Comments by zzo38computer

Perhaps a way to make "invulnerability" would be that: king, queen, and chancellor cannot be attacked by spinach moves longer than a sequence of five normal moves, unlessthe moving piece is a king, and any piece that has not yet moved cannot be attacked by spinach moves longer than two, unless the moving piece is a king or pawn. (This results in rules more complicated than you intended, but is another possible subvariant anyways.)

It looks like a pawn can move two spaces forward from its initial position with a spinach move, in order to avoid being captured by en passant, although this doesn't seem it would be a very good move in most situations (although maybe there is one; make up a chess problem if you know of one such situation).

This game was also called "Imperial Fiddlesticks" in a text called "Curiouser&Curiouser" (also by V.R.Parton).

Castling is not mentioned, but I should assume castling is still allowed, although the restriction of not moving into/out of/through check is ignored.

However I may suggest a further variant: You don't lose if you have more than one king. (For example, if you have only one pawn and no king, then promoting will cause you to use instantly, but if you have one pawn and one king, then if you promote, you won't lose until either one of your two kings is being captured.)

According to other comments here, and some of my own ideas, we can write about other kind of subvariants.

It is not mentioned what happen in case of attacking partner's pieces. Here is some possibilities of subvariants to deal with it

1. Attacking partner's pieces is not possible.
2. It is allowed to capture partner's pieces (other than a king), but you are not in check if being threatened only by a partner's pieces.
3. It is allowed to capture partner's pieces (including a king), and there is no check/mate (and therefore you can castle into/out of/through check); you need to capture the king.

One kind of subvariants can be the teams:

1. Partners across from each other.
2. Partners next to each other.
3. No teams.
4. Washizu-style: Partners next to each other, with one player of each team designated as "leader" and one as "supporter". The "supporter" plays immediately after the "leader". Your team instantly wins if the opposing leader's king is captured (or checkmated, depending on subvariant).

And there is winning conditions:

1. Russian: If you are checkmated, all of your pieces (including a king) is removed from the board and you take no more turns. Any team out of players loses.
2. West-European: You pass if you have no legal moves (for kick-the-king variant, also if you have no king). Checkmating both opponents simultaneously wins.
3. My rule: If you have no legal moves or if your king is captured, you pass your turn. If both players of a team pass without either making a move in between, that team loses the game.
4. Bradley's rule: Whenever a player is checkmated (or loses his king in kick-the-king), all of that player's pieces are changed to the color of whoever checkmated/captured him.

Lastly, there are subvariants dealing with fortress:

1. Normal: Each player deploys his three additional pieces inside the fortress at his own choice and it is then visible to everyone.
2. Hidden version: Variant of the above where you cannot see opponent's pieces in opponent's fortresses (you can see them in your/your partner's fortresses, or when they are in the main board, and you can still see your and your team's pieces in any fortress).
3. Random hidden version: Further variant where in addition to hiding pieces like that, the pieces are deployed at random in valid positions (individually for each player).

This looks good to me, however I do not quite understand the rules for castling. Does the raven and tower simply exchange positions (unlike FIDE)? Does the space in between need to be vacant? What about castling through or out of check? Is this allowed in both 2-players and 3-players game, or is only possible with 3-players game?

If you like the game where colorbound pieces are available for all color of cells, you may play the variant:

1. Replace the witch in the starting position by a third broom.
2. Once per game you can change one of your own brooms into a witch, either as a move or immediately after moving that broom (whether or not it is a capturing move).
3. Even by promotion you cannot have more than 1 witch, 3 brooms, 3 towers, and 3 bats, on the board at one time.

How is pawn promoting in this variant?

Do you have the list of games mentioned (with both English and Dutch names of the game)?

Note that a few of them use dice only for indication and not for randomness. (Most of them do use throwing the dice for random, though.) Also note that a few of them have only a very minor randomness (and that I consider a coin toss to be a "d2" dice roll).

(I don't mind either way; I think good games can be made in both ways.)

I didn't say I didn't like the game with dice. It is good too (and better than some of the similar games). However, I find that if there is some form of hidden information which is partially decided at random, that makes it more interesting too (although Game Courier and Zillions do not support this).

George Duke: You say I have "over half a dozen of the fifty", but I don't know if you want to count 123456 Chess because although it uses dice, it doesn't use randomness (Chaotic Chess is the other game with this property) (maybe randomness should be another category).

I do like this idea. However, here is one possible subvariant: A pawn can promote to a queen if you have no queen (so you cannot have two queens). A pawn dropped on the last row still won't be a queen, though. (The game is probably good whether or not you use this subvariant)

I agree this one looks better than the other chess game involving the dice. However, a variant would be to use cards to involve some hidden information into the game too. (Example: Make a deck four cards of each rank one to ten, each player get one card. On your turn pick up another card (if there are no more left, mix up the cards that have been discarded to renew the draw pile), and then choose one card to play. If you pass you have to discard both and pick up one card (so you always end up with one card); you pass only if neither card gives you a capability to move. This game would affect the probabilities though; maybe you have a different idea to make with cards.)

The rules should be made more exact rather than "All disagreements at the table (about the rules) are solved by casting votes among the three participants."

Here are some of my ideas about the movements (I don't know if they are correct):

• Orthogonally is to a cell sharing an edge and having a different color.
• Diagonally is to a cell sharing a corner but not an edge, and having the same color.
• The forward lines for red are: QR1-QR2-QR3-QR4-KR4-KR3-KR2-KR1, QN1-QN2-QN3-QN4-KN4-KN3-KN2-KN1, QB1-QB2-QB3-QB4-KB4-KB3-KB2-KB1, Q1-Q2-Q3-Q4-K4-K3-K2-K1, K1-K2-K3-K4-Q4-Q3-Q2-Q1, KB1-KB2-KB3-KB4-QB4-QB3-QB2-QB1, KN1-KN2-KN3-KN4-QN4-QN3-QN2-QN1, KR1-KR2-KR3-KR4-QR4-QR3-QR2-QR1, Q4-Q3-Q2-Q1, QB4-QB3-QB2-QB1, QN4-QN3-QN2-QN1, QR4-QR3-QR2-QR1, K4-K3-K2-K1, KB4-KB3-KB2-KB1, KN4-KN3-KN2-KN1, KR4-KR3-KR2-KR1.
• Knight moves one space orthogonally followed by one space diagonally, jumping over whatever is in between, but cannot land on a cell orthogonally adjacent to the cell it is moving from. (Example: a knight on K4 may move to QB3, Q2, KB2, KN3, QN4, QB3, K3, Q4, and KB4.)
• A diagonally forward move (for pawn captures) is a diagonal cell which is the forward cell of a orthogonally adjcent cell and which is also orthogonally adjacent to the forward cell. (Example: a red pawn on K4 can capture on QB3 but not on K4; a red pawn on Q4 may capture on K3 and QB3.)
• Note that K4 isn't adjacent to Q4 (it doesn't meet the definitions above), so only knights can move from one to the other.
• A move from Y to Z is in the same direction as a move from X to Y iff all of the following conditions hold:
  1. X isn't adjacent to Z.
  2. X, Y, and Z are all different cells.
  3. X and Z must touch opposite sides/corners of Y.

Here are some of my other ideas of rules (some of which differ from the article and/or comments):

• A pawn promotes if there is no more cells directly forward of it.
• There is no check (this also means you are allowed to castle through check, and move into check, etc).
• If you capture a king, all other pieces of the same color as the king change to your color, except for pawns, which are removed.
• If all pieces are one color that player wins.
• You may pass if you have no legal moves, but if there are two successive passes (not three) then the game ends in a draw; if one player is already eliminated, that player still loses though.

The problem is missing.

It seems like it is supposed to say, you are not allowed to name the same city as before.

The Maharaja is really powerful because it can move twice as often as the other pieces! One possible subvariant to weaken a bit is if you roll a 1 then you can only move 1 space. (This is untested.)

Continuous-space (non-discrete) games will require the correct number of dimensions. Geometry certainly does help to explain many things, and it is very convenient, as well as possibly supporting certain kind of mathematical generalizations; I am just saying that it is equivalent to the game with the different number of dimensions.

I myself would think this game described as one-dimensional with four boards sharing a square seems best, although doing it on a chessboard doesn't seem to work very well like that. But it could be described as 4D, if the rules for some of the pieces movements is corrected to make a bit more sense.