Comments/Ratings for a Single Item

With all due respect ...

I cannot yet find a *.zrf which plays 'Capablanca Random Chess' or
'FullChess' available upon major chess variant outlets such as Zillions
Of Games or The Chess Variant Pages.  [Incidentally, a *.zrf which plays
'Fischer Random Chess' is available to use as a pattern.]

Short of owning a physical board game enabling its play (which in fact
can
be purchased as a 'Gothic Chess' set), gameplayers who own the Zillions
program must have a means of conveniently experiencing this 'game which
includes many permutations'.

I would strongly recommend, however, that EVERY single,
comparatively-symmetrical permutation be thoroughly tested and analyzed
to
eliminate ALL flawed formations (including irrefutable, quick checkmates,
unstable defenses or  decisive advantages to the player with the first
move of the game [white]) as well as eliminate ALL undesirable formations
(complicated, game-specific preferences agreed to by a consensus of
knowledgeable players).

Only after all of this intensive groundwork is accomplished should the
game be released since it would be terribly inconsiderate to subject many
players to ANY worthless and undesirable permutations (which I suspect
comprise the vast majority of those theoretically available).

I wish you all of the best in your worthwhile yet formidable endeavor.

I disagree with Robert Fischer.  It is not possible to test EVERY possible
permutation, since there are tens of thousands of them.  And I think it is
totally unnecessary.  Fischer Random Chess has 960 different set-ups, and
probably, by now, all have been played at least once, but *NO* organized
effort has been made to test each and every one by high-level Chess
players for playability.  In fact, FRC doesn't even try to avoid
unprotected pawns.  Considering the number of FRC enthusiasts, I doubt
it's necessary.  Besides, as Fergus recently pointed out in a different
thread, the most common opening move in Chess is pawn to king-4, which
creates an unprotected pawn!  As long as white and black have identical
setups (mirror-symmetry, not rotational-symmertry) the game appears to be
fair.  I see no reason why Capablanca Random Chess would not be fair,
especially since Dr. Scharnagl has excluded all setups with unprotected
pawns.

Regarding Capablanca Random Chess:
Good work, Reinhard!  You have taken the concept of crossing Capablanca
Chess with Fischer randomizations and done a fantastic job of identifying
the issues that need to be addressed, such as castling and proper
notation.  I look forward to adding proper support for this game to
ChessV.

You asked how a game becomes 'recognized'.  I suspect you didn't mean
to use that word.  There are only about 30 variants that are listed as
Recognized Variants on this site, and it takes a while to become a
Recognized Variant.  What I suspect you want is to have a web page created
on this site for the game.  There are well over a thousand such games, and
it is not hard for yours to be added.  Technically, all that is required
is that you submit an e-mail to the editors.  They are very busy, though,
and they can get it up much faster if you have it already formatted into
their HTML template.  If you would like, I would be happy to take the
description you submitted, and format it properly, and submit it for you. 
I can also create a Game Courier preset for you if you like.

P.S.  Is your book on Fischer Random Chess available in English?

Sincerely,
Greg Strong

Greg, let me add some statements to the discussion here: it is essential,
like in FRC/Chess960, that the drawing of the to be played starting array
is done immediately before the game will begin. Thus possible weaknesses,
also because equally distributed on both sides, could not be used to
prepare a strategy based on unbalances. Avoiding uncovered pawns only will
exclude obviously weak points and have the players focus on a good
positional and creative play.

The FEN notation is subject to be improved. There seem to be some
incompatibilities using '0' for ten empty squares compared to a FFEN
approach. Thus I am thinking to not encode any empty squares finishing a
line (for 10x8 boards only). It seems as if that would cause less
problems.

There is already a homepage of CRC / FRC and more, see
http://www.chessbox.de/Compu/schachcrc_e.html and a lot pages around.

My book on Fischer Random Chess is available in German language only. And
because it is not a bestseller (9 copies have been sold in quarter 3 2004)
I doubt whether a translation into English would pay.

Reinhard Scharnagl

'It is not possible to test EVERY possible permutation since there are
tens of thousands of them.'

I have not made even a wide estimate of the number of select permutations
(with undesirable starting arrays excluded via game-specific criteria)
which would require some study to expediently determine whether or not
they are probably of decent quality.  Perhaps, Dr. Scharnagl has
calculated a firm estimate.  Even if it is not easily possible for one
person to playtest and analyze, for wild example, 10,000 select
permutations in less than 2.7 years (arbitrarily calculated at a rate of
10 per day), this is totally beside the point.

The point is that the quality of EVERY chess variant given to the public
which can legally be played must be protected.  'Bad game pollution' is
already a serious problem in the chess variant world.  If you wish to, in
effect, give the public 10,000 unique games to play, then it is your
responsibility to adequately playtest and analyze every damned one of
them.  Otherwise, please do NOT do it!  To be sure, noone has the right to
carelessly or lazily give the public a package of, in effect, 10,000 unique
games ... knowing that the vast majority of them are defective.  Hence, a
program designed to play 'Capablanca Random Chess' must NOT be capable
of loading ANY flawed starting arrays for people to waste time and
frustrate themselves playing.

I realize this entails a vast amount of work but if someone wishes to be
rightfully known as the inventor of, in effect, a couple-few thousand
unique, 'high-quality' (rather than 'low-quality') games, then this is
what they must do to earn it.

'And I think it is totally unnecessary.  Fischer Random Chess has 960
different set-ups, and probably, by now, all have been played at least
once, but *NO* organized effort has been made to test each and every one
by high-level Chess players for playability.  In fact, FRC doesn't even
try to avoid unprotected pawns.  Considering the number of FRC
enthusiasts, I doubt it's necessary.'

Popularity is no guarantee or proof of quality.  Do you also doubt that
quality is necessary?

Notwithstanding, what 'Bobby Fischer' and others have been disgracefully
derelict in doing should not be imitated.

'As long as white and black have identical setups (mirror-symmetry, not
rotational-symmertry) the game appears to be fair.  I see no reason why
Capablanca Random Chess would not be fair, especially since Dr. Scharnagl
has excluded all setups with unprotected pawns.'

Unavoidably, fairness must be proven individually for every unique game
studied as a permutation of the set.

I am very pleased to find that Dr. Scharnagl has already established rules
for determining a select list of permutations.  Perhaps, he is already far
more aware than I of what he needs to do in the future to achieve success
for his project.

George, you have completely mischaracterized the analogy I made, and your
belief that a 'devil-may-care' and 'no holds barred' attitude rules
here is completely mistaken.

The James Bond analogy has nothing to do with a loose attitude toward
quality. This analogy was given in the context of two other analogies, a
Romeo analogy and a Don Juan analogy. Each man was characterized by a
different attitude toward women, which translated into a different
attitude toward Chess variants. A Romeo sticks to one variant; a Don Juan
seeks after the perfect variant; and a Bond enjoys different variants
without trying to seek after the perfect variant. But this does not
translate into unconcern over quality. James Bond may sleep with many
women, but they are always beautiful women. I enjoy many variants, but I
prefer good variants to bad. The main thing that characterized the James
Bond attitude is the ability to appreciate quality in a diversity of
forms.

Also, when I create my own variants, I pay lots of attention to quality,
and I encourage everyone else to do the same. I normally playtest my games
against Zillions and make changes based on playtesting. And when I can't
make a game good, I avoid releasing it.

'I think what Robert and George are arguing for is: quantity with
quality.
Whereas, Fergus, Roberto and Michael are arguing for is: quality with
quantity.'

Yes, this summarizes the main argument elegantly.

When dealing with reasonably small numbers, it remains possible to
overcome the inverse relation between quantity and quality thru much hard
work (playtesting, analysis, research, theory, invention, etc).  In fact,
some [but not all] so-called prolific game inventors (usually less than
50
games) have done admirably well at maintaining quality throughout their
body of work.

Unfortunately, when dealing with astronomical, combinatoric values such
as
8!-10! or even much greater, the people who have undertaken this work
(even
though they comprehend the mathematics behind their craft) are, by their
own realistic admission, helpless to adequately playtest and analyze all
of the individual games they nonetheless encourage people to play.

At this extreme, the inverse relation between quantity and quality
becomes
inescapable and uncontrollable.  Furthermore, quality will exist at its
lowest possible average within the greatest permutative set (quantity).
A randomization of quality itself actually occurs with a maximization of
quantity.  Contrary to popular belief, a randomization of quality does
NOT
approach an average.  Instead, it approaches the lowest possible value
since the definable nature of quality involves order and structure. 
Chaos
and statistical randomization are the opposite process- inverse or
destructive to isolated existences of order and structure.  Finally, it
is
prohibitively labor-intensive to filter-out the works which are defective.

So, they must remain as hazards for gameplayers.

My final conclusion- I think it is wise to avoid this unmanageable,
permutative extreme of game invention entirely and forever.

'I think any conscientious game developer is concerned about quality.'

Yes but to varying extents.

On one's list of priorities, if quality is not #1.  Say, for example,
quality is considered important but ranks #2 with quantity ranked as #1
... in practice, quality may suffer greatly.

'I'm quite certain that he is not.'

You are correct.
I have even criticized Bobby Fischer in one message dated 2004-11-29.

I think one way we may be able to test the quality of variants is with
computer analysis.  So, if we have a shuffle variant with, say, 100,000
possible permutations, we can have the computer play itself all of those
permutations, In order to find one that is playable.  

Of course, computers play Chess differently than humans, so a computer's
idea of playability may be different than a human's.  And, computers are
generally programmed to find the best move in a given position, so
translating that to playest a game may be tricky.  

Greg Strong is much better qualified to say whether this is feasable or
not.  :-)

- Sam

It was classic, is vintage, popular, aclaimed, and various so on and so forth. As the only surving form of shogi, and as it has international tornament like support, i propose that it be in....