Major Cat wrote:
wrote:
I mean that the computer can number crunch
way in advance (left on overnight, etc.), and find the
best opening books, looking ahead many more moves
than they normally would in regular play. They could
then store these opening moves as a database for
each Fischer random position, and thus have
a time advantage during the openings. The human
would take much more time analyzing an opening position
for perhaps the first time.


Your point is well taken, provided there is no serious, organized
and competent human interest to critically review and counter such
computer-generated opening theory. Basically, it is an empirical
and open-ended question. Time will tell...

P.S. I have never come across any published reference regarding
GMs' assessments of how well powerful computers play their first
10 moves or so in Chess960! 8)

That's a very interesting question.

I have posed this question to this newsgroup on a number
of occasions. Unfortunately, I got no answers whatsoever.


What's the general opinion that GMs have
of computer opening play without the
database on.

For FIDE chess, it has been said that GMs would come out of
the opening with some positional advantage or another. Mind
you, I have no idea how credible this apparent consensus is...


In that case, the computers must better middle and
end-game chess, because they consistently are
beating the GMs. One the world champion and the
very top players are drawing against the machines.




And, have computers verified or
shown erroneous the human research
that has occurs for the past several
hundred years or so, in terms of opening
moves, etc. Certainly computers will
be able to look quite a bit farther down the
game.

I am no expert in FIDE chess opening theory. My understanding
is that computer analysis of openings has contributed to the
re-evaluation of quite a few opening lines due to uncovered
"deep" tactical opportunities or flaws.

Do you have a link to more info?

If computers have created a re-evaluation
of opening moves, that's great, and shows
that the computer can certainly improve on
chess theory.

But the advantages of the computer
analysis will diminish, unless the ability
to memorize more and more moves
is increased in humans.

Hence, my argument that
a sizable part of a great chess
champion is pure, brute force
memorization.


SLick

wrote:
What's the general opinion that GMs have
of computer opening play without the
database on.

For FIDE chess, it has been said that GMs would come out of
the opening with some positional advantage or another. Mind
you, I have no idea how credible this apparent consensus is...


In that case, the computers must better middle and
end-game chess, because they consistently are
beating the GMs. One the world champion and the
very top players are drawing against the machines.

Actually, computers do use their openings database(s) when play-
ing against GMs in "celebrated" matches. Therefore, the consensus
that I alluded to may be more apparent (like an urban tale) than
real...



And, have computers verified or
shown erroneous the human research
that has occurs for the past several
hundred years or so, in terms of opening
moves, etc. Certainly computers will
be able to look quite a bit farther down the
game.

I am no expert in FIDE chess opening theory. My understanding
is that computer analysis of openings has contributed to the
re-evaluation of quite a few opening lines due to uncovered
"deep" tactical opportunities or flaws.

Do you have a link to more info?

Not offhand! You see, FIDE chess opening theory ranks rather low
in my priorities, being so much into Chess_18. 8) Keep in mind
though that, due to rather widespread "macho" attitudes and pre-
dispositions, not that many analysts and GMs are likely to ever
admit that their brilliancies, refutations and so on are the...
brainchildren of powerful computers.


If computers have created a re-evaluation
of opening moves, that's great, and shows
that the computer can certainly improve on
chess theory.

This is generally acknowledged. However, specifics are hard to
come by, or so it seems to this hobbyist...


But the advantages of the computer
analysis will diminish, unless the ability
to memorize more and more moves
is increased in humans.

This is a very good point. I wonder if GMs may already be
experiencing significantly diminishing returns with respect to
a relentlessly expanding opening theory. Nevertheless, GMs
may tacitly "agree" to play only a subset of opening lines as
an attempt to limit the risk of coming across something that
they are not routinely familiar with. In fact, risk minimization
may be an important factor behind opening fads in GM play. I
freely admit that this is only "wild" speculation on my part! 8)


Hence, my argument that
a sizable part of a great chess
champion is pure, brute force
memorization.

If you stick around this forum for a while, you will come across
quite a bit of coverage regarding this. As a Chess_18 enthusiast,
I cannot get too much excited about such memorization feats. But,
the fact remains, FIDE chess at the highest levels seems to call
for heavy duty memorization...

Major Cat wrote:
wrote:
What's the general opinion that GMs have
of computer opening play without the
database on.

For FIDE chess, it has been said that GMs would come out of
the opening with some positional advantage or another. Mind
you, I have no idea how credible this apparent consensus is...


In that case, the computers must better middle and
end-game chess, because they consistently are
beating the GMs. One the world champion and the
very top players are drawing against the machines.

Actually, computers do use their openings database(s) when play-
ing against GMs in "celebrated" matches. Therefore, the consensus
that I alluded to may be more apparent (like an urban tale) than
real...



And, have computers verified or
shown erroneous the human research
that has occurs for the past several
hundred years or so, in terms of opening
moves, etc. Certainly computers will
be able to look quite a bit farther down the
game.

I am no expert in FIDE chess opening theory. My understanding
is that computer analysis of openings has contributed to the
re-evaluation of quite a few opening lines due to uncovered
"deep" tactical opportunities or flaws.

Do you have a link to more info?

Not offhand! You see, FIDE chess opening theory ranks rather low
in my priorities, being so much into Chess_18. 8) Keep in mind
though that, due to rather widespread "macho" attitudes and pre-
dispositions, not that many analysts and GMs are likely to ever
admit that their brilliancies, refutations and so on are the...
brainchildren of powerful computers.


"Chess 18"?



If computers have created a re-evaluation
of opening moves, that's great, and shows
that the computer can certainly improve on
chess theory.

This is generally acknowledged. However, specifics are hard to
come by, or so it seems to this hobbyist...


But the advantages of the computer
analysis will diminish, unless the ability
to memorize more and more moves
is increased in humans.

This is a very good point. I wonder if GMs may already be
experiencing significantly diminishing returns with respect to
a relentlessly expanding opening theory. Nevertheless, GMs
may tacitly "agree" to play only a subset of opening lines as
an attempt to limit the risk of coming across something that
they are not routinely familiar with. In fact, risk minimization
may be an important factor behind opening fads in GM play. I
freely admit that this is only "wild" speculation on my part! 8)


This is conservative versus aggresive play,
and Fischer used it against Spassy during
the world championship. Spassy assumed
that Fischer would always used the openings
he had used in the past, and then was surprised
(as was the chess world) when Fischer pulled
out this English Opening, something he had
never used before in tournament play:

http://www.chessgames.com/perl/chessgame?gid=1044367




Hence, my argument that
a sizable part of a great chess
champion is pure, brute force
memorization.

If you stick around this forum for a while, you will come across
quite a bit of coverage regarding this. As a Chess_18 enthusiast,
I cannot get too much excited about such memorization feats. But,
the fact remains, FIDE chess at the highest levels seems to call
for heavy duty memorization...

Oh definitely, and i've heard that especially with
speed chess, ya gotta know your opening
inside out if you want to do well.

Slick

wrote:

"Chess 18"?

Chess_18 is a subset of Fischer Random Chess (aka Chess960). The
King and Rooks always start on their "orthodox" squares. Castling
is "orthodox" through and through as well. Chess_18 comprises 18
starting arrays, all enjoying independent theoretical significance.
The "orthodox" (FIDE) array is one of them! 8)


This is a very good point. I wonder if GMs may already be
experiencing significantly diminishing returns with respect to
a relentlessly expanding opening theory. Nevertheless, GMs
may tacitly "agree" to play only a subset of opening lines as
an attempt to limit the risk of coming across something that
they are not routinely familiar with. In fact, risk minimization
may be an important factor behind opening fads in GM play. I
freely admit that this is only "wild" speculation on my part! 8)


This is conservative versus aggresive play,
and Fischer used it against Spassy during
the world championship. Spassy assumed
that Fischer would always used the openings
he had used in the past, and then was surprised
(as was the chess world) when Fischer pulled
out this English Opening, something he had
never used before in tournament play:

http://www.chessgames.com/perl/chessgame?gid=1044367

Yes, _one_ interpretation is that Mr. Fischer broke a gentleman's
agreement of sorts in deviating from "generally expected" (accept-
able?) GM behavior... 8)


Hence, my argument that
a sizable part of a great chess
champion is pure, brute force
memorization.

If you stick around this forum for a while, you will come across
quite a bit of coverage regarding this. As a Chess_18 enthusiast,
I cannot get too much excited about such memorization feats. But,
the fact remains, FIDE chess at the highest levels seems to call
for heavy duty memorization...

Oh definitely, and i've heard that especially with
speed chess, ya gotta know your opening
inside out if you want to do well.

Yes, this seems very logical to me. An equally important, perhaps,
skill is to be able to handle opponents' deviations from the "book"
effectively under the severe time pressure involved... To this ef-
fect, some posters have suggested that a "proper book" should contain
the refutations of all sub-optimal deviations as well!

wrote:
David Richerby wrote:
As I said elsewhere, the problem with this idea is that computer
opening books are made by humans, not by computers. If you assume
that computers are unable to play the opening well, you're also
assuming that they're unable to generate good opening books.

I mean that the computer can number crunch way in advance (left on
overnight, etc.), and find the best opening books, looking ahead
many more moves than they normally would in regular play. They
could then store these opening moves as a database for each Fischer
random position, and thus have a time advantage during the openings.

Sure -- just buy a thousand computers and leave them chugging away for
a month, or a hundred computers for a year. Expensive but not
implausible. This would only buy you several extra ply of analysis
and the problem is in evaluating the existing positions so it's not
clear that this would lead to dramatically better play.

It's possible that I'm underestimating the effect of these few extra
ply, of course.


If you and I sit down on opposite sides of the chess board, you
might as well save some time and resign off the bat!

Aw, don't you want a draw? :-)

Naw, I'll pull a Fischer and go cut-throat!

I'LL KICK ANYONE'S ASS HERE!

Resigns.


Dave.

--
David Richerby Miniature Generic Newspaper (TM):
www.chiark.greenend.org.uk/~davidr/ it's like a daily broadsheet but it's
just like all the others and you can
hold in it your hand!

David Richerby wrote:
Sure -- just buy a thousand computers and leave them chugging away
for a month, or a hundred computers for a year. Expensive but not
implausible. This would only buy you several extra ply of analysis
and the problem is in evaluating the existing positions so it's not
clear that this would lead to dramatically better play.

What I forgot to mention is that, as far as I'm aware, nobody has
thought it worthwhile to take one of these computers and have it chug
away at the standard opening position for a month or so. This would
suggest that the approach is not seen as being very valuable.


Dave.

--
David Richerby Erotic Disposable Tool (TM): it's like
www.chiark.greenend.org.uk/~davidr/ a hammer but you never have to clean
it and it's genuinely erotic!

Major Cat wrote:
Yes, _one_ interpretation is that Mr. Fischer broke a gentleman's
agreement of sorts in deviating from "generally expected" (accept-
able?) GM behavior... 8)

Not really... Nobody goes into a big match without some opening
surprises prepared. Admittedly, it's usually a new idea on move
fifteen of the Ruy Lopez rather than playing a totally unexpected
opening but that's just a difference of scale.


Dave.

--
David Richerby Miniature Devil Chainsaw (TM):
www.chiark.greenend.org.uk/~davidr/ it's like a lethal weapon that's
possessed by Satan but you can hold
in it your hand!

wrote:
Spassy assumed that Fischer would always used the openings he had
used in the past, and then was surprised (as was the chess world)
when Fischer pulled out this English Opening, something he had never
used before in tournament play:

http://www.chessgames.com/perl/chessgame?gid=1044367

That link is to game 8 of the match. The surprise was in game 6 and
it wasn't really 1.c4 but 3.d4 that was the shock, transposing into a
queen's gambit.

Fischer quite often played Reti and KIA systems early on and
occasionally used 1.c4 throughout his career --- a well-known example
is his game against Polugayevsky at the 1970 Inter-Zonal. (There's the
story that Polugayevsky arrived a couple of minutes late for that game
and thought he'd gone to the wrong board because Fischer had gone for
a walk and 1.c4 had been played.) To the best of my knowledge,
though, Fischer had never transposed from 1.c4 into a queen's pawn
opening before the sixth game of the Spassky match.


Dave.

--
David Richerby Confusing Edible Chicken (TM): it's
www.chiark.greenend.org.uk/~davidr/ like a farm animal but you can eat it
and you can't understand it!

David Richerby wrote:

David Richerby wrote:
Sure -- just buy a thousand computers and leave them chugging away
for a month, or a hundred computers for a year. Expensive but not
implausible. This would only buy you several extra ply of analysis
and the problem is in evaluating the existing positions so it's not
clear that this would lead to dramatically better play.

What I forgot to mention is that, as far as I'm aware, nobody has
thought it worthwhile to take one of these computers and have it chug
away at the standard opening position for a month or so. This would
suggest that the approach is not seen as being very valuable.

Yes, but this may be due to the fact that most serious analysts
appear to believe that "orthodox" opening theory in its present
state is superior to almost anything that a powerful computer can
come up with on its very own...

As an aside, could a program specifically designed to evaluate
openings make some difference? If so, what would its salient
features be and how would it differ from existing all-purpose
programs?

David Richerby wrote:

Major Cat wrote:
Yes, _one_ interpretation is that Mr. Fischer broke a gentleman's
agreement of sorts in deviating from "generally expected" (accept-
able?) GM behavior... 8)

Not really... Nobody goes into a big match without some opening
surprises prepared. Admittedly, it's usually a new idea on move
fifteen of the Ruy Lopez rather than playing a totally unexpected
opening but that's just a difference of scale.

It is precisely this difference in scale that may be important
to speculate about here. I have a hunch that GMs tacitly agree
to minimize certain types of risk while fully embracing others...