Computer chess history : Julio Kaplan interview (late 1985)

[Vinvin]

I plan to spend a lot more time on chess history for next weeks (or months).
I saw a lot a French articles with details I don't see on English sites.
BTW : who is admin on http://chessprogramming.wikispaces.com ?
I could write articles directly on this site if I'm allowed to ...

- I could add some reports from Aubervilliers tournament (a lot of comp-humans games) in the 80-90's.
- Games Humans vs Hitech from http://hobbiesdejp.free.fr/Sommaire/Ech ... atique.pdf
- Details from : Deep Thought vs humans at Boston open 1988 ; Hitech vs Arnold Denker
- ...


For now, here's an Interview from Julio Kaplan by Pierre Nolot in "Jeux & stratégie" magazine from dec 1985.
Scanned original is here (bottom and right) : http://download.abandonware.org/magazin ... -01%29.jpg
second page here (first column) : http://download.abandonware.org/magazin ... -01%29.jpg

Transcript with OCR + manual corrections is here : http://home.scarlet.be/vincentlejeune/l ... ammeur.txt

Translation with google :

----------------------------------
THE CHAMPION BECOME PROGRAMMER

It is now a well-known paradox
chess programmers are generally
no strong players. David Kittinger
Ulf Thomas Rathmann and
Nitsche worth around 1800
ELO is not to say
more than their program, and
even quite a bit less
anymore. (They are the fathers
respective Constellation
Expert Princhess and Orwell). The
height is achieved with Kathe
and Dan Spracklen worth

1500 ELO, while their Elite
Challenger can be evaluated
1800! This is when their
program plays the "parents"
are not always able
whether their "child" is
"Fairly well" or not on the board.
And Amsterdam, we saw
David Kittinger, Richard Lang
or ask Ulf Rathmann
often one of their colleague
programmer what to think
position. The latter,

----------------------------
-Picture-
Kaplan: "My program will
make serious progress in
next two years!"
----------------------------

nice redhead 35 years
cast a quick glance at
the board, and explained almost
immediately forces
and weaknesses of each
camp. It must be said that
Scisys programmer, which
designed the Turbostar is not
a stranger in the world of
failures: it is Julio Kaplan.
Kaplan was at 17 in 1967,
world junior champion in
the colors of Puerto Rico.
International Master, his ELO
reached in 2480 (no French
never happened in this figure).
Now, Julio Kaplan
completely stopped practicing
competition, and since 1980.
He devoted himself entirely to the
programming. We
met this player programmer,
Amsterdam during the
World Championship
electronic chess.

Games & Strategy: Why
Have you completely stopped
competition here 5 years, while
Were you one of the best
U.S. players, and
you seem to still have a
beautiful room for improvement
before you?

Julio Kaplan: I was born in
Argentina in 1950. When
I was 13, my parents
parties in Puerto Rico. This is
Puerto Rican as I am
become world champion
Junior 17 years that got me
the title of International Master
elsewhere. Then I went to
United States continue
computer studies. I took
U.S. citizen, my
since my third nationality
birth then! And then I
'm set to make the big tournaments
Americans like Lone
Pine example.
But in 1980, he took me to
a choice: do I spend
as failures, become Grand
Master, rise above
2500 ELO, and for this I
had to sacrifice everything, including
including my family, or
well lead a more normal life
and in this case, have an
profession less random.
This second choice I made.
I then completely
abandoned failures
competition, because there continue
devoting some
hours per week would
resulted only cons-
performance. It would have been very
hard to bear especially when
that was one of the best ...

J & S: In 1980, then you are
programmer become failures.
How this transition is
made?

J.K.: I studied
top computer.
With my level chess,
this interested first firm
American Mattel, which I
asked him to pro-
grams from 1979 to 1981. Then
firm in Hong Kong, Scisys
I was hired in 1981 to
enough to make a program
modest Sensor-Chess.
Then, in 1982, there has been President
still playing very badly,
it must be admitted, then
Superstar, the first program
I've been pretty satisfied, and
Finally Turbostar that does not
bursts in the championship
Amsterdam is true, but
which achieved 6.5 in 12
the U.S. Open this year.
This gives it a ranking
approximately 2150 ELO. Not
wrong in four years, right?

J & S: To a very good
chess program, it is
better be a good programmer
or a very strong chess player?

J.K.: Until 1983, it was clear
it better be primarily
one as a programming
chess master. But
Things began to change
two years. Take
for example the programs
Kathe and Dan Spracklen of: they
were the best, but they
are no longer, because they
can no longer grow. Their
authors are not enough
strong players! Now, I
seems that there is a big advantage
to fully understand the
strategy of the game, so to be a
strong chess player.

J & S: Do you power
much better
your program?

J.K.: Yes, I still
many ideas to make
Turbostar stronger. I know
weaknesses: its lack
aggressiveness, lack of

the protective function of the King
etc.. I will remedy this, and
seriously improve in
next two years.

J & S: The world champion
Garri Kasparov, works for
Scisys, at least it lends
behalf. Is it really serves
to that / that thing in
since it is not a programmer?

J.K.: There were two Turbostar
championship in the United States
Open 1985. One was the
normal model, but turning
50% faster, the other turned
to the normal speed of
4 MHz, but it was
module with Kasparov. They
both have marked
6.5 points out of 12, but the
Turbostar-Kasparov played against
opponents much more
strong. Conclusion: the classification of
Turbostar speed is pushed
2050 ELO, one of Turbostar-
Kasparov, 2150 ELO. This is
no comment.
Kasparov is certainly not
programmer, but it looks
much to this program
it is relevant ideas,
this is a fantastic collaborator.

J & S: The Championship
world of microphones is
triumph for programs
Mephisto designed by Richard
Lang, how do you explain
their overwhelming superiority?

J.K.: Richard Lang has achieved a
beautiful combination of "hard"
and "soft". The program has
no major weaknesses,
plays a very safe and
advantage of mistakes opposing
is efficient but not spectacular.
But I recall that
against the simultaneous
Sosonko the Turbostar have
festival by beating Grand
Master in three parts, then
Mephisto that have marked
half a point. My pro-
program is mainly
to play against players
not against other machines.
One day or another, it will be
make programming
16-bit is certainly
much more efficient for
chess concepts, but
we will do in Scisys
that when it is possible
to offer these programs to the public
for a reasonable price, this is
far from being the case today.
Turbostar has not won
Amsterdam, but the ratio
quality / price it is certainly
better now.

Pierre Nolot
----------------------------------

My best,
Vincent

[IGarcia]

Very nice and interesting!
Thanks

[Steve Maughan]

Truly fascinating! It's interesting to see that back then they thought chess knowledge would be much more important than it has turned out to be. Also it's interesting hearing them talk about the speed of the processors - a typical smartphone is now x160 faster!

Thanks for posting,

Steve

[Vinvin]

Google makes the same translation mistake many times : "failure" instead of "chess". Because in French both can be written "échecs"

[Gerd Isenberg]

I plan to spend a lot more time on chess history for next weeks (or months).
I saw a lot a French articles with details I don't see on English sites.
BTW : who is admin on http://chessprogramming.wikispaces.com ?
I could write articles directly on this site if I'm allowed to ...

My best,
Vincent

Hi Vincent,

I happen to be one organizer in cpw. You are very welcome to contribute. If you join cpw (username without spaces), you can edit and add new pages etc...

Some pages in context ...
https://chessprogramming.wikispaces.com/Pierre+Nolot
https://chessprogramming.wikispaces.com/Julio+Kaplan

and some guidiance ...
https://chessprogramming.wikispaces.com ... +New+Pages

Cheers,
Gerd

[Vinvin]

Thanks, I'll try to make a test page !

But first, I'd want to make a big time-line for the period 1950-2000 with mainly :
1) comp-human events and games
2) release/first appearance of engines/chess computers/programs.

Vincent

[Vinvin]

Google makes the same translation mistake many times : "failure" instead of "chess". Because in French both can be written "échecs"

Text corrected :

THE CHAMPION BECOME PROGRAMMER

It is now a paradox
well-known chess
programmers are generally
not good players. David Kittinger,
Ulf Thomas Rathmann and Thomas
Nitsche worth around 1800
ELO is not to say
more than their program, and
even quite a bit less
anymore. (They are the fathers
respectively : Constellation,
Expert Princhess and Orwell). The
height is achieved with Kathe
and Dan Spracklen worth

1500 ELO, while their Elite
Challenger can be evaluated
1800! This is when their
program plays the "parents"
are not always able
whether their "child" is
"Pretty good" or not on the board.
So, in Amsterdam, we saw
David Kittinger, Richard Lang
or Ulf Rathmann
often ask one of their colleague
programmer what to think
position. The latter,
-----------------------------
Picture
Kaplan: "My program will
make serious progress in
next two years! '
-----------------------------
nice redhead 35 years
cast a quick glance at
the board, and explained almost
immediately forces
and weaknesses of each
camp. It must be said that
Scisys programmer, which
designed the Turbostar is not
a stranger in the world of
chess : it is Julio Kaplan.
Kaplan was, at 17, in 1967,
world junior champion in
the colors of Puerto Rico.
International Master, his ELO
reached 2480 (no French
never reached this high).
Now, Julio Kaplan
completely stopped practicing
competition, and since 1980.
He devoted himself entirely to the
programming. We
met this player-programmer,
in Amsterdam during the
World Computer
Chess Championship.

Jeux & Stratégie : Why
Have you completely stopped
competition here 5 years, while
Were you one of the best
U.S. players, and
you seem to still have a
beautiful room for improvement
before you?

Julio Kaplan: I was born in
Argentina in 1950. When
I was 13, my parents
parties in Puerto Rico. That was
in Puerto Rican where I am
become world champion
Junior at 17 ; that got me
the title of International Master
too. Then I went to
United States continue
computer studies. I took
U.S. citizen, so that's my third
nationality since mybirth !
And then I start to play major
American tournaments as
Lone Pine for example.
But in 1980, he took me to
a choice: do I devote my life
to chess, become Grand
Master, rise above
2500 ELO, and for this I
had to sacrifice everything, including
including my family ; or
well lead a more normal life
and in this case, have an
profession less random.
This second choice I made.
I then completely
abandoned chess
competition, because there continue
devoting some
hours per week would
resulted only disappointing performance.
It would have been very
difficult to bear especially
when one was among the best ...

J & S: In 1980, then you became
chess programmer.
How this transition is
made?

J.K.: I made high studied
in computer.
With my level chess,
this interested first firm
American Mattel, which I
asked him Programs
from 1979 to 1981. Then
firm in Hong Kong, Scisys
I was hired in 1981 to
to make a modest program :
Sensor-Chess.
Then, in 1982, there has been "President",
still playing very badly,
it must be admitted, then
Superstar, the first program
I've been pretty satisfied, and
Finally Turbostar that does not
bight in the Amsterdam
championship, that's true, but
which achieved 6.5/12 in
the U.S. Open this year.
This gives it a ranking
approximately 2150 ELO. Not
bad in four years, isn't it ?

J & S: To make a very good
chess program, it is
better be a good programmer
or a very strong chess player?

J.K.: Until 1983, it was clearly
better to be primarily
one as a programming
chess master. But
Things began to change
two years. Take
for example the programs of
Kathe and Dan Spracklen : they
were the best, but they
are no longer, because they
can no longer grow. Their
authors are not enough
strong players! Now, I
feel that's a big advantage
to fully understand the
strategy of the game, so to be a
strong chess player.

J & S: Do you thing
you're able to improve
your program a lot ?

J.K.: Yes, I still
many ideas to make
Turbostar stronger. I know
weaknesses: its lack
aggressiveness, lack of

the protective function of the King
etc.. I will remedy this, and
seriously improve in
next two years.

J & S: The world champion
Garri Kasparov, works for
Scisys, at least it lends
behalf. Is it really serves
something in the
since it is not a programmer?

J.K.: There were two Turbostar
championship in the United States
Open 1985. One was the
normal model, but turning
50% faster, the other turned
to the normal speed of
4 MHz, but it was
module with Kasparov. They
both have scored
6.5 points out of 12, but the
"Kasparov Turbostar" played against
opponents much stronger.
Conclusion: the performance of
the overclocked Turbostar is
2050 ELO, the Turbostar-
Kasparov, 2150 ELO. This is
no comment.
Kasparov is certainly not
programmer, but it looks
closely this program
he have relevant ideas,
he is a fantastic collaborator.

J & S: The Championship
world of microphones is
triumph for programs
Mephisto designed by Richard
Lang, how do you explain
their overwhelming superiority?

J.K.: Richard Lang has achieved a
beautiful combination of "hard"
and "soft". The program has
no major weaknesses,
plays a very safe and take
advantage of opponent's mistakes
it's efficient but not spectacular.
But I recall that
in the simultaneous against
Sosonko the Turbostar made
a firework by beating Grand
Master in three games, in the
same time Mephisto marked only
half a point. My program
is mainly made
to play against players
not against other machines.
One day or another, we will have
to start programming in
16-bit, that is certainly
much more efficient for
chess concepts, but
we will do it in Scisys
when it is possible
to offer these programs to the broad public
for a reasonable price, this is
far from being the case today.
Turbostar has not won
Amsterdam, but the report
quality / price it is certainly
better now.

Pierre Nolot