Comments by CharlesGilman

Well it seems to have worked - even though the preview always shows the user ID.
 	I am not convinced that this kind of variant is as bad for forward-only pieces as you make out. Thre is still a race to get pieces to the far end and promotion, and players will have to choose between pushing them to promotion and holding the stronger ones back to ambush enemies as they cross their paths. The Bishop's quickest route to promotion is on the enemy files, so deployment of suitable pieces to prevent this could be advantageous. A more unusual feature is the relationship between opposing part-symmetric pieces. These are now placed to best attack each other by sidling up, and will be able to make non-mutual threats.

It is true that many variants have produced one-off pieces beginning with O, but none were quite what I was looking for. Orphan might have worked, but at the time of writing it was off my radar - I only resorted to Joker out of desperation for a J. On the other hand Orphan has a strong resonance with Friend, and I was already using F for Foxhound as part of a larger family of pieces. Perhaps I could have had have an Orphan without a Friend based on the precedent of an Infanta without an Inquisitor. Hm, perhaps I'll add a subvariant with that, called Missing X Chess. It's probably not worth its own page.
	Having one of each piece would have violated my idea for full use of 4 distinguishable FIDE sets on 4 FIDE boards.

This is an interesting theory, but details of how the pieces were chosen belie it. Certainly part of the motivation was an exercise in choosing pieces subject to a not uncommon restriction, but I would hope the game more playable than one where each letter's piece were indeed chosen randomly.
	The criteria in the introduction rule out most combinations fitting the broader Duke criteria. Using 'as many FIDE piece types as practicable' restricts B, K, P, Q, and R to the FIDE pieces, while 'grouping piece types as much as possible with relatively few one-offs' also rules out vast numbers of combinations. Insisting on 'pieces representable by 4 distinguishable FIDE sets on 4 FIDE boards' rules out too many weak pieces, as pieces of which there are only 1 or 2 should be able to hold their own on a large board. Thus I have chosen to have 8 each of the 4 piece types least suited to being present in small numbers.
	All the same, any comment giving new ideas is welcome. I have been inspired to review my choice of pieces in light of additions to the pool since this idea first came to me, and replaced the Hood, which is really too short-range for a lone piece on a large board, with the long-range Harrower. One of the suggestions for A, Antelope, might be an improvement on Ambrose, both thematically and perhaps in terms of understandability. Coincidentally the Antelope has a subset of the Ambrose destinations but on an unblockable path - like the Nightrider relative to the Rhino. I am deferring highlighting the page as changed until I have decided whether to make that change as well.

I concur with you on Storm the Ivory Tower. Given that the Fortress remains 3x3 the obvious rule would be that:
	orthogonal pieces can move diagonally one step to/from, or two straight through, the centre of the fortress (existing Jang Gi rule);
	diagonal pieces can move orthogonally one step to/from, or two straight through, the centre of the fortress (central orthogonals of Fortress).
	Compound pieces would of course be unaffected.

If I took all criticism of my variants personally I'd have pretty low self-esteem!
	I agree that compound oblique leapers can be overpowering on so modest a board, which is why I use them so sparingly. Perhaps I should have gone for a bigger variant and more piece types. Still, you hav einspired me to go ahead wth a comment of my own on Falcon Chess.
	Castling is a special move, but one that in FIDE Chess (and Yoto) can be replicated in just three ordinary moves if the square in front of the Rook's destination is also empty. A corresponding move can be replicated in Shogi only by dropping a captured piece, and in Xiang Qi not at all. In Xiang Qi on a Shogi Board replication takes five moves but without clearing first. So why exclude it in Yoto, even if its use would probably be rare?

I am further inspired to write in defence of the Falcon piece, at least, by comments on some variants of mine that do not use it. Yes, the Falcon is weaker than the Bison, but too much of a strong piece is not always a good thing.
	Comments on variants using compounds of two oblique leapers have made me reluctant to use them further unless a theme calls for them. They can just about get by on a board of squares, or more sparingly on a hex-prism board, but on a cubic board they can be overpowering. A Gnu, Gazelle, or Bison in the centre of an 8x8x8 board can reach 48 cells, and a Buffalo 72. The same could of course be said of the Churchwarden, Samurai, Overon, and Canoe but at least that lot are confined to the second preimeter.
`	Being blockable a Falcon does not dominate even the cubic board to the same extent, and suggests a logical set of fellow pieces. Where, by mixing Wazir and Ferz steps, it complements the Knight corresponding 3rd-perimeter steppers can be devised mixing Wazir and Viceroy steps to complement the Sexton - call it the Vulture - and mixing Ferz and Viceroy steps to complement the Ninja - call it the Kite. Even their own compounds are not unthinkable with sufficient blocking pieces - say Merlin for Falcon+Vulture, Kestrel for Falcon+Kite, Osprey for Vulture+Kite, and Eagle for the triple compound. In fact I might try out a cubic variant with the compound pieces, if George Duke does not object.

The illustrations of sets do a lot to put this game into its historic and geographic context.

Has anyone else noticed that the Bare Facing rule is an example, many centuries before the rise of music downloads, of a restriction on file sharing?

He is quite right that my piece cataloguing has moved on considerably since my last comments, and I now have variants using the pieces mentioned. The Guru and Sadhu are indeed described in Man and Beast 03: From Ungulates Outward, and the Guru is used in one army each on pages 2 and 6 of my Armies of Faith series. The Sahib, Memsahib, and Nabob are described in Man and Beast 11: Long-nosed Generals, and are demotees on page 5 of Armies of Faith.

I notice that Luiz Carlos Campos has yet to clarify the Camel/Giraffe ambiguity - or correct his Brahmin description.

I'll certainly consider putting something in about that, although it may take time. Your comment reminds me of some classic examples of Jewish humour involving visits to China, which I might add at the same time.

Having seen your comments elsewhere I can see why a variant with Cannons and a Cannonade would appeal to you!

One way to represent all possible pieces would be with 32 identical dice. They could be placed with faces parallel to the cell edges for one player and at 45° to them for the other. 1 would represent Pawn, 2 Knight, and so on upward.
	This could also be applied to Mortal Shogi, but with 40 dodecahedral dice. In that case the top face could be treated as the conventional (though irregular) pentagon that Shogi pieces are with a side facing its own player and a corner the enemy. Of course faces 11 and 12 would nveer be needed. Face 10 would be used only for array Kings, which would stay at that number - likewise face 6 in Mortal Chessgi.
	That gives me an idea for further variants. Start with either array, use dodecahedral dice for capturable pieces and something completely different for Kings, and have pieces return by the Mortal Shogi sequence but with the 'missing' Chessgi pieces inserted appropriately. Intuitive positions are Queen at the top, Knight just below Bishop, and Pawn second to bottom - numbers 12, 7, and 2 on the dice with other pieces upped by 1 or 2. Pawns would be promoted to Queen (possibly with the alternative of Knight as Knights are also unpromotable) and the rest as in standard Shogi. As Queens would have so far to fall before being lost, these variants might be called Vivat Regina Chessgi and Vivat Regina Shogi.

'Pawn promotion can only happen on back hexagons previously occupied by opposing forces.'
So, if they get to the end of file a or l (in the 2 player game) they have to wait until there's an enemy to capture to get them somewhere where they can be promoted? This is a major departure from square-cell and even other hex variants. For myself I consider McCooey's game a greater improvement on Glinsky's, and suspect that others will agree.

A possible solution has occurred to me to the complications of certain Xiang Qi pieces being restricted or having no FIDE Chess counterpart:
	The Elephant is barred from entering enemy territory, but its positional counterpart the Bishop is not, leaving it turning into an Elephant that shouldn't be there. On the other hand the Cannon can enter enemy territory but, in your variant, has no counterpart to become when it gets there. How about making the Bishop transform to a Cannon and vice versa? That way both pieces can continue in a relatively normal way. This also echoes the Bishop's position in Shogi, which could be considered equivalent to one of the Cannons.
	The King and Queen can enter enemy territory, but the reason why the General and Ferz cannot is that the Fortress bars them getting anywhere near it. This could be seen as the Fortress being the real barrier for them, in which case the King and Queen should transform to a General and Ferz unable to enter the Fortress. As they therefore cannot give check they are sufficiently weak pieces to not upset the balance.
	I hope that these ideas go some way to propelling this variant toward excellence.

'The Feeble Knight, on b1, g1, b8, and g8, is initially able to leap in the forwardmost Knightly direction towards the center line (from b1 to c3), and turns 45 degrees.'

Approximately 37 or 53 degrees, actually. The directions of the Knight are at 45 degrees to those of the Camel, not to each other. Likewise its compounds.

It has just occurred to me that strengthened-FIDE-array variants, such as this one and my own recent Overkill Chess and Quadripunch Chess, are particularly suited to combining with my Nearlydouble concept, so that the stronger pieces have a larger board to make better use of their greater powers. Would you be happy for me to include a properly-attributed Nearlydouble Tripunch Chess among a page of such variants?

Some other orthogonal/diagonal pairs of animals that might be added are
Panda and Bear (former distinguished by patch of opposite colour around
eye) and Sow and Boar (latter distinguished by tusks). Along with the
suggested Wildebeest character it might be worth including a slimmed-down
version for for Antelope or Gazelle depending on context. If some kind of
'striping' is to be applied to the Knight character for the Zebra, it
could also be applied to the Camel for the 5:1 Zemel, and to the Elephant
for the piece distinctive to Korean Chess.

The labels for the directions are somewhat confusing as 'oblique' usually indicates a direction such as that of the FIDE Knight, going through intervening cells off-centredly. A more accurate description for the directions of each colour are forward/backward hex-diagonal, sideways orthogonal, forward/backward orthogonal, and sideways hex-diagonal. The linepieces in these directions I term Unicoranker, Rookfiler, Rookranker, and Unicofiler. These definitions also work on a Glinsky board, but on that it is the first two that have four directions and the second two only two.

Hugo is wrong. It is true that any piece moving to threaten an Orphan is automatically threatened by it, but what if the piece is also protected by an ally? Then if the Orphan captures it the Orphan itself can be captured as it has no time to capture the next piece. For example, there is a Black Orphan on a4 and a White Rook moves to d4, where a Bishop on b2 protects it. If the Orphan captures the Rook, the Bishop can capture the Orphan.
	Is this a record for the time taken to reply to a comment? My excuse is that I have only recently become interested in pieces which imitate.

Some other images that could be worth including are:
Aanca/Anchorite - perhaps punningly represented by an anchor
Crooked Bishop
Crooked Rook
Fox
Gryphon
Kangaroo
Squirrel
Tank
Wolf

Notes:
¹same as Base in Prince, but name changed to avoid confusion with suffix -base meaning Man and Beast 12 downward-orientated piece.
²differs from Scientist in Prince in lacking 3d-specific Technician move.
³differs from University in Prince in lacking 3d-specific Technician move

'Really diagonal is just orthogonal on a different, bigger board'
	This is something that I have illustrated with my Nested series of variants. For the implication in 3d, see my comments on Tetrahedral Chess.

'Knights are diagonal but use 2 different diagonals together that make them not colorbound'
	Not technically diagonal but I see what you mean. The moves of the Veering Knight and Backing Knight are again the orthogonals of a smaller board:
	.*....*.. ..*....*.
	...*....* *....*...
	*....*... ...*....*
	..*....*. .*....*..
	....@.... ....@....
	.*....*.. ..*....*.
	...*....* *....*...
	*....*... ...*....*
	..*....*. .*....*..

'hunters (pieces that move and capture in diferent ways)'
	It is snipers that have different noncapturing and capturing moves; hunters have different forward and backward moves (and no same-rank ones).

'The precedent for the Ajax-Pieces not being able to capture with their adopted Commoner moves is the Pawn.'
	Not really, the Pawn uses one type of direction in which it can move without capturing, and one type in which it can capture, but none in which it can do both. Likewise the Yeoman, Steward, and other offshoots.
	All the traditional 'crownings' of linepieces (Shogi, Duke of Rutland, Wellisch hex &c) include the ability to capture with the extra move. Indeed you use images whose usual meaning is a piece that can capture in all its directions. Your new images could prove more popular for straightforward Rook+Knight+Ferz and Bishop+Knight+Wazir. As it happens I have been writing a page whose introduction mentions Rook+Knight+Ferz, although as far as I know it has yet to make it into any actual games.

I am hesitant to criticise a variant by one of the Polgar family, but a talent for playing on square-cell boards does not necessarily imply one for designing games for hex ones. This does look very muvch like a game by someone who has not made a great study of hex variants, as it addresses several issues of the hex board less well than variants on these pages do.

A severely bound Rookranker is really a very poor analogue to the Rook. A better piece to complement the Rookfiler here (or the Rookranker in the Wellisch orientation) would be the Moorhen - a hex piece moving straight forward/backward/left/right regardless of which two are orthogonal and which hex-diagonal. This is bound to alternate files here and alternate ranks on Wellisch boards. However it would then be logical for the Queen analogue to also include the straight sideways directions. As regards subdividing of just Rook directions, my own approach to this in Altorth Hex Chess avoided severe bindings and was also Migrant-based.

It is also odd that Migrants line up with their own edge of the board rather than - as in Glinsky's game - the far edge to which they are aiming. It would make more sense on a star-shaped board to arrange a row of Pawn analogues with the middle one furthest back rather than further forward, as in my own Flatstar. At first I thought that a 37-cell might be too small for that, but it could be done with six spaces behind to fill, in two blocks of three - rather than a single back row of five. Ther weakest piece would be doubled in number - the Rookfiler in the case of Mr. Polgar's own choice of pieces. The array prior to placing the back pieces would be (excuse the crude colouring):

Now that I think about it I haven't devised names for pieces moving at least two staps along one kind of radial and at most two along another, but I can see that they are interesting pieces. Pieces that could be seen as Mansion+Ferz and Dean+Wazir are intermediate between the Mansion and Dean and corresponnding enhancements of full linepieces such as the Infanta and Inquisitor - whicvh could be seen as Mansion+Wazir and Dean+Ferz. If this inspires any ideas for names I would be interested to hear them.

Well most of my Quadruple Besiege variants have at least twice as many pieces with a Rook move, so that will help Checkmate to happen. Remember that the minimum to Checkmate on a board with edges is a Rook plus one's own King, so an extra Rook to effect a virtual edge should allow the same on this board.
	The geometry is not quite Moebius, it's a bit more complex than that. A single orthogonal step across a horizontal join appears as a 10:9 leap. Bishops really are colourbound and, as I say in the text, each visible 10-cell diagonal loops round dircetly on itself.

The point is that the Huntsman and Hawksman are defined on a corner orientation. In this context the forward diagonal is toward the opponent's corner and the backward one toward one's own corner. The directions at right angles to these I term sideways diagonals. There are also two forward and two backward orthogonals in this orientation. Thus to sum up the differences between linepieces with 5-6 directions they divide into:
	Goldrider (face-to-face) - 4 orthogonal and 2 diagonal;
	Goldrider (corner) - 4 orthogonal and 1 diagonal;
	Silverider (face-to-face) - 4 diagonal and 1 orthogonal;
	Silverider (corner) - 4 diagonal and 2 orthogonal;
	Huntress and Hawkress (face-to-face) - 3 orthogonal and 2 diagonal;
	Huntsman and Hawksman (corner) - 3 diagonal and 2 orthogonal.
Would diagrams help? If so I will endeavour to add them when I have more time.

None of this page's long-range pieces are switching. The Rhino's first three destinations are those of the Wazir, Knight, and Camel. Knight plus Camel equals famously triangulating Gnu. Likewise the even destinations (exactly as with the Mirror Rhino) are destination of the Nightrider - a straight linepiece like the Bishop and Rook and so able to make two moves in the same direction and return in a single move the same length as the two together.
	Indeed not even a Waverer, a Rhino restricted to moves of odd numbers of steps, is switching as a Camel move can be reversed in four Wazir ones. Nor is a Feverer, a Mirror Rhino so restricted, as a Ferz move can be reversed in two Zebra ones. It may be more difficult when what I am for short calling Camel/Zebra moves are stepping ones here, but it is posible.

I would grateful if some editor could make the correction - and correct 'aranged' to 'arranged' while we're at it.

As I understand it there were royal and non-royal caliphs, just as there
are royal and non-royal governors. Caliph has the advantages that it can be
extrapolated, giving along with Bishop+Knight=Cardinal names for all Bishop
compounds with all coprime oblique leapers. Thus Zebra gives Zerdinal,
Giraffe Girdinal, Antelope Nardinal, Zemel Zeliph, Satyr Sardinal, Gimel
Giliph, Rector Rerdinal... If anyone can think of a better alternative that
can be extrapolated as obviously I'm eager to know it. Likewise for the
Rook compounds Canvasser gives Rook+Zemel=Zenvasser,
rook+Gimel=Ginvasser...