Comments/Ratings for a Single Item

Here's a (hopefully complete) list of armies:

Fabulous FIDEs
Colorbound Clobberers
Colorbound Clobberers II (alternate setup)
Remarkable Rookies
Nutty Knights
Forward FIDEs
Meticulous Mashers
All-Around Allstars
Amazon Army
Amazon Army II (Crabs replace Knights, Amazon replaces Queen)
Avian Air Force
Spacious Cannoneers
Amontillado (7 different knights, 2 queens for 14 total armies)
DemiRifle
Cylindrical Cinders
Colorbound Clobberers with Doublemove F instead of FAD
Colorbound Clobberers II with Doublemove F instead of FAD
Fighting Fizzies
Pizza Kings
Seeping Switchers

(I've got all of the above implemented in a ZRF that just needs some
polishing before I release it.)

Jupiter
Mannis Manglers
Nattering Nabobs of Negativity
Fabulous FIDEs with Iron Ferz instead of Queen
Fabulous FIDEs with Iron Crab instead of Queen

An 'Excellent' to the editor!

Several excellent people have also given excellent ratings for my game, for
which I thank.

It is common for the neophyte chess variant author to invent his first game
and tout it as the inevitable replacement for Chess. We all laugh at
this.

It is uncommon for somebody who has authored thousands of highly-regarded
chess variants to refer to one of his inventions as the most likely
evolutionary future of the game of Chess. I hope we all take this
seriously.

I do not expect that CwDA will become widely played, much less overtake
FIDE Chess, within my lifetime; nor do I expect that when it does the same
primitive armies that I designed will be used. However!

However, it was 1976 when I first conceived of the game, and 1996 when I
composed the first succcessful army (Colorbound Clobberers). Twenty years.
My first attempts were so bad; and I realized that in order to creat this
game I needed to explore the problem of the values of chess pieces. And so
I did.

Twenty years. A large part of one's life. Don't imagine that I thought
about the problem every day of every year, no, that's not how it went at
all! I worked on it, and I gave up in bafflement, and I came back to it
after a few years of not thinking about it, and then I gave up and came
back and tried again and gave up and came back and tried again and so on.
Not so much brilliant as really stuborn.

Remember that I am a genuinely certified master of FIDE Chess: I know and
love the openings, endgames, midgames. Chess with Different Armies has
satisfied my expectations of what Chess should be -- it has openings,
endgames, midgames, all with the general feel of real serious FIDE Chess,
but of course it's different. Someday, the Grandmasters will begin to play
my game, and because they are so strong they will find imbalances in the
particulat armies I designed -- and I don't care, because once they start,
they're hooked. Meanwhile, nobody can design any chess variant without at
least thinking about different armies! I am pleased to see this, because I
had expcted that my mind's greatest invention would not be recognized so
soon; and yet I always hope for more. Chess with Different Armies (together
with the essential work on piece values) is, I think, a really
revolutionary idea even though my own work on these subjects is so
hopelessly bad (I look good now, but when real mathematicians take a run at
the val
use and real Grandmasters start to evaluate my armies, watch out!) What an
accomplishemnt, and did you know that with that accomplishment I only need
a buck fifty to ride the subway?

Chess with Different Armies is certainly a very enjoyable set of games. I
particularly liked playing the Remarkable Rookies, perhaps because of their
mutually supportive and jumping capabilities. On the other hand, I had a
much more difficult time with the Colorbound Clobberers. Before I knew it,
I had trapped myself in an off-balance position. 

The overall idea of CWDA is very clever. The idea of balanced, yet
different, armies should see more use in Chess variants development. But,
as remarked by Ralph, this is not so simple and takes quite a bit of work.
On the other hand, it has endless possibilities with a simple theme. One
thought, would it be possible for players to 'assemble' an army from
'equivalent' sets?

Something else, there is a playful character to the armies, which is a nice
touch.

> This really does sound like a great game.

It is. At least one pair of armies has been tested with players who were
uscf master strength.

How about the Separate Realms Amazon- m(DDAAfbN)cQN - on the Queen squares,
to complete the theme? 

(I couldn't resist...)

I should think before I post--a fully thematic army is easy:

For Knights and Bishops, use their Separate Realms counterparts, for Rooks
use SR Queens and use a SR Amazon as the Queen.  This should be balanced
against the CWDA armies with top-heavy material (weak minor pieces, strong
major pieces.)

I've been playing around with variations on the Separate Realms pieces for
use in CWDA. The last army I proposed is much too strong--in fact the FIDE
team get slaughtered regularly.

The regular SR army is too weak -- every piece is mesurably weaker than
its FIDE counterpart.  The FIDE army won 9 out of 10 in Zillions vs itself
testing (the tenth game involved a huge blunder). Yet the SR team often
holds out for 60-70 moves even though I would guess its value at 3/4 of
the FIDE army.

What I must be seeing is the effect of the SR army's rapid
development--the SR team jumps the pawn line and FIDE is fearful of
devoloping its Bishops and Knights, least they be exchanged for their
weaker counterparts.  (A stunning confimation of Ralph's 'levelling
effect' -- the SR pieces are strong because of their weakness.)  This also
shows the value of leaping moves in permitting rapid deveopment (and
'undevelopement': the SR pieces are good at jumping back behind the pawn
line when needed.)

> A stunning confimation of Ralph's 'levelling effect'

What? You didn;t believe me?

I would guess that you are right in thinking that the rapid development
causes the army to get a great initial advantage which compensates for the
weakness of its individual pieces.

The Remarkable Rookies have perhaps the highest absolute value and the
slowest development (until you learn the trick of using the HFD to harass
the foe and give the WD and the R4 time to get into the game). For a
contrast in extremes, you shoul try the Rookies against your army!

Remember also that the results you get with a computer are not the same
results you would get if the game was played between two human chess
masters.

Ralph, I did indeed believe you.  Your work on the value of chess pieces is
the best in the game's history.  What surprised me is how strong the
synergy of rapid developement and leveling could be.

Zillions tries too hard to avoid exchanging FIDE minor pieces for their SR
equivalents--the material difference is only about a pawn and many times
the sacrifice is worth making for better developement.
 
Zillions is a flawed oracle, but a good strating point--if a game is
hopelessly unbalanced on Zillions, it is more likely to be worse than
better among masters.

Another experiment I've tried on Zillions:  The Separate Realms army vs
the Reverse Separate Realms army (Rook is mRcDD, etc.) Zillions is
notorious for undervaling capture and rates the reverse SR pieces as
nearly as strong as their FIDE counterparts.  Consequently it will glady
trada a SR Rook for a RSR minor piece.  Yet the game is no contest--the
reverse SR army is slow to develop and the pieces don't defend each other
well--its over in 18-24 moves.

Just as your work suggests, a divergent piece with more capturing power is
more valuable than a divergent piece with more non-capturing movement. In
any army I'll take mDDcR over mRcDD.  The above experiment suggested one
possible exception, the reverse SR King (FmW) seemed stronger than the SR
King (FcW).  The King's ability to avoid capture is more important that
its ability to capture.

I tried Ralph's suggestion and pitted the Separate Realms army against the
Remarkable Rookies--the games lasted about the same length as SR vs
Fabulous FIDEs, but had a different flavor--SR stayed competitive longer,
but once the slaughter started, it was over quicker. Makes perfect sense
as the Rookies are stronger the the FIDEs in absolute value but slower to
develop.

In another experiment, substituting the SR Rook (mDDcR) for the R4 seems
to strengthen the Rookies a fair amount (which would unbalance the Rookies
for CWDA). It can be devoloped to a square where it attacks the center in
two moves without moving a pawn; also it defends the first rank virtually
as well as a full FIDE Rook--the limitation on the squres it can occupy is
compensated by its ability to leap over obstacles.

The ancient different armies from 1977 or so are as balanced as they could
be with the knowledge and technology available at the time. It was the
failures of these armies that caused me to go into a shell for twenty
years, and come out only when I had a workable (though sketchy and
primitive) theory of piece values.

The Clobberers and the Rookies are amazingly well balanced despite the
fact that they are extremely extreme. Why did I start with such
difficult-to-balance armies? In retrospect, it seems foolish; but I do
love these two armies.

Without a theory of values, the only way to balance an army would be by
trial and error, with thousands of hours of playtesting per army. 

I do not have Zillions. I will not buy it, and I will not accept it as a
gift from anyone but its makers. If SUSE can send me their linux for free,
zillions can send me their product. Therefore I cannot comment on these
games because I cannot read them.

Actually, .zsg files are plain text and are not covered by licenses, any
more than the data files created by a word processor.  The text is mainly
the move list in full algebraic notation with a small amount of easily
ignored bookeping data.

I would urge anyone who wants to examine these saved games to download the
files even if you don't currently own Zillions--you will be able to read
them.

Looking back, I noticed that I had not rated this game, so I now correct
this oversight with an 'Excellent' rating for my personal favorite
variant.

I hope that Ralph re-emerges soon, because I am concerned that his
inventions might become overlooked without his continued input.  But I
will continue to do what I can to promote CWDA, though, such as voting for
it's inclusion in Game Courier tournaments, and providing the best
possible CWDA support to ChessV, for analysis of different army match-ups.
 Sometime in the not-too-distant future I will provide a great deal of
information here on what I have learned from computer analysis of the
major CWDA armies.

Regarding the Pawn promotion rule:  I would recommend a change to this
rule.  The current rule says a pawn may promote to any piece in either
army at the start of the game.  Here's the problem:  What about the
match-up of Nutty Knights vs. Nutty Knights?  Since no piece in that
entire army may move backward faster than one square at a time, even if a
pawn promotes to a (very powerful) Colonel, it still probably can't move
back into the frey quickly enough, seriously decreasing the value of pawn
promotion.  I would suggest the alternate rule:  A pawn may promote to any
piece (other than Pawn or King) in the player's army at the start of the
game, or in the standard Orthodox Chess army (Fabulous FIDEs).  This
always provides the option of promotion to Queen.

I always thought that Ralph preferred that both sides not use the same
army, which would make the NN vs NN case moot.  But even if you do allow
duplicate armies, promotion to Colonel might make for a slower end-game,
but not necessarily an undecisive one.

See Ralph Betza's 1996 Index to Articles about Pieces including the ones used in CWDA and others. Incidentally, Dai Shogi fans will find short articles on the Phoenix (under the name WA or Waffle) and the Kylin (under the name FD).