Comments by johnnyluken

Does Freelings claim that Grand Chess is an inherent improvement of FIDE Chess stand up to scrutiny?

Removal of castling

Castling is an artificial but effective rule that serves more than one purpose-the ability to switch an immobile king to prevent lopsided enemy attacks, and increased ease of rook developement. Staggered rooks as an alternative accomplish the latter, a more centralised king on a more spacious board (mostly) negates the former.

Extension of material

The premise in conventional chess variant wisdom seems to be that the choice of the RB compound is arbitrary, and the RN/NB are the natural "missing" extensions. However the FIDE Queen is arguably the most conceptually fundamental piece in the game; its movement on an empty board can be described in 3 words; "it moves straight." The same certainly cannot be said of RN/RB.

K/B/R are restrictions, rather than fundamental building blocks. Similarly the knight is really a special case that subsets the more obscure 2-1 slider. It, and not the "mad Queen" is the true wildcard of FIDE.

Either way, the knight complements the FIDE array perfectly, and gives the ensemble a high degree of balance for such simple pieces - 8 pawns, 4 minor pieces, 2 major pieces, one 1 "master" piece, thats difficult to better, and in my opinion distributing its most obscure movement type in new combinations is not sound grounds for doing so.

In truth, the weakening of the knight move of a 10*10 board aids GC somewhat declustering the pieces and producing a clearer hierarchy, but not enough, and the final ensemble is undeniably lopsided.

I do feel that FIDE is missing a piece (and just one), but I would consider the 2-leaper (a piece so neglected among variants that it barely has a name) to be that piece. Its conceptually simple and bridges the gap between Queen and Rook almost perfectly, being in almost exactly equal power ratio to each.

Aesthetics

FIDE is played on a lower base board (2 vs 10), with perfect 50% piece density.

Pawn promotion

This is where I feel Freeling makes a real mistep. The optional promotion of the 8/9/10th rank is slack and the restriction of promotion to a captured piece is an archaic throwback to precomputerised chess. Freelings defense of the unnecessary complications that arise (pawns on the 9th ranks can give check while immobile) by pointing to the case of pinned pieces in FIDE yet giving check is at best a case of two wrongs not making a right. Why not enforce promotion to the RNB and complete the (R, N, B) power set?

Stalemate, pawn first move and en passant

The primitive stalemate rule of FIDE is left unchanged (piece vs bare king still irrationally given as a draw), and convuluted pawn behaviour is left as it was.

Conclusion

I don't doubt that GC is still an excellent game and most likely the best of its type, but its just not a game that can be considered a clear forward step from FIDE. It extends in an abritrary manner, improves in some areas, loses in others and leaves other chess conventions unchallenged.

Freeling showed an ability to distill the chess paradigm to clear endpoints in Rotary, Shakti and Chad, but ultimately GC can't be considered in that group.

Like the swapper, the withdrawer can be extended quite logically.

Allow it to capture pieces in its line of sight as many squares distant or less as it moves when making the capture.

It already does this when capturing adjacent pieces.

In this way, the withdrawer could capture pieces as many as 4 squares distant using the outer edge of the board.

A variant that has no doubt been reinvented several times.

The archer is certainly more canon to chess than the "camel" and like pieces, which serve testament to the underwhelmingly barren nomenclature generally exhibited in greater chess (Ren Chess being one of the few variants that made a genuine attempt to expand chess in a faithful manner).

Personally I've always viewed the archer as the triangular analogue of the chinese cannon, though other interesting interpretations exist, typically revolving around short range rifle capture.

I see little need to force promotion into this game. It is sufficiently complex.

If an extra dimension really needs to be added, then conversion capture is a much more natural fit, with only the king currently using replacement capture.

I'm not really sympathetic in general to attempts to turn Ultima and its derivatives back into FIDE.

More interesting to purify the concept, which would be a game in which all pieces have equal movement, differentiated only by behaviour.

This games prides itself on conceptual simplicity, and indeed the 5 most fundamental Ultima types manifest themselves in some form or other, 

Replacer (x, y) => (0, x)

Neutral (x, y) => (x, y)

Codestructor (x, y) => (0, 0)

Archer (x, y) => (x, 0)

Swapper (x, y) => (y, x)

although only the latterly added Archer exists as a standalone representation.

Aronson attached codestruction to his swapper, but both concepts are better fleshed out separately in my opinion.

My amendment to the swapper is specified below, while the Queen from Atomic Chess as representative of the latter concept seems a very natural addition to this game (albeit it requires staggered opposition pawn rank to prevent early slaughter).

Its worth pointing out that the immobiliser in this game is stronger than the original concept of Abbots game. Abbott floated the idea of a "Neutraliser" 

(x, y) => ((y, x') => (y, x'))

as a separate piece, which acts as the twin of the immobiliser (one disables movement, the other the effect). As its able to not only freeze but entirely disable the chameleon, swapper and archer, the Rococo immobiliser is really a compound of both, which goes towards explaining why its so overpowered.

Some thing that mystifies me, is why the preference for a randomised setup in Chess 960.

A game that had a "zeroeth move" allowing both players to choose their preferred 960 array has potentially richer strategy, and the potential for black to counterract white advantage, by having the "answering" array.

A classic example of a game whose popularity exceeds its actual quality.

The addition of conversion to chess is a worthwhile pursuit, but the brainless mechanic of dropping a piece wherever you want, is the least imaginitive possible implementation.

More specifically the freedom to drop pieces produces a higher level of convergence in the game tree versus more restricted implementations, reducing strategic connotations of moves. Piece drops in Crazyhouse are always done on primitive grounds, check blocks, pawn promotion threat etc.

That it merely borrows this from Shogi is not a defense; those games are 1) somewhat aged, 2) purposely designed towards such a mechanic.

Were Chess capture performed by nonreplacement, pieces could simply be converted immediately and this would likely work well.

As it is, I believe there are two main implementations.

1. allow pieces of like colour to occupy common space with immediate conversion. This is not satisfactory as it simply allows the second player in a trade cycle to gain all the pieces.

2. my proposition. Captured piece is immediately converted, continues to occupy its cell and can be played on as usual. However it may not be captured at this point, and may not capture on its first move after conversion. 

On playtesting this idea, I further propose that converted piece must wait one turn before being played into a game-this avoids attritional cycles with little change to the board.

I would have to extend my criticism of Crazyhouse to this game, and echo Mr Dukes sentiments.

While I like the concept, I don't think its an actually good game.

My proposed amendment of the drop in rule (pieces are played in with non capture from the spot they were captured) likely brings an improvement here as well.

Non immediate promotion is also unsatisfactory; why not allow pieces to promote immediately on rotationally symmetric opposite squares?

To me these are the most logical ways of importing Shogis mechanics into the more energetic game of FIDE.

This game is conceptually appealing yet famously poor.

An improved subvariant is what I would call "Spotters" Chess, where a piece captures only when another friendly piece has the piece in its sight (the spotter).

On playtesting this subvariant is decent, but still not great.

Making captured pieces be converted adds complexity. Maybe that would make a worthwhile game for this concept.

In truth, Ultima pieces don't really have fixed values but rather discrete sets of values that vary based on the nature and composition of other pieces.

Looping around to give the weakest pieces resistance against the strongest only enhances this "rock-paper-scissors" dynamic (not that that necessarily makes it a worse game).

Regression to predetermined strategy would be countteracted by Blacks freedom to choose an answering array.

This gives 921,600 starting positions.

Depending on how much of a counteradvantage Black gains, moves could be staggered-White places a piece, Black answers.

A constructive phase would add to Chesses itinerary, though how much genuine extra depth is added by such pregame metastrategy and how much real nuance these premoves would have I'm not sure.

I think users should freely be allowed to add piece concepts to this page.

Rather than having adding to the heap of variants as the only outlet, why not allow
users to extract the component parts and simply write on them, exposing
more information that other designers can build off in the process?

I myself have concepts for hundreds of pieces (mostly Ultima like), but no
real interest in posting the endless variants that permute from them.

I would however submit entries on the piece concepts.

The piececlopedia page seemed to have been arbitrarily frozen at a certain
point, as if its list is somehow exhaustive, and should no longer be
tampered with.

This goes against the spirit of this website.