Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Deep Blue and the

Author: Robert Hyatt

Date: 18:27:27 11/15/98

Go up one level in this thread


On November 15, 1998 at 19:53:32, James B. Shearer wrote:

>On November 14, 1998 at 10:11:16, Robert Hyatt wrote:
>
>>So you think the original post by Amir (written by Shay) was "ok"???  And
>>didn't deserve any sort of response?
>
>       I think the original post contained some debatable statements.  I think
>Bruce Moreland in his first responses made some effective counterpoints.  My
>memory of your posts is that you basically repeated Bruce's arguments in a more
>aggressive way.  I do not think this contributed much to the discussion. If
>someone feels compelled to post but doesn't have anything original to say, I
>would prefer a simple I agree with Bruce (or whoever).
>        I also think Bruce's counterpoints applied more to the details of the
>original post than to the overall thrust.  I believe it is fair to say that:
>     1.  The victory over Kasparov was a tremendous PR coup for IBM.
>     2.  IBM has been reluctant to have Deep Blue demonstrate its strength by
>taking on all comers.
>Was the original post so different as to justify vitriolic attacks?
>                          James B. Shearer



I would have to look back, but I believe that I posted something before Bruce
did on that issue...  I ignored the stock business as that was silly but not
insulting in any way.  But the *rest* of the "article" was outright insulting
to the DB guys...  And I pointed out exactly why...  And asked for some sort of
explanation about how the "evaded further competition" statement could possibly
be made and justified...  I haven't seen an answer to that yet, since there were
no computer events they could have participated in...

As far as deep blue "playing a lot"  I'd simply remind everyone that it is
very easy to "think small".  Most of you haven't *been there*.  I, on the
other hand have, year after year.  It is *not* easy to arrange to use a
multi-million dollar computer to play whenever an event is available.  Cray
Research sponsored me for many years (and probably still would if I asked
again) but getting machine time was *always* difficult.  And I did good to
get dedicated machine time once a year for an ACM or WCCC event (if you look
back in history you will notice we *never* played in both the ACM and the WCCC
if both were held in the same year...  Not because we didn't want to, but be-
cause we couldn't get a 60 million dollar computer dedicated for enough time to
play...)  I'd suspect that IBM *never* gave the DB guys a full SP2 machine for
months on end, to test and play with.  They probably could break out with DB
junior, but if you had a quad-xeon to run crafty, would you want to do
exhibitions using a pentium-90?  So it is easy to say they should have
participated more, but in reality, it would be *very* difficult to justify
sticking a 32 node SP over in the corner for months when it could be sold for
something way over a million dollars, just as it was difficult for us.

Thinking of micros is one thing, and that's why Crafty is now a micro-based
program...  but getting time on a multi-million dollar computer is something
else...  and don't be mislead by the huge PR bonus IBM received...  They would
*still* think a long time about allocating that kind of resource for the amount
of time needed to really "break DB out and let it play."

Ed mentioned they had declined many one-on-one challenges, including one from
Rebel.  Again, same reasoning.  Cray Blitz would have *had* to decline such
challenges, because why in the world would Cray give us a 60 million dollar
computer to compete against a microcomputer? IE *what* would they get in return
for that?  If we won, zilch... wouldn't be news.  If we lost, it would be big
news and it would be all bad from their perspective.  Ditto for IBM.




This page took 0 seconds to execute

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

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