From: Loyd Willaford <loydw <at> DARKWING <dot> UOREGON <dot> EDU>
Date: 2 sep 1998
Subject: Re: Feng-hsiung Hsu

>>Feng-hsiung Hsu, the man whose program defeated the world chess
>>champion is looking for new worlds to conquer.  In this article, he
>>indicated that shogi might be next....
>
>What do people think about this?  My personal opinion is that while the
>possible media attention would be nice in countries like the US, if a
>similar loss occurs for shogi professionals, much of the artistic and
>traditional value of shogi would be lost.  (Can you imagine professional
>shogi players using computers during their matches like some chess
>tournaments are organized now?)  That is, of course, if the computer
>would win ;)  (Unfortunately, I also think the brute force method would
>work - even if the results are rather uninspiring.)
>
>As i said at the top, are there any other opinions on this?
>
>Patrick Lacz
>
>
>______________________________________________________
>Get Your Private, Free Email at http://www.hotmail.com
>
>

        I rather doubt that any shogi program will even come close to
beating a shogi professional for a while, it took 25 years of intense
development to create the machine that eventually "beat" Kasparov (in the
third incarnation), and even then they programmed it with his games and did
not allow him to see training matches. This deprived Kasparov of an
important tool for finding weaknesses in the program's game, this was also
one reason the match was short (only six games, ending 3.5-2.5).  I have no
doubt that in a full 24 game match for the World Championship that Kasparov
would have and still could rout Deep Blue.  Deep Blue might have a made an
impact on the world chess scene for a few months, but without full time
tinkering would get beaten even by mediocre Grandmasters after a few full
tournament appearances, that is how much effort is  required to stay on top.
I think
        My limited understanding of shogi and  computer shogi leads me to
believe that a similar pattern would hold.  I believe (please correct me if
I'm wrong) that computer shogi is really in in its infancy.  While many
lessons from the chess programs apply, there are a number of unique features
of shogi(especially dropping pieces back in the game) that I imagine
computers will have a very hard time with.  "Brute Force" is simply not an
option is complex positions.
        In any event, computers have not ruined chess, and database programs
have really equalized the availibility of theoretical knowledge. Computers
are not normally used during games.  I would guess that many Japanese shogi
professionals already use electronic databases and will use programs that
play in a manner similar to the way chess professionals use chess programs,
as a supplement to human analysis.
        In conclusion, I think we have very little to fear from further
development of computer shogi and a great deal to gain, provided we are
cautious and if we have human vs computer matches that we make that
correspond as closely to human vs human match conditions, so as to avoid the
hype that surrounded Kasparov's "loss"

Just a few naive comments,

Loyd Willaford