208,18" HREF="/reference/html/i.html" ALT="Reference" onMouseOver="self.status='Reference' ;return true">
<AREA SHAPE=RECT COORDS="215,0,292,17" HREF="/feedback/html/k.html" ALT="feedback" onMouseOver="self.status='Feedback' ;return true">
<AREA SHAPE=RECT COORDS="294,0,378,19" HREF="/guide/html/j.1.html" ALT="Site guide" onMouseOver="self.status='Site guide' ;return true">
<AREA SHAPE=RECT COORDS="506,9,591,41" HREF="http://www.ibm.com/" ALT="IBM home" onMouseOver="self.status='IBM home' ;return true">
</MAP>
 
 
 
<TABLE ALIGN=LEFT WIDTH=600>
<TR>
<TD>
<center>IBM creates and hosts this web site, which is powered by an IBM <A HREF="http://www.rs6000.IBM.com/">RS/6000 SP</A> - the same technology that's behind Deep Blue. 
Kasparov vs. Deep Blue: the rematch is
under the auspices of <A HREF="http://www.acm.org/">ACM</a>
 
*Java is a registered trademark of Sun Microsystems, Inc.
 
<A HREF="http://www.IBM.com/Legal/">© Copyright IBM Corporation 1997. All rights reserved.</A>
</center>
</TD>
</TR>
</TABLE>

</BODY>
</HTML>

--

Ralph Silverman
z007...@bcfreenet.seflin.lib.fl.us

Paul Hsieh

unread,

May 27, 1997, 3:00:00 AM5/27/97

to

On Tue, 27 May 1997, Michael LaLena mla...@geocities.com said:
> Spending years writing an intelligent chess program is fine, but
> when you can develop a computer fast enough that it can win at chess
> with a 100 line program that only contains the rules of chess and was
> written in 1 day, then you have accomplished something.

Huh? I'd say these IBM researchers have put considerably less effort
into their chess program then Kasparov has in becoming the World's Best
player. Furthermore, where do you get off implying that what they did
was not an amazing accomplishment.

> At that point, without the help of humans, the computers are smarter.

Easy, just write a class in Java to re-use the alreaddy encapsulated Deep
Blue library ... ok, at this point I'm babbling ...

Axel-Stephane Smorgrav

unread,

May 28, 1997, 3:00:00 AM5/28/97

to

In article <mkl.86...@rob.cs.tu-bs.de>, m...@rob.cs.tu-bs.de (Mario Klebsch DG1AM) writes:
|> Almost everyone old enough in the western world can drive a car,
|> although only a few people driving a car only have the chance to beat
|> Kasperov in chess. So it must be much easier to do for a computer. But
|> it isn't. In fact, I naver sa a computer driving a car.

Only a fraction of people driving cars can fly an airplane, and most
will agree that flying (in their opinion) is more difficult than driving
a car. Still, computers that you claim are not able to drive cars, can
control an aircraft and even bring it safely back to the ground...

-ascs

ko besuijen

unread,

May 28, 1997, 3:00:00 AM5/28/97

to

Maybe flying isn't simple enough? :-)

Meaning: there is a flaw in your logic:

the argument is: driving a car is easier than beating Kasparov, but
computers can't drive cars

you say: driving a car is easier than flying an aeroplane, but computers
can fly

So in fact you agree, but you are not aware of it.

Ko Besuijen

Gordon D. Pusch

unread,

May 28, 1997, 3:00:00 AM5/28/97

to

In article <mkl.86...@rob.cs.tu-bs.de> m...@rob.cs.tu-bs.de
(Mario Klebsch DG1AM) writes:
[snip]

> Almost everyone old enough in the western world can drive a car,
> although only a few people driving a car only have the chance to
> beat Kasperov in chess. So it must be much easier to do for a
> computer. But it isn't. In fact, I naver sa a computer driving a car.

Sorry, but you're a bit out of touch --- the US military has already
developed a truck that drives itself on standard marked roadways using
visual cues and pattern-recognition techniques (albeit rather slowly);
the US auto-companies are likewise researching this area...

The problem is not the relatively low level of *reasoning* required to
drive a car --- it's the ability to sort out the _visual_ cues involved...

-- Gordon D. Pusch <pu...@mcs.anl.gov>

Disclaimer: I'm a consultant collaborating with Argonne researchers;
I don't speak for ANL or the DOE --- and they *certainly* don't speak
for =ME= !!!

Jim Webber

unread,

May 28, 1997, 3:00:00 AM5/28/97

to

The "impressiveness" of Deep Blue's victory lies with the fact that it's
alleged "Brute Force" approach is computed so quickly.

Whilst the kind of algorithms used in computer chess games have been in use
for some length of time, I would imgaine that highly parallel versions of
these algorithms remain a research topic.

Now imagine giving the program a (possibly parallel) knowledge base too.
That is a serious software project. Perhaps people should consider the
metrics of such an undertaking before deciding that the Deep Blue victory
is worthless.

Jim Webber
Parallelism Research Group
University of Newcastle upon Tyne, UK

R!ch

unread,

May 28, 1997, 3:00:00 AM5/28/97

to Mario Klebsch DG1AM

On 27 May 1997, Mario Klebsch DG1AM wrote:

> Almost everyone old enough in the western world can drive a car,
> although only a few people driving a car only have the chance to beat
> Kasperov in chess. So it must be much easier to do for a computer. But
> it isn't. In fact, I naver sa a computer driving a car.

You mean you've never seen Kitt in Knight Rider? :-)

--
R!ch

Christian Bau

unread,

May 28, 1997, 3:00:00 AM5/28/97

to

In article <01bc6aac$1dc992c0$598e...@dpenning.b24domain.ingr.com>, "Dale
Pennington" <dpen...@ingr.com> wrote:

> Actually, as a friend of mine commented, Big Blue had a major unfair
> advantage. It (actually, it's programming team including at least one chess
> grand master) had plenty of time to analyze Kasparov's playing style and
> determine the weaknesses in his game. It could look at his past history of
> playing, as well as the books he had written. How much do you think
> Kasparov knew of Big Blue's style prior to the match (can you say 0). So
> Big Blue started with a major strategic advantage that no human player in
> that class would ever have.

I remember from the previous tournament what Kasparov said why he managed
to beat Deep Blue except in the first game: He said that Deep Blues ELO
strength was very different depending on the situation. So instead of
trying to do the best moves, he would try to use moves that would lead to
a situation where Deep Blue didnt play well.

It seems that Deep Blue did not try to avoid positions where it played
badly. The obvious thing to do for the programmers is to change Deep Blues
valuation function to give a lower rating to positions where the computer
plays less good.

I am wondering (and I have never seen this discussed) how much the
computers playing "style" is determined by the bit of code that computes
the value of each position. If this has a lot of effect, it might be that
different valuation functions would produce different, but equally good
styles. And if that is the case, I wonder if Deep Blue could just switch
between different playing styles at any time in the middle of a game. That
would be something that could be quite upsetting for the human opponent.

So there is the possibility that Deep Blue doesnt have any fixed "style",
that it can randomly pick a different style for each game it plays. If
that is the case, it would be an advantage, but not an unfair advantage.

-- For email responses, please remove the last emm from my address.
-- For spams, please send them to whereever...@but.not.here

Jonathan Kirwan

unread,

May 28, 1997, 3:00:00 AM5/28/97

to

On 27 May 1997 21:11:39 GMT, jggo...@oakland.edu (Monument)
wrote:

>No, but Kasparov, being the intelligent person that he assumedly is,
>should have been able to discern that Deep Blue would most likely
>follow a style of play similar to the people who taught it.

Actually, they "randomized" the play before every game. Science
News has an article about this, quoting Deep Blue's manager
(Chung-Jen Tan) as saying that this was one of three improvements
they believe made the difference. This made it more like trying
to play a tournament against a team of players, one after another.

Former US Champion, Pat Wolff, said ".. analysis of the games
shows very clearly that cognitively Kasparov is still the better
chess player. What shocked me and most chess players who followed
the match was how Kasparov simply fell apart at the end."

Don't make more of this than it is, just yet.

Jon

Markus Becker

unread,

May 28, 1997, 3:00:00 AM5/28/97

to

Have to jump in here...

Peter Harrison wrote:

> >Alicia Carla Longstreet <ca...@ici.net> writes:
> >
> >>When IBM comes up with a computer that can beat me at Bridge, I will be
> >>impressed. I am NOT a great Bridge player, merely a decent one.

> A human DOES NOT 'look ahead' at all, they develop stratagies, and use
> previous experience to apply to the current situation - even though
> that situation is not identical.

Seems like you have never played one game of chess. The "average"
human chess player looks ahead approximately 4 to 5 moves. In an
end game sometimes I (being a merely average human player) look
ahead up to 15 moves.

> However, with Bridge a computer cannot use brute force 'looking
> forward' in order to calculate the best move. It requires something
> more.

What does it require more? Please elaborate. Card counting? I think
every pocket calculator can do this better than any human.

Markus

Patrick Juola

unread,

May 28, 1997, 3:00:00 AM5/28/97

to

In article <338C1A...@zess.uni-siegen.de> bec...@zess.uni-siegen.de writes:
>Have to jump in here...

>Seems like you have never played one game of chess. The "average"
>human chess player looks ahead approximately 4 to 5 moves. In an
>end game sometimes I (being a merely average human player) look
>ahead up to 15 moves.
>
>> However, with Bridge a computer cannot use brute force 'looking
>> forward' in order to calculate the best move. It requires something
>> more.
>
>What does it require more? Please elaborate. Card counting? I think
>every pocket calculator can do this better than any human.

Chess is a perfect-information game, meaning that everything that
you know is also known to your opponent, and vice versa. Bridge
is not -- at least two hands are unknown to you, and your hand is
not known to anyone else.

Similarly, your bidding strategy may not be known to your opponent
(and vice versa).

This means, among other things, that it's possible to bluff.

-kitten

Markus Becker

unread,

May 28, 1997, 3:00:00 AM5/28/97

to

Patrick Juola wrote:

> >What does it require more? Please elaborate. Card counting? I think
> >every pocket calculator can do this better than any human.

> This means, among other things, that it's possible to bluff.

I have no problems writing a program that bluffs. Even Microsoft
can do that! ;-)

M.

Patrick Juola

unread,

May 28, 1997, 3:00:00 AM5/28/97

to

Yes, but can you write a program that bluffs *well* -- or that
detects bluffing?

-kitten

Casper H.S. Dik - Network Security Engineer

unread,

May 28, 1997, 3:00:00 AM5/28/97

to

pat...@gryphon.psych.ox.ac.uk (Patrick Juola) writes:

>Chess is a perfect-information game, meaning that everything that
>you know is also known to your opponent, and vice versa. Bridge
>is not -- at least two hands are unknown to you, and your hand is
>not known to anyone else.

>Similarly, your bidding strategy may not be known to your opponent
>(and vice versa).

That's not correct; the bidding strategy used must be made known to
your opponents. You can't tell your partner anything about your cards
that your opponent can't understand.

>This means, among other things, that it's possible to bluff.

Not really.

Casper
--
Expressed in this posting are my opinions. They are in no way related
to opinions held by my employer, Sun Microsystems.
Statements on Sun products included here are not gospel and may
be fiction rather than truth.

Dale Pennington

unread,

May 28, 1997, 3:00:00 AM5/28/97

to

Monument <jggo...@oakland.edu> wrote in article

<5mfimb$78j$1...@news2.acs.oakland.edu>...

> : > Point #4: Kasparov DID know how Deep Blue played: they have played
> : > each other once before.
> : Kasparov has a few games experience with Deep Blue before the current
> : reprogramming. What 10 games before a radical change vs 100's of
recorded
> : games.
>
> It's just the same as if someone does a little bit of training for a
> boxing match, gets their ass beat on severely, and then actually
> trains for the followup match.

So Kasparov planned to get beat in this match, so he could play better
later ?

>
> : I still see the strategic advantage. I am saying that Deep Blue winning
is
> : not such a surprising thing, given the advantages it had in knowing its
> : opponent, while its opponent had very little knowledge on it.
> : Sort of like the advantage we had in Desert Storm. (only maybe not
quite so
> : extreme).
>
> Hopefully Deep Blue won't carpet bomb the tournament site into the
> stone age. ;)

You never know what they have those machines hooked to these days :)

David Welton

unread,

May 28, 1997, 3:00:00 AM5/28/97

to

Do you suppose Big or Deep Blue or whatever is intelligent enough not
to cross post to unrelated groups?:P
--
David Welton
dav...@efn.org dav...@freenet.hut.fi http://www.efn.org/~davidw
Se quest'email e` in Italiano, mi dispiace per gli errori:-) FORZA PANTANI!
--Debian GNU/Linux--

Monument

unread,

May 28, 1997, 3:00:00 AM5/28/97

to

In comp.ai.alife Dale Pennington <dpen...@ingr.com> wrote:

: Monument <jggo...@oakland.edu> wrote in article
: > It's just the same as if someone does a little bit of training for a

: > boxing match, gets their ass beat on severely, and then actually
: > trains for the followup match.
: So Kasparov planned to get beat in this match, so he could play better
: later ?

No, Deep Blue got beaten when unprepared to challenge effectively. IT
was trained to win, after then first ass-wupping it received.

: > Hopefully Deep Blue won't carpet bomb the tournament site into the

: > stone age. ;)
: You never know what they have those machines hooked to these days :)

I just can't get "James Cameron"-esque about Deep Blue... ;)

Jim Balter

unread,

May 28, 1997, 3:00:00 AM5/28/97

to

Peter Harrison wrote:

> A human DOES NOT 'look ahead' at all,

[...]

> However, with Bridge a computer cannot use brute force 'looking
> forward' in order to calculate the best move. It requires something
> more.

This thread surpasses the usual crap seen in comp.ai.philosophy
(where I am reading this; I have removed the idiotic crossposts to
the programming groups) in that not only do most of the correspondents
know nothing about AI, but they know nothing about chess as well.

--
<J Q B>

Paul Hsieh

unread,

May 28, 1997, 3:00:00 AM5/28/97

to

On Thu, 29 May 1997 09:42:43 GMT, Peter Harrison (p...@ihug.co.nz) said:
> A human DOES NOT 'look ahead' at all, they develop stratagies, and use
> previous experience to apply to the current situation - even though
> that situation is not identical.

Excuse me? I'm a human, I play chess at an acceptable level. *I* look
ahead. All my chess playing friends all look ahead. The strong the
player, it seems to me, the more they look ahead! (Of course the more
they know, and use their experience too.)

> However, with Bridge a computer cannot use brute force 'looking
> forward' in order to calculate the best move.

That is debatable.

> [...] It requires something more.

A probablistically based search tree. Its just a generalization.

Peter Harrison

unread,

May 29, 1997, 3:00:00 AM5/29/97

to

On 25 May 1997 22:22:25 -0400, hu...@pegasus.montclair.edu (H. M.
Hubey) wrote:

>Alicia Carla Longstreet <ca...@ici.net> writes:
>
>>When IBM comes up with a computer that can beat me at Bridge, I will be
>>impressed. I am NOT a great Bridge player, merely a decent one.
>

>Most people probably won't be.

I think that the point about Bridge Playing has been missed. The
sterotype of chess is that it requires the quality of 'intelligence'
to play - and of course, win.

Now the situation has occured where the computer beats the most
skillfull human - and this is perceived to be more 'intelligent', but
only because it is able to physically look ahead so many moves.

A human DOES NOT 'look ahead' at all, they develop stratagies, and use
previous experience to apply to the current situation - even though
that situation is not identical.

However, with Bridge a computer cannot use brute force 'looking
forward' in order to calculate the best move. It requires something
more.

The Human Brain and Computer are different. I don't expect that
computers will be act like natural brains for a long time yet.

Peter

Barton Chittenden

unread,

May 29, 1997, 3:00:00 AM5/29/97

to

> > Hopefully Deep Blue won't carpet bomb the tournament site into the
> > stone age. ;)
>
> You never know what they have those machines hooked to these days :)

Would you like to play global thermonuclear war? ;-).

\ /~
~ / =================
~ O ~ Barton Chittenden -- to reply, remove 'bounce.'
\/|\ ti...@iglou.com from return address.
/ \\ =================
/ \~

Jim Balter

unread,

May 29, 1997, 3:00:00 AM5/29/97

to

Patrick Juola wrote:

> Chess is a perfect-information game, meaning that everything that
> you know is also known to your opponent, and vice versa. Bridge
> is not -- at least two hands are unknown to you, and your hand is
> not known to anyone else.

Lack of perfect information is not a particular problem for game-playing
programs. Even chess-playing programs which plan ahead are planning
against imperfect information -- they don't know where the opponent will
move, so they account all possibilities. Imperfect information
translates into breadth in the game tree.

> Similarly, your bidding strategy may not be known to your opponent
> (and vice versa).
>

> This means, among other things, that it's possible to bluff.

Bluffing is an integral part of game theory -- a game strategy must
not be predictable if no pure strategy dominates.

--
<J Q B>

Antoine Leca

unread,

May 29, 1997, 3:00:00 AM5/29/97

to

Gordon D. Pusch wrote:
>
> Mario Klebsch DG1AM writes:[snip]

> > In fact, I naver sa a computer driving a car.
>

> Sorry, but you're a bit out of touch --- the US military has already
> developed a truck that drives itself on standard marked roadways using
> visual cues and pattern-recognition techniques (albeit rather slowly);
> the US auto-companies are likewise researching this area...

This is off-topic, but can you withdraw "US" in the above sentence.

Or do you imply that US companies are researching while other countries
companies are demonstrating? ;-)

Thanks in advance.

Antoine

Antoine Leca

unread,

May 29, 1997, 3:00:00 AM5/29/97

to

Casper H.S. Dik - Network Security Engineer wrote:

>
> pat...@gryphon.psych.ox.ac.uk (Patrick Juola) writes:
>
> >Chess is a perfect-information game, meaning that everything that
> >you know is also known to your opponent, and vice versa. Bridge
> >is not -- at least two hands are unknown to you, and your hand is
> >not known to anyone else.
>

> >Similarly, your bidding strategy may not be known to your opponent
> >(and vice versa).
>

> That's not correct; the bidding strategy used must be made known to
> your opponents.

The general principes, yes.
The details, however, aren't, if you play with your regular partner.

And good bridge playing generaly involves a _pair_, not just one
player alone.

> You can't tell your partner anything about your cards
> that your opponent can't understand.

Oh yes you can.
And even if your adversaries (opponents if it isn't a English word)
can ask you for the meaning of a bid of your partner, every word you
speak at a bridge's table convey a lot of information beyond its
direct meaning. Example: if I'm asked about the meaning of a bid,
I will explain what I think my partner said; so my partner will know
what I understood, and then can change its subsequent bids.

Bridge is essencially a 4- then 3-players game, as opposed to chess
which is only a 2-players game. And when something involve
communication with more than 2 persons, you've indirect effects...
That's why computers aren't very good at bridge: this is the same
as with planets: the "3-bodies' problem" is much more difficult than
the "2-bodies'" one.

> >This means, among other things, that it's possible to bluff.
>

> Not really.

Oh yes. I used to named it psychic (I don't know if this is used by
English spoken persons).

Psychics are bids that are false (based on your actual game), but
only you know that. They are used for bothering your opponents.

In tournament, it's forbidden for your partner to magically understand
that it's a psychic, it must be able to demonstrate the forgery based
on all bids and its proper cards. But you still can do psychics.

Antoine

Ralph Silverman

unread,

May 29, 1997, 3:00:00 AM5/29/97

to

Alicia Carla Longstreet (ca...@ici.net) wrote:
: When IBM comes up with a computer that can beat me at Bridge, I will be

: impressed. I am NOT a great Bridge player, merely a decent one.

: Actually, if the people at IBM can produce a computer that is even

: barely competent at Bridge, I will be impressed.

: As for 'Big Blue' beating Kasperov at Chess, this is nothing more that
: using brute force to solve a problem, it is most UNimpressive.

: Actually, I AM imressed that Kasperov was able to beat the machine,
: considering that he could not begin to approach the computational speed
: of the computer.

: To me, the 'Big Blue' vs Kasperov match is comparable to racing a human
: on foot, to a remote computer controlled 100 horse power race car, I
: will be impressed if the human manages to beat the moterized computer
: controlled car even once.

: --
: ********************************************
: * Alicia Carla Longstreet ca...@ici.net
: ********************************************
: Be Good,
: If you can't be good, be careful,
: If you can't be careful, be lucky,
: and if you're not lucky...
: You damn well better be good!

with due respect, alicia,
a 486 probably can
beat you now !

From: Jim Loy <jim...@mcn.net>
Subject: Review of Bridge Programs for PC Compatibles - June 16, 1996
Date: Sun, 16 Jun 1996 22:25:01 -1000
Organization: Montana Communications Network

Review of Bridge Programs for PC Compatibles
by Jim Loy (jim...@mcn.net)
June 16, 1996 edition of this review
Copyright 1996, Jim Loy - Copy & distribute this all you want
This review is also at http://www.mcn.net/~jimloy/review.html
-----------------------------------------------------------------
CHANGES SINCE LAST REVIEW:

I'm testing Bridge Parlor
I'm testing Eddie Kantar's Bridge Companion
I'm testing MVP Bridge Deluxe for Windows 2.1
I'm testing Let's Play Bridge 96.7
I'm testing Bridge Assistant 3.1
On the skill list, I've moved Bicycle Bridge up a ways
I've added Finesse Bridge to my list of favorites

When counting conventions, I've ignored Blackwood and Stayman,
as these are required and not optional in many programs, and are
sometimes not even mentioned in the documentation. And I've
ignored conventions that are not optional, for example Bridge
Buff 3.0 makes you use 1NT Forcing.

-----------------------------------------------------------------

NOTE: There's STILL "confusion" over the top spot. Micro Bridge
7 seems to be the strongest program, when it plays at its higher
skill levels. But, then it is too slow for my 33 Mhz machine.
I'll do more testing. In order to break the tie, I'm going to
have to have them play some long matches, maybe some other day.

NOTE: Concerning skill, Bridge Baron has a standing $2500
challenge. So far, no other program has accepted the challenge.

NOTE: A number of the shareware programs are available from the
rgb (rec.games.bridge) FAQ archives:

http://rgb.anu.edu.au/Bridge/FAQ/Programmes/
ftp://rgb.anu.edu.au/pub/Bridge/FAQ/Programmes/

-----------------------------------------------------------------

MAC VERSIONS: The following programs come in Mac versions. I do
not know if this list is complete:

Bridge Baron (no 2/1 Forcing, may be old version)
Bridge Deluxe with Omar Sharif (may be old version)
Bridge Master? (Capstone) (I'm not sure of this one)
Bridge 8.0 (peeks at the cards)
Perfect Partner

-----------------------------------------------------------------
The better programs bid pretty well, and play OK (with lapses)
Here, the ranking (awarding of *'s) and criticism are based on my
own estimates of 1-bidding and playing strengths, 2-user
friendliness, 3-number of conventions, 4-extra features
(takeback, custom hands, can I play N or E or W?, skill
levels, closed room...), 5-bugs. I'm not very impressed by
looks, sound, copy protection, and my partner's personality.
I have not tested the Mac versions of any program.
I've spent over $1500 on these programs, so far.
-----------------------------------------------------------------

MY FAVORITES:
Bridge Mate 2 (Precision, many conventions, good bidding)
Bridge Baron 6 (Windows & OK play & good bidding)
Micro Bridge (Good play)
Oxford Bridge 4 (Acol & modify conventions)
Meadowlark Bridge (Excellent bidding)
Finesse Bridge (No-nonsense simplicity & ease of use)

Choose the features you like, as there is no perfect program.

=================================================================
SKILL

Here's how I roughly rank the skill of the top 24 programs today:

rank bidding (1-10) play (1-10) total
*****
1. Bridge Mate 2 9 6(peeks) 15
1. Bridge Baron 6 9 6 15
1. Micro Bridge 7 8 7 15
1. Meadowlark 9+ 6- 15
5. Bridge Buff 3.0 9- 5(peeks) 14-
****
6. Bridge Parlor 8 4? 12?
7. Oxford Bridge 7 4 11
7. Br. Master Class (Omar) 7 4 11
7. Br. Deluxe II (Omar) 7 4 11
7. Finesse Bridge 8 3 11
11.Blue Chip Bridge 7 3 10
12.Grand Slam Bridge II 6 3(peeks) 9
12.Br. Master (Capstone) 6 3 9
***
14.WinBridge 5 3 8
14.Bridge Brain 5 3 8
14.Bridge Assistant 3.1 5 3 8
14.Let's Play Bridge 5 3 8
14.Bicycle Bridge 5 3 8
**
19.Perfect Partner 4 3 7
19.Contract Bridge 2.5 4(peeks) 3(peeks) 7
21.Bridge Olympiad 3 4-(peeks) 7-
22.MVP Bridge Deluxe 2.1 3 3? 6?
22.Eddie Kantar's Br.... 3 3? 6?
24.Bridge 8.0 3(peeks) 3(peeks) 6

Other Programs (with less skill than the above programs):
* Randy's Bridge
* MVP Bridge (the Deluxe version is listed above)
* Hoyle Bridge
* Hoyle Classic Card Games
* Classic 5 Games
Bridge Pal
Bridge for Windows
Bridge Partner Alice
TRBridge
Expert Cardgames Classics

I think that the top six programs bid about as well as the
average club player and play worse than the average club player.

=================================================================
FEATURES
! = Great feature, + = Good feature, - = Bad feature

--------------------------5 stars--------------------------------

***** Bridge Mate 2
Bridge Mate, 8845 42nd Ave. S.W., Seattle, WA 98136
email: bo...@aa.net
DOS - $59.95 + $5 s&h
! 237+ conventions, much more than any other
! Systems = 5-card majors, 2/1 forcing, Goren, Kaplan/Sheinwold,
Precision (bids Precision very well), Forum-D
- Peeks at the cards, after opening lead
! Deal types = Really tremendous variety
+ Skill levels = 5
+ Closed Room
+ Defensive signals
! Explains the bids in numerous ways, including alerting
- Nice-looking text, rather than graphics

***** Bridge Baron 6 (was called Micro Bridge Companion)
Great Game Products Inc., 8804 Chalon Drive, Bethesda, MD 20817
(phone: 1-800-GAMES-4-U)
email: great...@delphi.com
DOS/Windows/Mac/or CD (DOS & Windows together) - $59.95 + $5 s&h
(I don't know the CD price)
8 conventions
Systems = 5-card majors, 2/1 forcing
+ Skill levels = 4
Deal types = several
+ Duplicate matches
! Explains the bids in numerous ways, including interactive
flow charts
! Tutorial - Sheinwold Bridge Challenges
! You can compete in worldwide tournaments
+ Defensive signals

***** Micro Bridge 7
Tomio Uchida, 2-1-7 Mizonuma, Asaka-shi, Saitama-ken 351, JAPAN
email: uch...@rkmath.rikkyo.ac.jp
DOS - $85 including s&h
+ 52 conventions. Define your own conventions with difficulty
+ Systems = 4-card majors, 5-card majors. You can define 2/1
forcing, Acol.
! Deal types = Tremendous variety
! Skill levels = 9 (some very very slow on most machines)
+ Pairs and team matches
+ Estimates unseen hands with Monte Carlo sampling
+ Defensive signals

***** Oxford Bridge 4.5
Thinking Games, Cedar Lodge, The Crescent, Pattishall,
Northants NN12 8NA ENGLAND
phone/fax: +44-1327-830490
email: o...@thinkgam.demon.co.uk
Windows UKP59.95 (pounds) or DOS UKP49.95
DOS version has no graphics, to save memory
+ Systems: Standard American, Acol, European 5-card majors
25 conventions
! You can define your own bidding conventions and systems
! Explains bids in numerous ways
+ Count signals (only?)
+ Closed room
! Very very detailed manual (substitutes for help messages)

***** Meadowlark Bridge
Meadowlark Software, 2718 N. Broadway #283, Fargo, ND 58102
email: meado...@rrnet.com
http://rrnet.com/meadowlark
Windows - $59.95 + $5 s&h
+ 49 conventions
* Systems = 5-card majors, 2/1 forcing, Kaplan/Sheinwold, Yellow
card, Simple
+ Skill levels = 3 bidding, 3 play (various bird personalities)
! Tournaments
+ Deal types = great variety
- Slow loading. I erroneously said that it bid and played
slowly. The bidding is not overly slow, and the play is
actually fast.
+ No longer copy protected
- You must play South
+ Defensive signals
! A demo version can be downloaded from Meadowlark's WWW page
Allows you 3 24-hand tournaments before repeating the hands

***** Bridge Master Class with Omar Sharif
Oxford Softworks, Stonefield House, 198 The Hill, Burford,
Oxford, OX18 4HX, ENGLAND
http://www.demon.co.uk/oxford-soft/bridge.html
40 UKP + s&h - MSDOS CD
about 25 conventions
Systems: Acol & Standard American - 5-Card Majors
+ Defensive signals (rare?)
! Allows several players to play on a Local Area Network
Very fancy looking
+ Deal types = many combinations
- Loads slowly
- Both partnerships must use the same conventions
! CD has good lectures by Omar Sharif

***** Bridge Deluxe II with Omar Sharif
Interplay Productions, 17922 Fitch Ave., Irvine, CA 92714
email: orde...@interplay.com
http://www.interplay.com
$29.95 - MSDOS (Mac and Windows versions planned) - CD $29.95
Mac version is Bridge Deluxe with Omar Sharif, from MacPlay
(same address as Interplay)
25 conventions
Systems: Standard American - 5-Card Majors
+ Defensive signals (rare?)
! Allows several players to play on a Local Area Network
Very fancy looking
+ Deal types = many combinations
- Loads slowly
- Both partnerships must use the same conventions
! CD has good lectures by Omar Sharif

--------------------------4 stars--------------------------------

**** Bridge Buff 3.0
BridgeWare, PO Box 65077, 358 Danforth Ave., Toronto, Ont.
M4K 3Z2 CANADA
phone: 1-800-WEAK-1NT
$99.95 + $5 s&h - Windows or MSDOS
$49.95 upgrade if you registered an earlier version
+ 25 conventions
Systems = 5-card majors, 2/1 forcing
- Peeks at the hands during play
- You must play South
! Kibitzer explains bids, Post Mortem estimates how various
contracts would do
- Both sides must use the same conventions

**** Blue Chip Bridge
Blue Chip Bridge Ltd., P.O. Box 167, Waltham Cross, EN7 5GB
ENGLAND
I don't know the price - Windows
10 conventions
System = Acol
! You can define your own bidding conventions & systems
! Deal types = tremendous variety
- No takeback bid or play

**** Perfect Partner
Positronic Software Inc., 114 Thornhill Drive, Dartmouth,
Nova Scotia, Canada, B3B 1S3
phone: 1-800-565-4005
$59.95 - Windows or Mac
! Will, in time, learn your bidding system, if close to
Standard American
20 conventions
- Bids and plays and loads slowly
Systems: 5-Card majors, 2/1 forcing
- Bugs - Bids very poorly (half the time), won't print deal
- You must play South
+ Closed room
+ Skill levels = several
Defensive signals when you choose some bidding systems?

**** Finesse Bridge v1.5
Wild Card Software, P.O. Box 208, Hockessin, DE 19707-9954
email: wildc...@aol.com
! $19.95 + $5 s&h - Windows ($39.95 with Pinochle & Gin Rummy)
- 0 conventions
Systems: Standard American - 4 or 5-Card Majors

**** Grand Slam Bridge II
Electronic Arts, P.O. Box 7578, San Mateo, CA 94403-7578
http://www.ea.com
$49.95 - MSDOS or Windows (also on CD with other games)
8 conventions
Systems: 5-Card Majors & 4-card majors
- Peeks at the hands during play
! Deal types = really tremendous variety
- Bug = will bid Jacoby transfers over 2NT, but won't
recognize it when partner bids it

**** MVP Bridge Deluxe for Windows v2.1
MVP Software, 1035 Dallas SE, Grand Rapids, MI 49507-1407
email: 74777...@compuserve.com
$39.95 - Windows ($20 upgrade for registered users)
MVP Bridge (listed below) is shareware.
! Network/modem play
10 conventions
Systems: 4 (Goren) & 5-card majors (Standard American)
- Bugs in the bidding

--------------------------3 stars--------------------------------

*** Eddie Kantar's Bridge Companion
Lifeware, 63 Orange St., St. Augustine, FL 32084-3584
http://www.lifeware.com
About $12 - Windows CD
Systems - Goren & 5-card majors
9 conventions
Systems: 4 & 5-card majors
! Several cardplay tutorials
- Bugs in the bidding

*** Bridge Master (There's another product with this name)
There is a bridge tutoring program of the same name
Capstone, Intracorp, 7200 N.W. 19th Street, Ste. 500, Miami,
FL 33126
http://www.intracorp.com
$40? - MSDOS (It crashes under Windows on my computer)
There seems to be a Mac version
4 conventions
Systems = Standard American - 5-Card Majors
! Allows several players to play on local network or by modem
There is reportedly a new version. I haven't seen it.

*** Bridge 8.0
Artworx, Penfield, NY 14526
email art...@frontiernet.net
http://www.frontiernet.net/~artworx
$54.95 + $4 s&h - MSDOS or Mac
+ User friendly
2 conventions
System: Standard American (sort of) - 5-Card Majors
- Peeks at your cards during play and bidding

*** Contract Bridge 2.5
Robert Lindsay Wells, 1405 Lynn Ave., Clearwater, FL 34615
email Well...@aol.com
http://softsite.com/authors/auth100/auth157.htm
MSDOS - shareware - register $19
0 conventions
System: 4-card majors
Somewhat friendly - This is a new version, and I am testing it
- Crude text, no graphics
- Peeks during bidding and play

*** Bridge Olympiad
QQP (Quantum Quality Products), 495 Highway 202, Flemington,
NJ 08822
$59.95 - MSDOS
Systems: Standard American, Natural, or Precision (crude)
+ Nice looking
0 conventions
- not friendly (no play review after last trick, no
hand-entered deals, can't save hand, can't print hand)
Many psychic bids from some partners or opponents (the program
calls this "cheating")
- Peeks at the hands during play
- Copy protection - You must look up a word in the manual
Some people enjoy this program a lot

--------------------------2 stars--------------------------------

** WinBridge
Pik A Program, 13 Saint Marks Place, New York, NY 10003
Europe: Gerald Wilson, WinBridge orders, 36, Ferndale Rd.,
Church Crookham, FLEET, Hampshire, GU13 0LN, UK
email: 10004...@compuserve.com
Windows - Shareware - register $29.95 + $4 s&h
- Somewhat unfriendly (can't takeback bids, no play review,
can't print deal)
Registered version has more features
5 conventions
Systems: Acol system

** Let's Play Bridge - 96.7
Terry Inc., P.O. Box 21289, Carson City, NV 89721
email: l...@letsplaybridge.com
http://letsplaybridge.com
Windows - Shareware - $39.95 to register to unlock features
Somewhat unfriendly (no hints, no takeback, no play review,
can't save, can't print deal)
0 conventions
Free demo, only deals two hands
Deal types = 4 types
+ Suit preference signals
! Network/modem play

** Bridge Brain
George Bodnar, 1134 Harvard Circle, Pittsburgh, PA 15212
email: BOD...@duq3.cc.duq.edu
Windows - Shareware - register $6 for password to unlock more
features
Available free from rgb archives (URL at top of this review)
2 conventions
Systems: 5 or 4-card majors
- Unfriendly (can't takeback bids, can't see deal at the end,
no play review, can't print deal)
- A bug prevents you from bidding a retrieved or hand-entered
deal

** Bridge Parlor
Dynacomp, Inc., 178 Phillips Rd., Webster, NY 14580
phone: 1-800-828-6772
DOS - $69.95
286 computer = 8 mhz (probably won't run on a faster computer!)
Systems: 4 or 5-card majors
1 conventions
- Not very friendly (lots of features, but awkward)
- Text, no graphics
+ Defensive signals (several options)
NOTE: Some people recommend it highly

** Bridge Assistant 3.1
Public (software) Library, P.O. Box 35705, Houston, TX
77235-5705
UK: Softtech, 24 Rosenthal Rd., London SE6 2BY ENGLAND
email: 10011...@compuserve.com
Windows - Shareware - register $37.50 (post. pd) (UKP25)
0 conventions
- somewhat unfriendly (no rebid, can't see deal at end, no play
review, can't print deal)
NOTE: I'm testing this new version

--------------------------1 star---------------------------------

* Bicycle Bridge
Swfte, PO Box 219, Rockland, DE 19732
$30? - for MSDOS and Windows (in same package)
Is also packaged with other card games as well
0 conventions
Systems: Standard American (crude)
Somewhat unfriendly (no hints, no rebid or replay, can't
see deal at end, can't print deal)

* Randy's Bridge 1.2 (also called R-bridge)
Randall Shepardson, 662 Bates College, Lewiston, ME 04240
Windows - shareware - register $20
0 conventions
Systems: Standard American
Somewhat unfriendly (no hints during play, no rebid, no
replay, no takeback play, no play review, can't print deal

* MVP Bridge (see MVP Bridge Deluxe above)
Register MVP Bridge or order MVP Bridge Deluxe, listed above
MVP Software, 1035 Dallas SE, Grand Rapids, MI 49507-1407
MSDOS - register $20
1 convention
Systems: Goren (sometimes opens a 4-card major)
Unfriendly (no hints, no rebid, no takeback, no
replay, no hand-entered deals, can't save, can't print)

* Hoyle Bridge
Sierra On-Line, Coarsegold, CA 93614
http://www.sierra.com
About $12 - Windows CD
+ Nice looking
0 conventions
System: Standard American 5-card Majors?
- Unfriendly (No hints, no takeback, can't see deal at end,
no play review, can't hand-enter, can't print deal)

* Hoyle Classic Card Games (including Bridge)
Sierra On-Line, Coarsegold, CA 93614
http://www.sierra.com
$40.95 - MSDOS and Windows (in same package)
+ 8 card games (Crazy 8s, Old Maid, Hearts, Gin Rummy,
Cribbage, Klondike, Bridge, Euchre)
+ Nice looking, low resolution graphics
0 conventions
System: Standard American 5-card Majors
- Slow (even on last trick)
- Unfriendly (No hints, no takeback, can't see deal at end,
no play review, can't hand-enter, can't print deal)

* Classic 5 (Backgammon, Chess, Bridge, Go, Checkers)
Interplay Productions, 17922 Fitch Avenue, Irvine, CA 92714
$12? - MSDOS (May be on CD now)
1 convention
Systems: Acol
- Somewhat unfriendly (no takeback, no hints during play,
no play review, no hand-entered deal, can't print deal)

* Bridge Partner Alice
Dynacomp, Inc., 178 Phillips Rd., Webster, NY 14580
phone: 1-800-828-6772
DOS - $49.95
- Not friendly - very crude
0 conventions
! Learns your bidding system from scratch
- Crude text, no graphics

-------------------------no stars--------------------------------

Bridge Pal
WR Software, P.O. Box 4819, Walnut Creek, CA 94596
DOS - Shareware - Reg. $9.95
1 convention
System: 4-card majors
- Not friendly (no hints during play, no takeback, can't see
deal at end, no play review, can't print deal)
- Large text, no graphics

Bridge for Windows (also called Wnbridge under its icon)
Software Labs, 8870 148th Ave. NE, Redmond, WA 98052
Registered version: Full Circle Computing, 15 Greenridge Ave.
Suite 19, White Plains, NY 10605-1248
Windows - About $2.50 - shareware - register $49 + $4 shipping
0 conventions
System: Standard American
- You always get the good hand
- Opponents never bid
- Not friendly (no hints during bidding, no takeback bids,
takeback one card, can't see deal at end, can't print deal)

TRBridge - compiled version of Turbo Gameworks Bridge
Source code version discontinued by Borland, years ago
MSDOS - compiled version available free on the Internet
Several bugs, not friendly, few features

Expert Card Games Classics CD
Expert Software, 800 Douglas Rd, Executive Tower #750, Coral
Gables, FL 33134
phone: 1-800-759-2562
Windows CD - about $12 (bridge, poker, cribbage, solitaire)
0 conventions
- System: unknown
- Not friendly (no hints during play, no rebid, can't see deal
at end, no play review, can't print deal)
- Doesn't follow the rules

JTS Bridge Bidder - version 3.0
JTS Micro Consulting Ltd, RR#4, 10931 Lytton Rd., Ladysmith, BC,
CANADA V0R 2E0
Shareware - register $25 ($35 Canadian) & get password - Windows
- Does not play the cards (bids well)
- very slow dealer (a full minute to deal on my 33 mhz machine)
If it could deal and play cards, I'd give it 4 or 5 stars
0 conventions
System: 4 or 5-card majors - pt-count ranges for bids can be
adjusted.
Unregistered version won't load after 30-day pre-reg period

=================================================================

COMMENTS:

I've been searching for defensive signals:
Bridge Baron signals. Trump echo and attitude.
Bridge Mate signals. Attitude. Suit preference?
Meadowlark signals, some times. It gives count signals.
Bridge Buff signals, some times?
Micro Bridge signals. Attitude.
Oxford Bridge gives count signals only (according to manual)?
Perfect Partner may signal, when you choose some bidding
systems?
Bridge Deluxe 2 with Omar Sharif signals attitude?
Let's Play Bridge signals suit preference.
Bridge Parlor signals, with several options.

They all may ignore your signals.
None of the other programs seems to signal, ever.

Jim Loy (jim...@mcn.net)

--

Ralph Silverman
z007...@bcfreenet.seflin.lib.fl.us

Andrew Gierth

unread,

May 29, 1997, 3:00:00 AM5/29/97

to

>>>>> "Ralph" == Ralph Silverman <z007...@bcfreenet.seflin.org> writes:

[nothing worth reading, followed by a *huge* piece of HTML]

Sigh. Time to trot out an old favourite:

STANDARDIZED BONEHEAD REPLY FORM, Version 1.24
(c) copyright 1999 by
various people.

(check all boxes that apply)

Dear:

[x] Clueless Newbie [ ] Lamer [ ] Flamer
[ ] Loser [ ] Spammer [x] Troller
[ ] "Me too" er [ ] Pervert [ ] Geek
[ ] Freak [ ] Nerd [ ] Elvis
[ ] Racist [ ] Fed [ ] WaReZ D00D
[ ] Fundamentalist [ ] Satanist [ ] Revisionist
[ ] IBM/Mac Bigot [ ] Apologist [ ] Scientologist
[ ] "Expert" [ ] Hypocrite [x] Moron
[x] Evolutionary Dead End
[ ] Stereotypical AOLer
[ ] Unbearably self-righteous person

I took exception to your recent:

[ ] Email [ ] "News" article
[x] Post to _comp.unix.programmer_.
(newsgroup)

It was (check all that apply):

[x] Lame [ ] Stupid [ ] Abusive
[ ] Clueless [ ] Idiotic [ ] Brain-damaged
[ ] Imbecilic [ ] Arrogant [ ] Malevolent
[ ] Contemptible [ ] Libelous [ ] Ignorant
[ ] Nauseating [ ] Stupid [ ] Fundamentalist
[x] Boring [ ] Dim [ ] Cowardly
[ ] Deceitful [ ] Demented [ ] Self-righteous
[ ] Crazy [ ] Weird [ ] Hypocritical
[ ] Loathsome [ ] Satanic [ ] Despicable
[ ] Belligerent [ ] Mind-numbing [ ] Maladroit
[ ] Assinine [x] Pointless [x] Annoying
[x] Much longer than any worthwhile thought of which you may be
capable.

Your attention is drawn to the fact that:

[ ] You Spammed
[ ] You defended someone who Spammed.
[ ] You doubtlessly violated the Terms of Service of your ISP
[ ] You posted what should have been emailed
[x] You obviously don't know how to read your newsgroups line
[ ] You obviously know nothing about the topic at hand
[ ] You are trying to make money on a non-commercial newsgroup
[ ] You are placing your profits above fundamental ethics
[ ] You self-righteously impose your religious beliefs on others
[ ] You self-righteously impose your racial beliefs on others
[ ] You made a blanket statement without backing it up
[ ] You made unwarranted assumptions
[ ] You posted a "news article" without bothering to do basic research
[ ] You posted a binary in a non-binaries group
[ ] You don't know which group to post in
[ ] You made a post and didn't say anything
[x] You posted something totally uninteresting
[x] You crossposted to *way* too many newsgroups
[ ] I don't like your tone of voice
[x] You are the reason some groups are moderated
[x] You posted in HTML
[ ] You used the phrase "Information Superhighway"
[ ] You cannot recognize sarcasm
[ ] You made an error and blamed someone else
[ ] You flamed someone for a spelling error.
[ ] And you misspelled something in the process
[ ] What you posted has been done before.
[ ] and it was done better the last time.
[ ] You quoted an *entire* post in your reply
[ ] You quoted nothing, so nobody knows what you refer to
[ ] You messed up the attributions on your quotes.
[ ] You started a long, stupid thread
[ ] You continued spreading a long stupid thread
[x] You crossposted an existing thread into tons of irrelevent groups
[x] Your post is absurdly off topic for where you posted it
[ ] You posted a followup to crossposted robot-generated spam
[ ] You posted a URL that had nothing to do with the subject header
[ ] You posted a "test" in a discussion group rather than in alt.test
[ ] You posted a "YOU ALL SUCK" message
[ ] You posted low-IQ flamebait
[ ] You posted a blatantly obvious troll
[ ] You followed up to a blatantly obvious troll
[ ] You said "me too" to something
[ ] You asked questions straight out of the FAQ
[ ] Deliberately
[ ] You posted a "FAQ" containing deliberately misleading information
[ ] You make no sense
[ ] Your sig/alias is dreadful
[ ] Your sig is longer than the body of your message
[ ] It contains stupid song lyrics
[ ] It contains a stupid "witty" quote
[ ] It contains worthless ASCII graphics
[ ] Amiga tick [ ] Map
[ ] Map of Australia
[ ] including pointer to Perth
[ ] You have been warlorded
[ ] more than once
[ ] You posted a phone-sex ad
[ ] You posted a stupid pyramid money making scheme
[ ] You claimed a pyramid-scheme/chain letter for money was legal
[ ] You argued that spamming was legal, even though nobody claimed it
wasn't
[ ] Your margin settings (or lack of) make your post unreadable.
[x] You used unusual line formatting making your post confusing.
[ ] You posted in ELitE CaPitALs to look k3wL
[ ] You failed
[ ] Miserably
[ ] You posted a message in ALL CAPS, and you don't even own a TRS-80
[ ] Your post was FULL of RANDOM CAPS for NO APPARENT REASON
[x] your post contained no capital letters at all
[ ] You referred to the Usenet Cabal (There is no Cabal)
[ ] You attempted a deliberate invocation of Godwin's Law
[ ] You referred to Usenet as "the Internet", or vice versa.
[x] You have greatly misunderstood the purpose of this newsgroup.
[ ] You have greatly misunderstood the purpose of Usenet.
[ ] You have greatly misunderstood the purpose of email.
[ ] You have greatly misunderstood the purpose of the Internet.
[ ] You have greatly misunderstood the purpose of communication
[x] You are a loser.
[x] This has been pointed out to you before.
[x] more than once
[ ] You believe O.J. Simpson is innocent
[x] You have a direct genetic link to:
[ ] Sanford Wallace [ ] Phil Lawlor
[ ] Walter Rines [ ] Zach Everett
[ ] John Grubor [x] Steven Boursy
[ ] ___other___
[ ] You are one of the scuzzwipes from:
[ ] AGIS
[ ] the Woodside Literary Agency
[ ] the Copy Cat Shop
[ ] You didn't do anything specific, but appear to be so generally
worthless that you are being flamed on general principles.

It is recommended that you:

[x] Get a clue
[ ] Get a life
[x] Go away
[ ] Grow up
[ ] Act something close to your age
[ ] Never post again
[ ] Smoke less of what you are smoking
[ ] Consider how low the tolerance on the Internet is for Spam
[ ] Become Sanford Wallace's love slave
[x] Read every newsgroup you posted to for a week
[ ] stop reading Usenet news and get a life
[ ] stop sending Email and get a life
[ ] Bust up your modem with a hammer and eat it
[ ] Have your medication adjusted
[ ] Go to the top of a building and attempt to defy gravity
[ ] Jump into a bathtub while holding your monitor
[ ] find a volcano and throw yourself in
[ ] get a gun and shoot yourself
[ ] poke your eye out with a stick
[x] Actually post something relevant
[x] Read the FAQ
[ ] stick to FidoNet and come back when you've grown up
[x] Apologize to everybody in this newsgroup
[ ] Attempt to win this year's Darwin Award
[ ] Do not pass along your genes
[ ] consume excrement
[ ] consume excrement and thus expire
[ ] Perform solitary sexual congress
[ ] Post your tests to alt.test/misc.test
[ ] Put your home phone number in your ads from now on
[ ] Don't bother getting a new account when this one is pulled
[x] Refrain from posting until you have a vague idea what you're doing

In Closing, I'd Like to Say:

[ ] You need to seek psychiatric help
[x] Take your gibberish somewhere else
[ ] *plonk*
[x] Learn how to post or get off the usenet
[ ] Learn the purpose of email
[ ] Take a basic economics course and learn the costs of UCE
[ ] Go back to school and actually learn something
[ ] Your mother wears army boots
[ ] You are possibly the most annoying person currently alive
[ ] _____other______
[ ] Most of the above
[ ] All of the above
[ ] Some of the above, not including All of the above
[ ] You are so clueless that I didn't even bother filling in this form.

--
Andrew.

comp.unix.programmer FAQ: see <URL: http://www.erlenstar.demon.co.uk/unix/>

Alicia Carla Longstreet

unread,

May 29, 1997, 3:00:00 AM5/29/97

to

Paul Hsieh wrote:

>
> On 25 May 1997, H. M. Hubey (hu...@pegasus.montclair.edu) said:
> > Alicia Carla Longstreet <ca...@ici.net> writes:
> >

> > >When IBM comes up with a computer that can beat me at Bridge, I will be
> > >impressed. I am NOT a great Bridge player, merely a decent one.
> >

> > Most people probably won't be.
> >

> > That's why probably nobody will spend millions making
> > bridge-playing computers. The reason for picking chess is
> > obvious; it has been shown everywhere as the prime example
> > of an intelligence-requiring and an intellectual game.
>
> Actually, I've seen some very strong bridge programs (usually combined
> with a tutorial) out there. But as you suggest, nobody really cares, so
> these programs fell into relative oblivion.

Tell me about them. I have never seen a Bridge program that was much
above the level of a Novice. In Chess you can computationally test
every possible sequence. Further, using simple AI you can eliminate
some of the sequences early (the earlier the better, obviously). Bridge
is a far more complex game than chess, be several orders of magnitude.

> > >As for 'Big Blue' beating Kasperov at Chess, this is nothing more that
> > >using brute force to solve a problem, it is most UNimpressive.
>

> Your knowledge about the internal working of Deep Blue is what is most
> unimpressive. The Deep Blue team has in the past published several
> papers which have been considered great contributions to the field of AI.
> As far as I know they have continued to use some of their more successful
> techniques in the Deep Blue program that helped it be more than just a
> brute force search engine.

Obviously, did it occur to you that I was simply unimpressed with the
state of AI today. I have watched them go down so many dead end roads.
If they had a REALLY good AI Chess playing program: 1) They would have
beaten Kasperov with a PC, 2) Kasperov wouldn't have even drawn any of
the games. Again, solving Chess is 'childs play' compared to a real
game like Bridge or Go.

> > It also happens to be the case that Kasparov plays chess by
> > using brute force. But because it is done via the human
> > brain, we are told that it is intelligence.

Not very likely. I doubt that Kasparov examines even a small percentage
of the possible sequences.

> :o) An interesting way of looking at it. Many people believe that the
> pursuit of artificial intelligence must be to model the way the human
> mind works. But it also encompasses using the computer to think about
> problems in ways unfeasible to humans, such as was done with Deep Blue.

> > >Actually, I AM imressed that Kasperov was able to beat the machine,
> > >considering that he could not begin to approach the computational speed
> > >of the computer.

> > I am impressed that the machine beat kasparov since it is
> > obvious that it is not working in parallel like Kasparov's
> > brain :-)

Human brains do not work in parallel, at least in terms of the way
computers work in parallel. Well, at least, there is no evidence to
this effect.

> Actually, Deep Blue *is* a highly parallelized machine. Though, perhaps
> not in the sense you mean.

> > >To me, the 'Big Blue' vs Kasperov match is comparable to racing a human
> > >on foot, to a remote computer controlled 100 horse power race car,

> Your analogy belittles what chess is, and how hard it was for the Deep
> Blue team to create such a device with such capability.

You got it!

> > > I
> > >will be impressed if the human manages to beat the moterized computer
> > >controlled car even once.

> > Yep.
> > From now on, we will be impressed if humans can match machines.
> > And this is the beginning. :-)

> :o) You have that little faith in man kind? Computers ability to do
> some things better than humans will consequentially make humans better at
> doing other things. For example the computer's ability to calculate
> spread sheets faster than a human can do by hand, intrinsically made
> humans (at least some) better at designing spread sheets to tell them
> more, and give them more information.

> Even in Chess, as these programs have been released as commercial
> products, even the top grandmasters have been using them to check their
> analysis, keep a database of openings, special opening repetoires as well
> as being used as a way of studying the games of their opponents. It has
> also been used to further the state of opening theory in some cases. But
> they computer doesn't win the cash prize, the human who used it does.

Jim Balter

unread,

May 29, 1997, 3:00:00 AM5/29/97

to

Paul Hsieh wrote:
> The Deep Blue team has in the past published several
> papers which have been considered great contributions to the field of AI.

Citations, please? You may be confusing Deep Blue with Deep Thought.

--
<J Q B>

Patrick Juola

unread,

May 29, 1997, 3:00:00 AM5/29/97

to

In article <338D2B...@netcom.com> Jim Balter <j...@netcom.com> writes:

>Patrick Juola wrote:
>
>> Chess is a perfect-information game, meaning that everything that
>> you know is also known to your opponent, and vice versa. Bridge
>> is not -- at least two hands are unknown to you, and your hand is
>> not known to anyone else.
>

>Lack of perfect information is not a particular problem for game-playing
>programs. Even chess-playing programs which plan ahead are planning
>against imperfect information -- they don't know where the opponent will
>move, so they account all possibilities. Imperfect information
>translates into breadth in the game tree.

There's a big difference between "not knowing where the opponent will
move" and "not knowing what the opponent knows." A better description
is, in my opinion, that "imperfect information translates into
user modelling."

It's a collapse of the "X knows that Y knows that..." hierarchy; in
chess, by construction, for every fact that I know, I also know
that you know that fact, and so on into infinity. This doesn't
hold for bridge (obviously).

>> Similarly, your bidding strategy may not be known to your opponent
>> (and vice versa).
>>

>> This means, among other things, that it's possible to bluff.
>

>Bluffing is an integral part of game theory -- a game strategy must
>not be predictable if no pure strategy dominates.

Ah, but you see, a pure strategy *does* dominate chess; that's the
essence of the lookup-table algorithm.

-kitten

Shez

unread,

May 29, 1997, 3:00:00 AM5/29/97

to

jggo...@oakland.edu (Monument) writes:
|In comp.ai.alife Dale Pennington <dpen...@ingr.com> wrote:
|: Monument <jggo...@oakland.edu> wrote in article

..
|: > Point #3: Deep Blue would have adapted to Kasparov's playing style
|: > even if it hadn't been fed his books and previous games in a database.
|: > It was, to my knowledge, designed with a NNet/GA combination where
|: > they rebuilt the NN after each game if possible.
|: True, but it got a large amount of adaption to Kasparov's style before the
|: first game they played. If it had started with no Kasparov knowledge, then
|: I think Kasparov would have won this contest hands down, while the computer
|: developed the database to be a competitor later on.

I get the impression in this thread that people don't appreciate how
human grand masters operate. No human grandmaster would approach a
tournament by simply sitting down and playing. What they in fact do is
rather similar to how Deep Blue was prepared: they spend six months
going over every game their opponent had ever played, to get an insight
into their style.

During the tournament, just as Deep Blue was tweaked to take account of
the game just played, so a human such as Kasparov would spend long hours
in strategy meetings with fellow grandmasters who'd been enlisted as
advisers. When Nigel Short played Kasparov for the world championship,
he, like Deep Blue, had a room full of grand masters (eg. Jonathon
Spielman) advising him between games. Big league chess is a team thing,
with behind the scenes coaches and sparring partners, like any major
sport. It's only during the actual game that the player is alone.

In reprogramming Deep Blue between games with the help of grand masters,
IBM were doing no more and no less than Kasparov himself would have been
doing at the same time.

-Shez.
____________________________________________________
If replying by email delete Don't_Spam. from address

Peter Harrison

unread,

May 30, 1997, 3:00:00 AM5/30/97

to

On 28 May 1997 12:35:24 GMT, Caspe...@Holland.Sun.COM (Casper H.S.

Dik - Network Security Engineer) wrote:

>pat...@gryphon.psych.ox.ac.uk (Patrick Juola) writes:
>>This means, among other things, that it's possible to bluff.
>

>Not really.

I am not a bridge expert, but Patrick is correct in that body
langauge, and the reading thereof becomes a part of the game. The
ability to tell when your opponent is really bluffing.

The objective is a classic warfare situation - deceive the enemy to
win.

Computers have enough trouble with 'perfect' data, such as a chess
board. When dealing with data that is DELIBERATLY DESIGNED to deceive
the computer cannot recognise it.

For a computer to be programmed in the same manner as chess to play
bridge the 'search' explodes into a great many more possibilities - to
the point of being useless.

By the way, although humans can think of what may occur five or six
moves ahead, deep blue ACTUALLY PLAYS THEM, thus being able to
evaluate each move exactly. Do you honestly calculate every possible
eventuality of a move for the next four moves - or even two - before
moving? I think not. You have a specific 'game plan' in mind, and
when the opposition moves out of that game plan you revise it to
maintain the same goal - ie checkmate.

I don't believe Deep Blue thinks in terms of planning a specific set
of moves. Rather it simply calculates the best move with all the
available processing power - Raw Grunt processing.

Peter

RHA

unread,

May 30, 1997, 3:00:00 AM5/30/97

to

In article <338E12...@ici.net>,
Alicia Carla Longstreet <ca...@ici.net> wrote:
>Paul Hsieh wrote:
[cut]

>>
>> On 25 May 1997, H. M. Hubey (hu...@pegasus.montclair.edu) said:
>> > Alicia Carla Longstreet <ca...@ici.net> writes:
>> > >As for 'Big Blue' beating Kasperov at Chess, this is nothing more that
>> > >using brute force to solve a problem, it is most UNimpressive.
>>
>> Your knowledge about the internal working of Deep Blue is what is most
>> unimpressive. The Deep Blue team has in the past published several
>> papers which have been considered great contributions to the field of AI.
>> As far as I know they have continued to use some of their more successful
>> techniques in the Deep Blue program that helped it be more than just a
>> brute force search engine.
>
>Obviously, did it occur to you that I was simply unimpressed with the
>state of AI today. I have watched them go down so many dead end roads.
>If they had a REALLY good AI Chess playing program: 1) They would have
>beaten Kasperov with a PC, 2) Kasperov wouldn't have even drawn any of
>the games. Again, solving Chess is 'childs play' compared to a real
>game like Bridge or Go.

What I'd like to know is whether "Deep Blue" is capable of
"learning" to play better chess. Would two, or more, Deep
Blues playing against each other develop a stronger game?
Or would programmers have to tweak the code?
--
rha

Gordon D. Pusch

unread,

May 30, 1997, 3:00:00 AM5/30/97

to

[Note --- 'Followup-To:' field brutally trimmed...]

In article <MPG.df64fcfe...@news.zocalo.com>
nob...@chromatic.com (Paul Hsieh) writes:

> On Thu, 29 May 1997 09:42:43 GMT, Peter Harrison (p...@ihug.co.nz) said:
>

> > A human DOES NOT 'look ahead' at all, they develop stratagies,
> > and use previous experience to apply to the current situation -
> > even though that situation is not identical.
>

> Excuse me? I'm a human, I play chess at an acceptable level.
> *I* look ahead. All my chess playing friends all look ahead.
> The strong the player, it seems to me, the more they look ahead!
> (Of course the more they know, and use their experience too.)

Yes, You ``look ahead'' --- but not in the same sense that a =computer=
looks ahead: by exhaustively exploring =every= possible move in depth,
sometimes as far ahead as 15 moves...

Numerous studies have shown that a human player conciously explores
only a limited set of moves, of a size closely comparable to the number
of items that player can reliably hold in their short-term memory.
For most humans, this means a short list of between 4 to 7 different
initial moves, depending on the individual. Unlike a computer, even
a novice player is unlikely to even conciously *consider* most massively
stupid moves at all; furthermore, they are quite incapable of thinking
more than five or six moves into the future.

By comparison, similar studies on chess masters have shown that they
do =NOT=, on the average, have a larger short-term memory-list than
novice players. Instead, they appear to differ from novice player
in two significant ways:

1.) They have better subconcious ``tactical filters;'' hence, the
list of moves they consciously hold in their short-term memory are
all tactically better moves on the average than a novice player's
short list would be;

2.) They do not perceive the board as =individual= pieces on squares,
but rather as sets of 2 to 3 pieces organized into >>tactical units.<<
Hence, while their ``short list'' of moves may be no longer than
that of a novice players, their subconscious ``data compression''
into tactical chunks means that they are effectively considering
two to three times as many pieces.

[The above results come from studies of the comparitive ability of
novice and master players to reconstruct the positions of the pieces
on a board they have seen momentarily: masters are much better at this
than novices --- but only if the board comes from an >> ACTUAL GAME.<<
On =RANDOMLY= generated boards, however, chess masters do no better
at reconstructing the board than novices, because their ability to
detect tactical chunks and strategic configurations at a glance is
of little use on randomly-generated boards...]

So, no --- chess masters do not ``look ahead'' the way computers do.
(I believe either Spasky or Fischer once, when asked how many moves
they usually considered, answered ``One --- the _RIGHT_ move...'')

-- Gordon D. Pusch <pu...@mcs.anl.gov>

Disclaimer: I'm a consultant collaborating with Argonne researchers;
I don't speak for ANL or the DOE --- and they *certainly* don't speak
for =ME= !!!

RHA

unread,

May 30, 1997, 3:00:00 AM5/30/97

to

In article <phaflgj...@xanadu.mcs.anl.gov>,
Gordon D. Pusch <pu...@mcs.anl.gov> wrote:
>In article <mkl.86...@rob.cs.tu-bs.de> m...@rob.cs.tu-bs.de
>(Mario Klebsch DG1AM) writes:
>[snip]
>> Almost everyone old enough in the western world can drive a car,
>> although only a few people driving a car only have the chance to
>> beat Kasperov in chess. So it must be much easier to do for a
>> computer. But it isn't. In fact, I naver sa a computer driving a car.

>
>Sorry, but you're a bit out of touch --- the US military has already
>developed a truck that drives itself on standard marked roadways using
>visual cues and pattern-recognition techniques (albeit rather slowly);

Actually, talk to Alan Alda. He did a science show on PBS,
cruising down an interstate...hands off...IIRC, city driving
too. The system doesn't even need standard markings, either,
roadwear marks serve almost as well. The 3 miles/hour military
test vehicle is obsolete.

One robot is even able to play catch.

--
rha

Paul Hsieh

unread,

May 30, 1997, 3:00:00 AM5/30/97

to

On Thu, 29 May 1997 22:39:48 -0700, Jim Balter (j...@netcom.com) said:
> Paul Hsieh wrote:
> > The Deep Blue team has in the past published several
> > papers which have been considered great contributions to the field of AI.
>

> Citations, please? You may be confusing Deep Blue with Deep Thought.

Its the same team you doorknob. Feng, and Campbell have been there since
the "ChipTest" days. That "team" was the one to publish the Deep Thought
papers.

Paul Hsieh

unread,

May 30, 1997, 3:00:00 AM5/30/97

to

On 30 May 1997 01:52:47 -0500, RHA (ri...@praline.no.neosoft.com) said:
> In article <338E12...@ici.net>,
> Alicia Carla Longstreet <ca...@ici.net> wrote:
> >Paul Hsieh wrote:
> [cut]
> >>

> >> On 25 May 1997, H. M. Hubey (hu...@pegasus.montclair.edu) said:
> >> > Alicia Carla Longstreet <ca...@ici.net> writes:
> >> > >As for 'Big Blue' beating Kasperov at Chess, this is nothing more that
> >> > >using brute force to solve a problem, it is most UNimpressive.
> >>
> >> Your knowledge about the internal working of Deep Blue is what is most
> >> unimpressive. The Deep Blue team has in the past published several
> >> papers which have been considered great contributions to the field of AI.
> >> As far as I know they have continued to use some of their more successful
> >> techniques in the Deep Blue program that helped it be more than just a
> >> brute force search engine.
> >
> >Obviously, did it occur to you that I was simply unimpressed with the
> >state of AI today. I have watched them go down so many dead end roads.
> >If they had a REALLY good AI Chess playing program: 1) They would have
> >beaten Kasperov with a PC, 2) Kasperov wouldn't have even drawn any of
> >the games. Again, solving Chess is 'childs play' compared to a real
> >game like Bridge or Go.

So you know something about AI do you? I've strong bridge programs
already, the problem is highly uninteresting. Go is a larger game, and
indeed it is a different problem than chess (I'm sure somebody must have
told you this.) But solve chess? There is nothing child's play about
it.

> What I'd like to know is whether "Deep Blue" is capable of
> "learning" to play better chess.

I don't know, but there are well known AI techniques for doing this.
This avenue has been very well researched in the area of computer
Othello, where computers have learned their own opening books (and
revolutionized Othello opening theory in the process) as well as learned
their own heuristical weightings by neural net and simple least square
approximation methods.

> [...] Would two, or more, Deep Blues playing against each other develop
> a stronger game?

Oh, well, the work done in Othello suggests that you give sort of a
"heuristic space" to learn in, and the program optimizes its variables in
that space via a series of trial games, or imported master games. My
guess is that if Deep Blue was also designed this way, then this is
precisely what it did (also playing against players like Joel Benjamin I
guess) to train itself. But the limit of what the computer can learn is
limited by the predefined "heuristic space" of course.

But to critisize this as not AI due to its bounded nature, I think is
unfair. After all, humans have the same limitation. For example, in
Alicia's case, its clear that her pre-set mental limits preclude her from
any sort of original thoughts, limiting her only to headlines, things
she's heard other people say, reason by analogy, that sort of thing. And
limitations of some kind limit every human in one way or another.

> Or would programmers have to tweak the code?

Programmers can never resist the temptation to tweak code. :o)

Patrick Juola

unread,

May 30, 1997, 3:00:00 AM5/30/97

to

In article <phwwoh3...@xanadu.mcs.anl.gov> pu...@mcs.anl.gov (Gordon D. Pusch) writes:
>[Note --- 'Followup-To:' field brutally trimmed...]
>
>In article <MPG.df64fcfe...@news.zocalo.com>
>nob...@chromatic.com (Paul Hsieh) writes:
>
>> On Thu, 29 May 1997 09:42:43 GMT, Peter Harrison (p...@ihug.co.nz) said:
>>
>> > A human DOES NOT 'look ahead' at all, they develop stratagies,
>> > and use previous experience to apply to the current situation -
>> > even though that situation is not identical.
>>
>> Excuse me? I'm a human, I play chess at an acceptable level.
>> *I* look ahead. All my chess playing friends all look ahead.
>> The strong the player, it seems to me, the more they look ahead!
>> (Of course the more they know, and use their experience too.)
>
>Yes, You ``look ahead'' --- but not in the same sense that a =computer=
>looks ahead: by exhaustively exploring =every= possible move in depth,
>sometimes as far ahead as 15 moves...
>

>Numerous studies have shown that a human player conciously explores...

Need I point out, once again, that what a human player *conciously*
does and what a human player does are two distinct sets?

One of the reasons that a skilled human player sees only the good
moves is because in the process of learning, he's probably seen
(explored) most of the bad ones.

-kitten

Dale Pennington

unread,

May 30, 1997, 3:00:00 AM5/30/97

to

Jim Balter <j...@netcom.com> wrote in article <338E68...@netcom.com>...

> Paul Hsieh wrote:
> > The Deep Blue team has in the past published several
> > papers which have been considered great contributions to the field of
AI.
>

> Citations, please? You may be confusing Deep Blue with Deep Thought.
>

> --
> <J Q B>
>

42

But please, don't ask me what the question is.

Rork D. Kuick

unread,

May 30, 1997, 3:00:00 AM5/30/97

to

People appear to quickly want to get all mystical about our ability
to limit inspection of a large tree to the few branches that count.
When I play chess for each move I consider there are often only 1 or
two reasonable responses, particularly if I am "pressing", and I
don't consider the others. I do this not because it is optimal,
but because I have limited horse-power upstairs, and pruning is
pragmatic. I think I think in terms of "strategies" for the
same reason of being only a mildly smart ape.
The mystification surrounding bridge also has amused me greatly.
Oh yes, something more.

Perhaps folks need reminding that game-theoretically chess is
a FINITE game. It is a semi-famous example of one where no one
has yet shown what the "value of the game" is. That is, with
perfect players, either whites "should" win, or black "should"
force a draw. There is no need for strategies with random components
for perfect players, though they may help us imperfect players.
For a perfect player, chess would resemble tic-tac-toe - utterly
pointless and boring.
I'll be impressed when chess scientists prove what the value of
the game is, in this technical sense.
-rork

Andy Ylikoski

unread,

May 30, 1997, 3:00:00 AM5/30/97

to

In article <mkl.86...@rob.cs.tu-bs.de>, m...@rob.cs.tu-bs.de (Mario Klebsch DG1AM) writes:
|> Almost everyone old enough in the western world can drive a car,
|> although only a few people driving a car only have the chance to beat
|> Kasperov in chess. So it must be much easier to do for a computer. But
|> it isn't. In fact, I naver sa a computer driving a car.

Patrick Henry Winston's (excellent!) book Artificial Intelligence,
Third Edition mentions a neural net system called ALVINN which can
drive a car. (I understand it can only direct a car along a road but
not carry out more nontrivial tasks such as the correct behaviour at
traffic lights.)

all the best, Andy Ylikoski
graduate student
Helsinki Univ of Tech

Andy Ylikoski

unread,

May 30, 1997, 3:00:00 AM5/30/97

to

Mario Klebsch DG1AM:
>Oh, not at all. Playing (is the machine really playing?) chess is only
>one point, where the machine is getting better. But what about some
>really "simple" taskes, compared to chess.

In his book The Society of Mind, Marvin Minsky mentions that the
researchers who built the SHRDLU blocks-world system (which is
familiar to many, many AI researchers) came to the conclusion that
approximately a thousand micro-skills would probably be needed to
enable a little child fill a pail with sand.

But with the current pace of the development of computer science we
will probably rather soon see robot systems which have a thousand
nontrivial software routines.

Randy Kaelber

unread,

May 30, 1997, 3:00:00 AM5/30/97

to

Mario Klebsch DG1AM (m...@rob.cs.tu-bs.de) wrote:

> Almost everyone old enough in the western world can drive a car,
> although only a few people driving a car only have the chance to beat
> Kasperov in chess. So it must be much easier to do for a computer. But
> it isn't. In fact, I naver sa a computer driving a car.

I have. There are prototype systems for computer controlled cars that can
operate them far better than any human possibly could. Right now, they
operate in pretty controlled conditions, but it's only a matter of time
before this technology winds up in new cars. Imagine setting your car to
computer control and buzzing down the Interstate/Autobahn/Highway at
150km/hr with a following distance of only 5 to 10 feet between cars.

Oh, I forgot, people do that anyway. :) OK, Imagine that going on with no
fatal crashes. The technology is there, it just needs some refinement,
and economy of scale, and we'll have auto-drive.

Followups trimmed.... the language groups arewn't really relevant.
--
Randy Kaelber: kael...@muohio.edu http://avian.dars.muohio.edu/~randy/
DARS Programmer/Analyst, Miami University, Oxford, OH 45056 USA

Monument

unread,

May 30, 1997, 3:00:00 AM5/30/97

to

In comp.ai.alife RHA <ri...@praline.no.neosoft.com> wrote:
: What I'd like to know is whether "Deep Blue" is capable of
: "learning" to play better chess. Would two, or more, Deep

: Blues playing against each other develop a stronger game?

: Or would programmers have to tweak the code?

Well, it would actually depend on the initial structure and
programming of the programs themselves. If it were a GA with a
mutation rate that was set just right, it is possible for two GA's to
learn from each other without human influence, based on what each
other did.

It is my understanding, and I'm sure I'll be corrected if I am wrong,
that Deep Blue was programmed with a NNet/GA combination, therefore,
in my opinion, it is possible for Deep Blue to learn from playing
itself.

--
----------------------------------------------------------------------------
| Jeff Goslin - Monument | "Oh Bentson, you are so |
| jggo...@vela.acs.oakland.edu | mercifully free from the |
| | ravages of intellect." |
| http://www.acs.oakland.edu/~jggoslin | --Evil, The Time Bandits |
----------------------------------------------------------------------------
| how come everyone elses religion is a cult but your cult is a religion |
----------------------------------------------------------------------------

John J. Stafford

unread,

May 30, 1997, 3:00:00 AM5/30/97

to

Have the big blue developers ever pitted the machine against itself?

Gordon D. Pusch

unread,

May 30, 1997, 3:00:00 AM5/30/97

to

In article <5mm8d6$p...@news.ox.ac.uk> pat...@cogsci1.psych.ox.ac.uk
(Patrick Juola) writes:

> Need I point out, once again, that what a human player *conciously*
> does and what a human player does are two distinct sets?
>
> One of the reasons that a skilled human player sees only the good
> moves is because in the process of learning, he's probably seen
> (explored) most of the bad ones.
>
> -kitten

<*!ROTFL!*> You aren't seriously proposing that humans play chess
by __rote memory__, are you ?!?!?

Do you have any _idea_ how many possible legal chess moves there are?
The number of possible moves exceeds the number of synapses in the brain
by about __twenty orders of magnitude__ !!!!!!!

No --- as numerous experiments have shown, humans that play well do so
because they have built up a heuristic set of small, strategically-
important configurations of =good= moves, NOT because they've built
up an exhaustive mental catalog of ``bad'' ones. (For one thing,
there haven't even been enough seconds since the _Big Bang_ to explore
them all !!!)

The ONLY parts of chess where rote memory plays a role are the first
few moves of the opening, and the last few moves of the end-game ---
where (not coincidentaly!) the number of pieces are small, and the set
of possible legal moves is _highly_ constrained...

Chris Keane

unread,

May 30, 1997, 3:00:00 AM5/30/97

to

In article <338E12...@ici.net>, Alicia Carla Longstreet wrote:

> If they had a REALLY good AI Chess playing program: 1) They would have
> beaten Kasperov with a PC

Hmmmm. And I suppose that if Kasparov were a REALLY good chess player he
would have beaten Deep Blue with half a brain and no arms...?

Paul Hsieh

unread,

May 30, 1997, 3:00:00 AM5/30/97

to

On Fri, 30 May 1997, John J. Stafford said:
> Have the big blue developers ever pitted the machine against itself?

My theory on the matter is that they've done in thousands, if not
hundreds of thousands of times in order to get it to "stabalize" its
heuristic functions.

chand

unread,

May 30, 1997, 3:00:00 AM5/30/97

to

This chess/bridge stuff is being posted to the wrong group.
This is the Alife group. Thank you.

Jim Paris

unread,

May 30, 1997, 3:00:00 AM5/30/97

to

>Mario Klebsch DG1AM (m...@rob.cs.tu-bs.de) wrote:
>> Almost everyone old enough in the western world can drive a car,
>> although only a few people driving a car only have the chance to beat
>> Kasperov in chess. So it must be much easier to do for a computer. But
>> it isn't. In fact, I naver sa a computer driving a car.

Computers would have no problem driving if it weren't for all the
humans on the road. There's nothing difficult with following road
markings and deciding where you want to go. The hard part is to try to
process all of the stupid things that other drivers do (suddenly brake,
turn, crash, not follow signs, go through red lights, etc) and compensate
for that. That's why computers have no problem flying airplanes but you
don't see them speeding down a highway at 65 miles an hour.

-j...@jtan.com

Daniel P Hudson

unread,

May 30, 1997, 3:00:00 AM5/30/97

to

>>Chess is a perfect-information game, meaning that everything that
>>you know is also known to your opponent, and vice versa. Bridge
>>is not -- at least two hands are unknown to you, and your hand is
>>not known to anyone else.

I take it you don't play chess. The whole object is to make your
opponent think you're going to do one thing while you do another.
Your opponent, excluding foolish beliefs they may be psychic, does
not KNOW what you are doing, that can only make a guess at what your
strategy is. Players who concentrate of the move you just made, rather
than the formation of your pieces, are generally poor chess players.
They make mistakes like jumping at simple "baited" attempts
to gain their pieces, such as placing a semi-important piece
in an easy attack in order for your oponent to move a key piece
interfearing with your attack, etc..

It's hard to put someone in checkmate if they know what you are
doing unless you've eliminated all other pieces except the king,
which was in itself a strategy. You must also adapt to an ever
changing environment. One move by your opponent may cause your
current strategy to be utterly useless. Chess is not a aim for
the king and hope you are lucky game, at least not when you are
playing the game they were. Furthermore, depending on which rules
you use, you may be limited to a certian number of moves before
declaring a stale mate, which is chess's version of a tie which
may also be a strategy to use when you are loosing. Placing yourself
in stale mate by setting your king ina position where he may not move
without being in check, but is not in check yet declares a tie.
--
Do the world a favor, aggervate a nerd:
Name: Nudds, Scott Phone # (905) 388-2969
Address: 480 Upper Kenilworth Ave. Hamilton, Ontario Canada
Suggestion: Egg the house or prank call the dweeb.

ccr...@pacific.net

unread,

May 30, 1997, 3:00:00 AM5/30/97

to

In <338eb2af...@news.ihug.co.nz>, on 05/30/97
at 11:10 AM, p...@ihug.co.nz (Peter Harrison) said:

:For a computer to be programmed in the same manner as chess to play

:bridge the 'search' explodes into a great many more possibilities - to
:the point of being useless.

To the contrary, the number of possible "moves" in bridge is quite small
in comparison to chess. Consider the play of the hand. The maximun number
of possible plays at the start of the hand is only 13 per player, and this
number goes down by one after each play, until after 13 plays, the hand is
over. In practice, the number of possible plays is much less than this,
because players are required to follow suit. The only "gotcha" is that you
can only see two of the four hands, but this can be handled in the same
way human players do -- by calculating the percentages that key cards are
in one ot the other of the unseen hands.

The real issue is that search techniques can not be used in the bidding,
because there is no way to evaluate the nodes. However, over the years,
human players have developed "conventions" to describe their holdings to
their partners -- and the opponents get to listen in.

Best,
-- Chuck Crayne
-----------------------------------------------------------
ccr...@pacific.net
-----------------------------------------------------------

ccr...@pacific.net

unread,

May 30, 1997, 3:00:00 AM5/30/97

to

In <christian.bau-2...@christian-mac.isltd.insignia.com>, on
05/28/97
at 09:22 AM, christ...@isltd.insignia.comm (Christian Bau) said:

:I am wondering (and I have never seen this discussed) how much the
:computers playing "style" is determined by the bit of code that computes
:the value of each position.

Almost totally. The heart of the minimax search strategy is the evaluation
function. In the case where pieces are captured, the evaluation functions
work fairly well, as the relative value of the chess pieces has long since
been established. However, evaluating positional advantage is much more
difficult.

Jim Balter

unread,

May 30, 1997, 3:00:00 AM5/30/97

to

Paul Hsieh wrote:
>
> On Fri, 30 May 1997, John J. Stafford said:
> > Have the big blue developers ever pitted the machine against itself?
>
> My theory on the matter is that they've done in thousands, if not
> hundreds of thousands of times in order to get it to "stabalize" its
> heuristic functions.

Mr. Hsieh appears to be totally ignorant on the matter beyond the
fact that some of the people involved with Deep Thought were also
involved with Deep Blue. However, these two programs use quite
different technology. Therefore, Hsieh's speculations are worse than
random guesses.

--
<J Q B>

Jim Balter

unread,

May 30, 1997, 3:00:00 AM5/30/97

to

Paul Hsieh wrote:
>
> On Thu, 29 May 1997 22:39:48 -0700, Jim Balter (j...@netcom.com) said:

> > Paul Hsieh wrote:
> > > The Deep Blue team has in the past published several
> > > papers which have been considered great contributions to the field of AI.
> >
> > Citations, please? You may be confusing Deep Blue with Deep Thought.
>

> Its the same team you doorknob. Feng, and Campbell have been there
since
> the "ChipTest" days. That "team" was the one to publish the Deep Thought
> papers.

Look, you fucking asshole: Deep Blue and Deep Thought are designed
differently, intentionally. The mere fact that the same people are
involved doesn't mean that the same technology is. So provide some
citations concerning "Your knowledge about the internal working of Deep
Blue" or shut the fuck up.

--
<J Q B>

Jim Balter

unread,

May 30, 1997, 3:00:00 AM5/30/97

to

Patrick Juola wrote:

> >Bluffing is an integral part of game theory -- a game strategy must
> >not be predictable if no pure strategy dominates.
>
> Ah, but you see, a pure strategy *does* dominate chess; that's the
> essence of the lookup-table algorithm.

I never claimed that there is any bluffing in chess.

--
<J Q B>

Jim Balter

unread,

May 30, 1997, 3:00:00 AM5/30/97

to

Gordon D. Pusch wrote:

> In article <MPG.df64fcfe...@news.zocalo.com>
> nob...@chromatic.com (Paul Hsieh) writes:
>
> > On Thu, 29 May 1997 09:42:43 GMT, Peter Harrison (p...@ihug.co.nz) said:
> >
> > > A human DOES NOT 'look ahead' at all, they develop stratagies,

^^^^^^

> > Excuse me? I'm a human, I play chess at an acceptable level.
> > *I* look ahead. All my chess playing friends all look ahead.

> Yes, You ``look ahead'' --- but not in the same sense that a
^^^^^^^^^^^^^^^^^^^^^^^^^

Learn to read for comprehension.

--
<J Q B>

Andy Ylikoski

unread,

May 30, 1997, 3:00:00 AM5/30/97

to

Alicia Carla Longstreet <ca...@ici.net> writes:

>> > It also happens to be the case that Kasparov plays chess by
>> > using brute force. But because it is done via the human
>> > brain, we are told that it is intelligence.
>
>Not very likely. I doubt that Kasparov examines even a small percentage
>of the possible sequences.

Richard Re'ti was one of the best chess players in the world in the
beginning of this century. Somepne asked him how many moves he
usually considers when choosing his next move. Re'ti answered:
"Usually, I make no calculations."

Chris Adams

unread,

May 31, 1997, 3:00:00 AM5/31/97

to

On 28 May 1997 06:20:24 GMT, Axel-Stephane Smorgrav

<axel-stepha...@oslo.itservice.telenor.no> wrote:
>|> Almost everyone old enough in the western world can drive a car,
>|> although only a few people driving a car only have the chance to beat
>|> Kasperov in chess. So it must be much easier to do for a computer. But
>|> it isn't. In fact, I naver sa a computer driving a car.

>Only a fraction of people driving cars can fly an airplane, and most
>will agree that flying (in their opinion) is more difficult than driving
>a car. Still, computers that you claim are not able to drive cars, can
>control an aircraft and even bring it safely back to the ground...

This probably has more to do with the fact that driving a car involves
considerably more effort to avoid/counteract the flaws of the other humans
driving. By sheer lack of volume flying a plane would be easier...

Kaz Kylheku

unread,

May 31, 1997, 3:00:00 AM5/31/97

to

In article <5mmh7r$k...@huron.eel.ufl.edu>,

Daniel P Hudson <afn0...@freenet2.afn.org> wrote:
>Do the world a favor, aggervate a nerd:
>Name: Nudds, Scott Phone # (905) 388-2969
>Address: 480 Upper Kenilworth Ave. Hamilton, Ontario Canada

Careful: the Nudds we know could be an impostor, dragging through the mud the
name of the individual who lives at 480 Upper Kenilworth Avenue.

John Kugelman

unread,

May 31, 1997, 3:00:00 AM5/31/97

to

Axel-Stephane Smorgrav wrote:
>
> In article <mkl.86...@rob.cs.tu-bs.de>, m...@rob.cs.tu-bs.de (Mario Klebsch DG1AM) writes:

> |> Almost everyone old enough in the western world can drive a car,
> |> although only a few people driving a car only have the chance to beat
> |> Kasperov in chess. So it must be much easier to do for a computer. But
> |> it isn't. In fact, I naver sa a computer driving a car.
>
> Only a fraction of people driving cars can fly an airplane, and most
> will agree that flying (in their opinion) is more difficult than driving
> a car. Still, computers that you claim are not able to drive cars, can
> control an aircraft and even bring it safely back to the ground...
>

> -ascs

But they don't have to worry about rush hour!

Rotes Sapiens

unread,

Jun 1, 1997, 3:00:00 AM6/1/97

to

On Tue, 27 May 1997 09:36:17 -0500, "Dale Pennington"
<dpen...@ingr.com> wrote:

>Walter <wal...@and.nl> wrote in article <338AC...@and.nl>...

>> (Just my opinion..)
>> I think it is pretty impressive. Playing chess is something else than
>> calculating 1+1. Kasparov seemed very surprised by Deep Blue's
>> performance and he was very angry that he lost... :)

>> However, I also think Kasparov could beat Deep Blue in a next game
>> just the same. Deep Blue may not have won the war yet.

>Actually, as a friend of mine commented, Big Blue had a major unfair
>advantage. It (actually, it's programming team including at least one chess
>grand master) had plenty of time to analyze Kasparov's playing style and
>determine the weaknesses in his game. It could look at his past history of
>playing, as well as the books he had written. How much do you think
>Kasparov knew of Big Blue's style prior to the match (can you say 0). So
>Big Blue started with a major strategic advantage that no human player in
>that class would ever have.

Kasparov claimed he changed his game playing to match the computer's.
It seems he did worse after this. If somehow he hadn't had access to
the computer's games he wouldn't have changed his playing technique.
If he hadn't changed his playing technique he may have won.

Walter

unread,

Jun 2, 1997, 3:00:00 AM6/2/97

to

Just a remark: Computers that drive cars already exist. Mercedes has
manufactured a car with camera's all over it and the computers in it can
drive it. They tested it on the german autobahn at about 80 miles per
hour. It drove perfectly well, and did not cause any accidents.
It just wasn't any good at parking (yet).

In a way, driving is like chess. You have a set of rules, and computer
only needs to follow the rules. The computer also 'knows' how to reach
certain goals. Don't humans work that way too?
--
*** When you build it, they will come ***

walterj$and.nl walter$ronix.ptf.hro.nl
People should be able to e-mail me, spambots should not.
http://www.ptf.hro.nl/~walter/
Go to BrintaBBS: telnet://bbs.hro.nl where I am known as 'Walter'

Patrick Juola

unread,

Jun 2, 1997, 3:00:00 AM6/2/97

to

In article <5mmh7r$k...@huron.eel.ufl.edu> afn0...@freenet2.afn.org (Daniel P Hudson) writes:
>>>Chess is a perfect-information game, meaning that everything that
>>>you know is also known to your opponent, and vice versa. Bridge
>>>is not -- at least two hands are unknown to you, and your hand is
>>>not known to anyone else.
>
>I take it you don't play chess.

Actually, I do -- not anywhere near Kasparov's caliber, of course.

> The whole object is to make your
>opponent think you're going to do one thing while you do another.
>Your opponent, excluding foolish beliefs they may be psychic, does
>not KNOW what you are doing, that can only make a guess at what your
>strategy is. Players who concentrate of the move you just made, rather
>than the formation of your pieces, are generally poor chess players.

But my opponent has complete information about the formation of my
pieces.

>They make mistakes like jumping at simple "baited" attempts
>to gain their pieces, such as placing a semi-important piece
>in an easy attack in order for your oponent to move a key piece
>interfearing with your attack, etc..

And the consequences of these moves are all obvious "on the board."

I begin to suspect that you don't play chess yourself...

It's possible to try to psych out the opponent. I've never considered
it particularly effective (at either end).

Think of it this way -- if so much of chess is psychological, then
how come IRC or postal chess works?

-kitten

Patrick Juola

unread,

Jun 2, 1997, 3:00:00 AM6/2/97

to

Right. My point, if you want it in mathematical lingo, is that
(computer solution of) chess is simpler than bridge because chess is
provably dominated by a pure strategy and bridge isn't.

You might be amused by a recent poster who claimed that chess *does*
have bluffing.

-kitten

Chris Keane

unread,

Jun 2, 1997, 3:00:00 AM6/2/97

to

In article <5mmh7r$k...@huron.eel.ufl.edu>, Daniel P Hudson wrote:

> >>Chess is a perfect-information game, meaning that everything that
> >>you know is also known to your opponent, and vice versa. Bridge
> >>is not -- at least two hands are unknown to you, and your hand is
> >>not known to anyone else.
>

> I take it you don't play chess. The whole object is to make your

> opponent think you're going to do one thing while you do another.

> [etc., etc.]

All of what you say is absolutely true. It doesn't contradict what the
original posting said, though: in chess, _all_ the information you need to
know what your opponent _can_ do is available to you. What he's _going_
to do is another matter completely, but that's not the point of what was
being said.

Randy Kaelber

unread,

Jun 2, 1997, 3:00:00 AM6/2/97

to

John Kugelman (kuge...@mnsinc.com) wrote:

> But they don't have to worry about rush hour!

You've never flown through O'Hare or Dallas-Fort Worth, have you? :)

Trimmed followups greatly. Advice to others in this thread is to do
likewise.

Gary Forbis

unread,

Jun 2, 1997, 3:00:00 AM6/2/97

to

Chess is neither known to be won or lost for the first player. Opening
books lead one into the middle game where one generally relies upon
heuristics. Since the set of possible middle game positions are many
a lookup table is not viable at this time. A lookup table will only
become
viable when one side or the other may play a winning strategy that
involves
traversing a game tree. A single lookup table would need to include not
only position but legal castling, times entered, moves 'til stalled
game,
(etc.?) because these are sometimes important variables.

Rock, paper, scissors is a quite simple game. One who plays a pure
strategy
(other than random selection) is open to counters by an adaptable
opponent.

Since most middlegame positions encountered in tournament play are not
known win or loose positions what do you mean by "chess is provably
dominated by a pure strategy?" Such a notion is counter intuitive to
me.

--
--gary
for...@accessone.com

Chris Adams

unread,

Jun 3, 1997, 3:00:00 AM6/3/97

to

On Mon, 02 Jun 1997 09:07:38 +0200, Walter <wal...@and.nl> wrote:
>Just a remark: Computers that drive cars already exist. Mercedes has
>manufactured a car with camera's all over it and the computers in it can
>drive it. They tested it on the german autobahn at about 80 miles per
>hour. It drove perfectly well, and did not cause any accidents.
>It just wasn't any good at parking (yet).

>In a way, driving is like chess. You have a set of rules, and computer
>only needs to follow the rules. The computer also 'knows' how to reach
>certain goals. Don't humans work that way too?

Of course, with driving you have to deal with chaotic humans; perhaps
chess with a random change every 5 moves?

--
# <ada...@io-online.com.NoUCEsPlease> | <cad...@acucobol.com.NoUCEsPlease>
# (Remove everything to the right of .com to reply & sorry about the hassle)
"Businesses will continue to suffer intrusions until they successfully
remove the motive behind them. Usually this is reputation - for a
hacker taking down microsoft.com would be like blowing up the Death Star"

Hans Moravec

unread,

Jun 3, 1997, 3:00:00 AM6/3/97

to Andy Ylikoski

> Patrick Henry Winston's (excellent!) book Artificial Intelligence,
> Third Edition mentions a neural net system called ALVINN which can
> drive a car. (I understand it can only direct a car along a road
> but not carry out more nontrivial tasks such as the correct
> behaviour at traffic lights.)
- Andy Ylikoski

For RALPH, the NON neueral successor to this work, which learns
much faster and better, see:

http://almond.srv.cs.cmu.edu/afs/cs/user/tjochem/www/nhaa/nhaa_home_page.html

which describes an 98.2% autonomous, 5,000 kilometer drive from
Washington D.C. to San Diego, CA, done in July 1995. It think
this is the longest stretch of robotic driving to date. There
are autonomous vehicles driving the roads of Europe these days
as well.

-- Hans Moravec CMU Robotics

Tom Thornhill

unread,

Jun 3, 1997, 3:00:00 AM6/3/97

to

Andy Ylikoski <ylik...@alpha.hut.fi> wrote in article <svy3er5...@alpha.hut.fi>...

I think he meant that

"Usually, I make no [conscious] calculations."

--
Mailto: ttho...@best.ms.philips.com-nospam ( remove youknowwhat b4 replying)
Talkto: Tom Thornhill, Philips Medical Systems Nederland BV
Tel +31 4027 64080 ( Work ), +31 499 371230 ( Hotel - ask for room 19 )
************************ +44 3709 46045 GSM ****************************
**** IMPORTANT: please use the above numbers rather than the GSM *******
**** one if possible as I pay for incoming calls from outside the ****
**** Netherlands. *****************************************************

Patrick Juola

unread,

Jun 3, 1997, 3:00:00 AM6/3/97

to

In article <3392EE...@accessone.com> Gary Forbis <for...@accessone.com> writes:
>Patrick Juola wrote:
>>
>> In article <338F95...@netcom.com> Jim Balter <j...@netcom.com> writes:
>> >Patrick Juola wrote:
>> >
>> >> >Bluffing is an integral part of game theory -- a game strategy must
>> >> >not be predictable if no pure strategy dominates.
>> >>
>> >> Ah, but you see, a pure strategy *does* dominate chess; that's the
>> >> essence of the lookup-table algorithm.
>> >
>> >I never claimed that there is any bluffing in chess.
>>
>> Right. My point, if you want it in mathematical lingo, is that
>> (computer solution of) chess is simpler than bridge because chess is
>> provably dominated by a pure strategy and bridge isn't.
>>

>Chess is neither known to be won or lost for the first player.

True but irrelevant. Mathematical proofs don't generally pay attention
to minor details like the heat death of the universe, which are
considered to be nuisances but not interruptions.

>Rock, paper, scissors is a quite simple game. One who plays a pure
>strategy
>(other than random selection) is open to counters by an adaptable
>opponent.

Rock, paper, scissors is also not a perfect-information game. You'd
get a better metaphor examining noughts-and-crosses (aka tic-tac-toe).

>Since most middlegame positions encountered in tournament play are not
>known win or loose positions what do you mean by "chess is provably
>dominated by a pure strategy?" Such a notion is counter intuitive to
>me.

Um, "chess is provably dominated by a pure strategy."

"It can be (and has been proven) that there exists a pure strategy
that dominates chess."

The mere fact that it might take longer than the lifetime of the
universe to determine which strategy is not, by game theorists, considered
to be a significant drawback.

-kitten

Derek J. Witt

unread,

Jun 3, 1997, 3:00:00 AM6/3/97

to

Paul Hsieh wrote:

>
> On 25 May 1997, H. M. Hubey (hu...@pegasus.montclair.edu) said:
> > Alicia Carla Longstreet <ca...@ici.net> writes:
> >

> > >When IBM comes up with a computer that can beat me at Bridge, I will be
> > >impressed. I am NOT a great Bridge player, merely a decent one.
> >
> > Most people probably won't be.
> >
> > That's why probably nobody will spend millions making
> > bridge-playing computers. The reason for picking chess is
> > obvious; it has been shown everywhere as the prime example
> > of an intelligence-requiring and an intellectual game.
>
> Actually, I've seen some very strong bridge programs (usually combined
> with a tutorial) out there. But as you suggest, nobody really cares, so
> these programs fell into relative oblivion.
>
> > >As for 'Big Blue' beating Kasperov at Chess, this is nothing more that
> > >using brute force to solve a problem, it is most UNimpressive.
>
> Your knowledge about the internal working of Deep Blue is what is most
> unimpressive. The Deep Blue team has in the past published several

> papers which have been considered great contributions to the field of AI.

> As far as I know they have continued to use some of their more successful
> Actually, Deep Blue *is* a highly parallelized machine. Though, perhaps
> not in the sense you mean.

A single-processor machine can also be parallel. However, the
parallelism is implemented as time-shared slices. These slices are
executed sequentially in a cycle in such a way that the most important
processes (highest priority) are executed more often than the least
important ones (lowest priority). But, it is important to note that
this is NOT true parallelism. This may also be called non-preemptive
multi-tasking for those in the PC world.

Since Deep Blue is a multi-processor machine, it performs true
parallelism. The processes are shared across each processor. This
enables Deep Blue to process many future moves simultaneously without
any difficulty.

>
> --
> Paul Hsieh
> http://www.geocities.com/SiliconValley/9498/mailme.html

--
*************************************************************************
* Derek Witt * senior, computer science *
* 1515 Houston St * Kansas State University *
* Manhattan, KS 66502-4128* - d...@cis.ksu.edu - *
* USA * Home Phone: +1 (913) 537-4708 *
*-----------------------------------------------------------------------*
* "Resistance is futile. You will be assimilated." - Bill Gates / Borg *
*************************************************************************