Comments/Ratings for a Single Item

OK, I think we should agree to disagree about Capablanca/Carrera chess’ value as a variant, well except to point out it has, short of regional variants like Shogi and Chinese Chess, the most computer engine support of any variant; it’s one of the few variants where I can readily have, after a couple of weeks, some 30,000 games played with a given array to study its opening. One thing I wish Jeff Mallett added to Zillions was the ability to have the engine automatically play hundreds or thousands of games at a given engine strength/time control, so more variants could be better studied.

So, a question: What is your favorite variant right now? What do you like about the variant in question?

It's not an important issue being a more or less settled matter. The exact
RNBCeQKChBNR on 8x10 is 400 years old, given that differing arrays are
automatic. No one came to the defense of Centaur and Champion after
comments the first week of this month. C(RB) and C(RN) are of great
historical importance, no question. People now call OrthoChess 64 boring
and past its time. It's both taste and rhetoric to offer constructive
criticism, and this thread took a lot of thought. More importantly,
prioritizing is a beginning to make some progress for a change. That was
the purpose of NextChess threads. I think Schoolbook was a NextChess
nominee as a formality. I absolutely think it is commendable to run 10,000
games of one Carrera array. Doing so does not even attempt to address two
critical issues. One, proliferation of CVs to the point even computer
runs are impossible for as many as 1% of them. Two, in this case, that most do not rate the
awkward tandem RN-BN very high any longer and their being passe. Plain as day. Thanks for the
example of what can be done, the two conclusions for the particular
Schoolbook array. Certainly 8x10 is still the best try for those two after 4 centuries.
However, BN alone is probably better for example in both Modern of Maura
and Janus Chess of Schoendorf.

I started this thread, among a dozen other threads, over two years ago,
because I don't like to write articles for CVPage right now because of
their lack of direction. Anyway there are nine Demonstrations of Flawed
Centaur BN and Champion RN here. They warrant an article. The biggest flaw
is not even here yet thought sometimes mentioned. That is, they are so lame
as compounds they don't go paired as individuals but interlocked. You
can't have one without the other. Sort of like Schizophrenic Chess has
Left Schizzy and Right Schizzy.  One reasonable exception is Janus Chess,
trying paired BN not without interest. Modern designers avoid the Knight
compounds with sliders, because it ruins the regular Knight. Everybody
knows that. I lost what respect I had remaining for Seirawan when he and
some co-author began advocating their RN-BN drop on 8x8. Real original. To
each his own. When Capablanca's name begins to fade, they'll call the
things Seirawans, well-deservedly. If, incredibly unlikely, Centaur and
Champion turn out to be some next phase, too bad for the CVPage million
CVs, since 8x10 Carreras, and 10x10 by implication, have been around since
the end of the first full century of the fine small 8x8 Mad Queen, now
played out (if it wasn't already in Capa's day). Capablanca had nothing to do with either inception, did he? To
call these things Capablancas, in accepted practice, is lifting from prior
art.

I suppose it boils down to the strategical dimension. I, too, prefer
variants that are more strategical whereas the BN and RN pieces typically
enhance the tactical dimension. Capablanca Chess variants are not very
strategical but often result in brutal tactics in the middlegame.
Nevertheless, I have created a version which utilizes the 'relocation
method'. It implies a thorough kill of theory in Capablanca Chess. The
Capablanca piece array can in one move be rearranged by the players,
creating 144 different board positions. There is a strong Zillions
implementation of Capablanca Relocation Chess here. I don't know what
it's worth. I did it mostly to promote my relocation method.
http://hem.passagen.se/melki9/caparelocation.htm
/Mats

I prefer any of several dozen bifurcators to BN. There is a lot of beauty
on the chessboard down to Betza's Half-Duck and Berolina Pawn. I think I
should in fact differentiate the distinction H.G. draws. That 1-10, BN is
5.5, about average, among some CV-piece-type 500, and it is the RN, 9.5, that
draws down the average for the pair. Pritchard suggests RN to replace RB in
Intro 'ECV' but obviously on 8x8 with standard Pawns, BN is the one as good
as Queen. I wouldn't want BN on 8x10 or 10x10 to substitute for Queen. I
think the best Carrera-style outgrowth is Janus 8x10, dispensing with
Champion(RN). Obviously some few still like Centaur and Champion, but they
are average to most modern designers. [Sam will have noticed, this comment went up simultaneous as his.]

Betza didn't use these pieces together because his goal was to make a variant using the FIDE pieces that was as strong as the FIDE army. You can't have the RN and BN with the Queen if you want to do that. However, in CWDA, Betza added the RN to the “Remarkable Rookies” army and the BN to the “Colorbound Cobblers” army, so he obviously had no problem with the pieces, just with having too many of them on the board at once. Also, look at “Almost Chess”.

Speaking of Betza, I wonder about a Capablanca variant where the knights are different; one is a Betza Crab + Ferz; the other knight is Crab + Wazir. Instead of an Archbishop, we have a Crab + Wazir + Bishop; instead of a Marshall/Cardinal, we have a Crab + Ferz + Rook. Should be about as powerful as the Capablanca army, and allows 252,000 opening setups instead of “only” 126,000 opening setups.

The (BN) is actually one of the most wonderful pieces ever designed. The
exceptional synergy between the B and N move make it unexpectedly powerful,
and this power is wielded with exquisit grace when the piece is properly
used. It creates action and excitement on the board, wherever it goes. The
only name that does it right would be 'Dancer'.

The Mad Queen is a comparatively dull and Boorish invention compared to
(BN). The (RN) piece is not so hot. It is a bit clumsy. In Chess with FIDE
Pawns, open files tend to be scarce, and orthogonal movers are severely
hindered in the middle game, the phase of the game that counts. Diagonal
moves run the show. Anyone having watched sufficiently many high-quality
games involving these pieces will attest to that.

Quite possibly (RN) would be a wonderful piece in combination with
Berolina Pawns.

I have no idea why any variantist is remotely interested in using anymore
Carrera Centaur (BN) and Carrera Champion (RN). On a scale of 1-10 of the
major CV pieces with one at the top, RN and BN are ugly at about 8.0. I
think most designers would put them right below the middle about 5.5 or
6.0, not so low as I do. They ruin many a CV, and are about as interesting
and cumbersome as using mediaeval Alfil (2,2) leaper on a large board.
Dutch designer Freeling's Grand Chess, long referred to as great by
certain novices, gets no more mention here. Any bifurcator or ShortRange
Project piece or Betza ideal-and-practical-values compound lacking full range
tops RN and BN in beautiful play as a general rule. It would not be too hard nowadays to start to rank the best 500 CV pieces, and RN, BN and RBN have only historic interest. This thread describes
several specific flaws. Tutti-Frutti uses the three horrors together on the standard small board, Amazon, Centaur, Champion. I think Betza and Cohen are doing this more or less facetiously -- harbinger of Betza later post-2000, when he straddled important design and sarcasm.
http://www.chessvariants.org/dpieces.dir/tuttifr.html
In fact, you would be hard pressed to find two more Betza uses of RN and BN. He grew out of them. By the days of Chess Variant Page Betza stopped using them -- relatively speaking.

King and Queen are different, and most variants have only one of each. Does
that mean you also consider the Mad Queen game flawed, and advocate use of
King + Commoner?

Here second half of 17.January.2008 Comment> ''Let us end the misery
putting them down for the last time. Euthenize them, if it were
figuratively possible, on the supposition that an idea has life. Creative
Pietro Carrera's curiositites, Centaur(BN) and Champion(BN), original for
their time, contemporaneous with Shakespeare and Pocahontas, came on the
heels of 'defeat' of the Spanish Armada in 1588. Foredoomed in employing
overwrought, ineffectual chess-compounds,
the stream of copycats for 400 years, one and all, proved destructive of
critical skills and subtle play inherent in legendary stand-alone utility
Knight. R.I.P., Centaur. Requiescat in Pace, Champion.''

In support of DEMONSTRATION IX> There is no obligation to use Ultima
Chameleon, V. R. Parton Swapper, or Rococo Cannon Pawn with some separate
unit systematically. Each is satisfactory stand-alone chess piece for
implementations. Fourriere uses Cannon/Canon effectively as one piece in
Jacks & Witches. That cannot be done with RN-BN, or you get clumsy
over-strong, even more grotesque game-destroying Amazon(B+N+R). Altair two Grand Bishops, Sissa two of each, Centennial two Spearmen, Quintessential two Quintessences, Grande Acedrex two Gryphons -- each pulls itself along by their own bootstraps, unlike the derivational Centaur-Champion.
Significant others these RN-Marshall and BN-Cardinal, their peculiar symbiosis and state
of affairs until death do they part. You can't have one without the other.

Definitely okay but unaesthetic Centaur and Champion(RN) had their eulogy
announced 17.January.2008 at Grotesque Chess, where respectfully called 
venerable and creative for Carrera's time, setting up a 400-year shadow
Chess still widely regarded. However, there can be no respect for RN and 
BN overuse claiming new CVs, as Muller points out. This old thread
''Fatally Flawed'' has each of 10 Demos different, all so far using
board positions. DEMONSTRATION IX: Major lack of aesthetics in
BN(Cardinal) + RN(Marshall) is being two different pieces. Out of the
ordinary Janus Chess uses two BN on 8x10.  Where are two RN on 8x10?
Practically nowhere, you don't do that, because (notwithstanding
exceptions like Janus) the pieces cannot stand on their own hindlegs. Each
dependently needs the other somehow presumptively to balance right. What other different
pieces more or less are always mated up that way in designs? Very few. (We
can think of some later you forget.) For example, Berolina Pawn is not
matched 50-50 with Orthodox Pawns (except Overby's). If Templar is any
good piece, introduce two Templars (Templar Chess), and at least do not
require its accompanying inverse, contrapositive, companion, whatever
always. Dreadful Omega Chess even implements two ''Champions'' (WAD) and
two Wizards (Ferz+Camel), based on their strengths, such as they are. If
Winther's Mastodon Chess new-old piece up to two squares is great, it
does not need supporting cast mandatorily.

So (BN) is flawed because it cannot enforce mate on 3x3 now? What kind of
reasoning is that? Falcon cannot do that either. In fact, it cannot even
move on a 3x3 board. Does that mean it is even more fatally flawed?

I think it is a mistake to consider using a different array as creating a
new variant. I usually refer to such things as sub-variants. In the
WinBoard GUI, 10x8 games like Carrera, Bird, Embassy are all played as
'variant capablanca'. If not actually playing Capablanca, the user will
have to provide the opening setup (as FEN).

I guess (BN) and (RN) are so popular because their play is appeling in
practice, and they blend in well with the usual crowd of FIDE pieces.

The point you are trying to make escapes me. What the heck do you mean by
'with the latter determinin who moves first'? K + (BN) vs K is a totally
won end-game even on 16x16. With white to move, there are no draws. With
black to move, there are of course always draws when the blck King can
capture an undefended (BN), and it dos not care much if it was a (BN),
(RN) or Q it was capturing...

Some of your other aguments seem to be directed against (BN) and (RN)
occurring only once in the Capablanca (and related) setup. It says nothing
about the pieces per se. You could have said exactly the same thing about
Rook and Bishop being fatally flawed, if you have one of each (like in
Shogi). Or the Knight and the Falcon.

It just doesn't make any sense. So the Cardinal and Marshall are mixed
slider-leaper compounds. So what? You seem to judge pieces by the abstract
symmetries underlying their design. But the rest of the World judges them
for the beauty and marvelous complexity they display in action, when
participating in games. You seem to live on another planet...

Fatally flawed. This thread Joe Joyce and I had fun with for a month 2007.
I give five demonstrations that Marshall and Cardinal are fatally flawed.
Joyce concentrates on short-range substitutes like his (20.8.07)
''combination Dabbabah-Wazir.''

Over here 20.August.07 JJoyce says, 'Think of the genius who first
invented the Knight move; and then wonder what his/her friends said the
first time that Crooked jump move was used on them. How popular was that
piece in the beginning?'  Well, In the Beginning, were the Knight, Rook
and King -- to be brief and reductionist, not entirely inaccurate
metaphorically either. So, those semi-ideal 'friends' would have been
well used to the Knight, not the other way around. PAronson may know more
about appearance of Knight, or Rook, or King, in early board games Before
the Common Era. Pertinently, Shatranj, and Chaturanga before that from year 600, have only those three movers perfectly the same of the six-piece-type method. That generality was the main objection to couple of JJoyce offhand remarks. [Royal Falcons etc. being studied]

George, you just might be looking for shorter-range pieces!  :-)  
On a 2D square-units board, there are only 2 kinds of 'shortest
distance' infinite sliders, the orthogonal and diagonal, rook and bishop,
with the queen of course being both. [For a different take, Graeme has got
some interesting things going on a triangular board.] This doesn't leave
many options for variety. What you do have are 2 pieces that travel the
maximum linear board distance in the least number of squares possible, a
'straight' line of n-1 squares with n = board length. If you're going
to cover the same [maximum linear] board distance with a new slider, then
the path that slider takes must be longer and more complex that that of
the R or B. This makes the game harder, and your critique - who in the
real world will play it? - applies. So sliders may have to either be of
shorter range or dropped entirely. [One very simple longer-path slider is
the 'hook-rook', a bent rook that can cover the entire board potentially
in one move. But you don't want that kind of power, I suspect.]
Obviously, we both lean toward shorter range pieces, you with the Falcon
and me with a proliferation of pseudo-shatranj pieces. But I'm curious
just where you're taking this longrange thing, and I'd like to examine
the potentialities of the DW as a longrange piece - in another post.

DEMONSTRATION VIII: (Mate in Four) Black has just played R j5-j4 Check.
Black's only other piece is Rook at b8. White has full complement of ten
8  __R__ __ __ __ __ __ __ __      pieces/ten pawns. The Two Rooks Alone 
7  __ __ __ __ __K__ __ __ __      checkmate in at most four moves by 
6  __ __ __p__ __p__ __b__ __      Rook at j4 in turn capturing any
5 p__ __p__ __p__ __p__p__p__      interposer across the entire rank 4 
4 k__ __ __ __ __ __ __ __ __R     to White King.  At most, in turn,
3 r__ __q__ __ __ __ __ __ __      White Champion(BN), Marshall(RN), 
2 p__ __p__ __ __m__ __ __ __c     Queen and Pawn-c2 interpose only to
1  __b__ __n__ __ __ __n__ __r     be captured. So what?  Well, it is a
  a  b  c  d  e  f  g  h  i  j     rather straightforward unpeculiar
position(only omitting any other logical Black pieces as irrelevant).  Get a feel for how there is simply no subtle move-order for White to consider: Queen first only makes it mate in one or two instead. Whether Marshall goes -e4, -e5, or -e6 same outcome; ponderous, is that not so?  Just play abstractly the doomed line of bowling pins across Rank 4 for the awkward constraint RN/BN pair tend to impose. As infinite in variety as tired 8x8 itself they may be, yet Cardinal/Marshall 8x10 positions typically present such rather uninteresting interactions.  Cannot future composers find more promising piece material for their skills to work? Are there not some other optimizations than top-heavy ancient Carrera-Centaur and Carrera-Champion?

Subtlety. Subtle T, Sublet, Letups, Upset, Setup:  
        DEMONSTRATION V:                 Think of Centaur(BN) and
  n___n___n___n___ ___ ___n___n___n___n  Champion(RN)in their Knight mode
  P___P___P___P___P___P___P___P___P___P  Then from most of the starting
  x___N___C___x___x___x___x___M___N___x  arrays with these pieces, there
  a   b   c   d   e   f   g   h   i   j  are eight possible 'Knight'
   openings: N-a3 or -b3 or -c3 or -d3 or -g3 or -h3 or -i3 or -j3.
   DEMONSTRATION VI: Think of    p___ ___ ___ ___ ___ ___ ___ ___M___ 
Champion and Queen in their      k___p___ ___ ___ ___ ___ ___ ___R___
Rook mode. All four Ranks 1,      ___ ___p___ ___ ___ ___ ___ ___R___
2,3,4 are fully controlled in     ___ ___ ___ ___ ___ ___ ___ ___Q___
this (incomplete) endgame by     a   b   c   d   e   f   g   h   i   j
four 'Rooks' except squares a2 and b2.  Subtle.         
   ___ ___ ___ ___C___o___o       DEMONSTRATION VII
   ___ ___ ___o___B___B___o  o = control by diagonal w/o capture
   ___ ___o___o___o___Q___   Think of Centaur(BN) and Queen(RB) in
   ___o___o___o___o___ ___   their Bishop mode against White's King,
  o___o___r___o___ ___ ___   Queen, Rook, Rook. Four 'Bishops',
  q___r___o___ ___ ___ ___   four long diagonals fully under control.
  k___o___ ___ ___ ___ ___   Overkill.

Given that the Mad Queen as such is dead. Like many-headed Hydra, a new
Carrera derivative most every decade since year 1617. Is Carrera's worthy
companion piece? A Hera? Or only weak sister, lady-in-waiting? To extent
JJoyce's Comment about 'Rhino' relates to this M/C thread, it suggests
Mao and Moa compounded with Rook and Bishop respectively, instead of
Knight with either. No secret then where we may well he headed: in place
of awkward combination pieces NR and NB, instead overlapping sliders, like Queen itself, making true compounds. We accumulate the evidence in DEMOS,
no one necessarily more or less convincing in itself, toward a preponderance of evidence. DEMONSTRATION III: On 3x3 board and all larger
  3 K__ __c  (rectangular) boards from a legal position, Centaur(BN) and
  2  __ __   King cannot checkmate lone King, with latter choosing who
  1  __ __k  moves first. Champion(RN) can: like one of some seven  
             deadly sins, the fatal flaw asymmetry.
DEMONSTRATION IV(symmetry): Ignoring the other back-rankers, consider the
Pawns protected by Queen(RB), Champion, Centaur. There is no way to avoid
  a___a___P___a___aq__qm__q___m___P___m  a = pawn protected by Centaur,C
  x___x___C___x___x___Q __x___M___x___x  m = pawn protected by Champion
      b   c   d   e   f   g   h   i
at least two unsymmetrical unprotected Pawns, without positioning Champion and Centaur themselves unsymmetrically. That Pawns may subsequently be protected by other piece placement is irrelevant. This asymmetry here again being sign of certain imbalance, or disequilibrium, in power distribution among the pieces.

Waiting on DEMO III against Carrera stock Champion(R+N) and Centaur(B+N),
this thread wants to show they are incompletely-examined implementations to those (2,3) plus (1,2)(1,3)(1,4)...squares. JJoyce's generic DW,
Dabbabah-Wazir, he describes in number of different implementations.
Broadly, they are all 'sequential pieces' whether there are two or three
legs, whether repeat Dabbabah-Dabbabah is allowed, and so on. If memory
serves, Antoine Fourriere deleted a Comment in 2004 saying he prefers a
piece moving to Falcon (2,4 and 3,4) squares by way of two legs: Knight
leap then mandatory one-step outwardly to those (Camel) or (Zebra) arrival squares.  Differently, JJoyce's legs up to three for this DW are optional.
AF's description would also include recent Sissa in the same group of 'mandatory sequential', but 13th C. Gryphon would fall in JJ's 'optional'. Sequential piece, Multi-path piece, Leaper, Slider(one-path) are four fairly distinct categories. No full taxonomy attempted like RBetza or DHowe['A Taxonomy']: ''venture too deep into the jungle of classification''  we used in year 2000 FC article. Now for convenience, we call stock RN and BN 'combination pieces' because made of Leaper+Slider, functional opposites, but 'pseudo-compound' is descriptive as well. Any Leaper is a multi-path form with an infinity of paths, therefore no path at all in automatic shift to an arrival square(s).  Any Slider one-path has the potential for other pathways that could be added to piece-move definition. Therefore, multi-path is the linking category or organizing principle. Even the above sequential pieces are built according to definition of particular multi-pathers' units.

Thanks for excellent input and we agree 100% with JJoyce's except one
mis-read: instead, 'pseudo-compound' is nothing but innocent undiplomatic, vernacular synonym for combination piece. Yes, very good interesting piece. Chatham's replacements, in our system, are to be developed in Demos III through XX against Carrera's stock.

George, you've got some things going here that, naturally enough, I don't
fully agree with. Let me start by taking issue with the following quote:
'Disparagingly, 'pseudo-compound' fits also because of improbability
that movements combining powers at opposite extremes, namely leaper and
one-path slider, could be very effective within one piece. Hence their
unpopularity.'

Contingency, historic accident, is a poor base on which to build an
argument of inevitability, and buttressing it with an unsupportable slight
does not help convince the skeptical reader. That a piece is worthless, or
at least worth little, if it is not in current use by millions of people,
is at least historically demonstrable as false. New pieces do arise and
are not always popular at first. Think of the genius who first invented
the knight move; and then wonder just what his/her friends thought and
said the first time that crooked jump move was used on them. How popular
was that piece in the beginning?

The piece I wish to consider is the combination dabbabah-wazir [DW]. Let
me hasten to add that I don't expect this piece to take the place of the
knight in 100 years, or anything close. But I do wish to examine some of
the potential for this apparently unprepossessing piece, very specifically
because it combines a leaping and a non-leaping component. The DW has as
its basic moves a 1-square slide or a 2-square jump. If allowed to use
both halves of its move, it leaps and steps 1 to 3 squares, an inclusive
compound piece [DHowe 'A Taxonomy']. If the piece instead repeats its
basic move once or twice, it becomes a limited rider and may move up to 4
or 6 squares. The 3 step DW rider is an interesting medium-range piece
that I have not seen examined in any games. I don't have a real
combination leaper and multi-square slider there, but the piece may leap
once and step/slide 1 square twice in its move, or even slide 3 squares as
its move. Will this piece do as a counterexample, George?