Space Chess

by George Jelliss
From Variant Chess, Volume 3, Issue 23, Spring 1997, pages 52-53
Back to: Home Page
By Space Chess I mean any form of chess that uses a three-dimensional board and in which the rules for the pieces, other than the pawns, are the same in all lines and planes of the board. This latter condition is added to exclude such games as Parton`s Alice Chess, using two boards, or Beatty`s Total Chess, using four boards, in which the moves between the boards are by 'lift'.

For descriptions of the two main types of space chess known hitherto, those by Maack and Kobetliantz, and numerous others, I will have (for reasons of space) to refer the reader to David Pritchard`s Encyclopedia. Here I outline the general principles and follow this with a new game of my own, provisionally called Cuboid Chess.

Three-dimensional 'boards' are shown by means of a set of 2-dimensional boards representing successive slices through the 3-dimensional region, the labels of the boards, A, B, C, ... acting as the third coordinate. The boards used hitherto have nearly always been cubes n譶譶, for some number n, or a box-shape a譩譪. Cuboid Chess however is played within a very roughly spherical volume (a 6󬝲 cube with a 44 area added to each face).

A 4󫶘 space chess can be played on the normal 88 board, the four layers A, B, C, D being the four quarters, but the chequering may be misleading.

A move in three dimensions is represented by three coordinates (r, s, t), indicating that the move is equivalent to r (wazir) steps right, s steps forward and t steps up. (Or respectively left, backward and down if the numbers are negative.) We express the 'pattern' of the move by the corresponding positive or zero values {r, s, t}. The length of an {r, s, t} move is √(r + s + t).


Some moves of {1,2,3} X, and {0,1,2} * from the centre, Dd4, of a 7󬱟 board.

If all three numbers are different and none is zero an {r, s, t}-leaper can move in 48 different directions! The simplest leaper of this type is the {1, 2, 3}-leaper, or √14-leaper, called a Hippogriff in Kogbetliantz`s game. If one pair of the coordinates is taken to specify a move on one of the layers of the board, then the third coordinate represents a move up or down. Thus a {1, 2, 3}-leaper either moves vertically one layer and makes a {2, 3} zebra move or vertically 2 and makes a {1, 3} camel move or vertically 3 for a {1, 2} knight move, as illustrated by the marks X in the figure. In each case there are 16 choices of move (8 leaps 2 vertical ways) whence 48. Since this piece has an even number of odd coordinates it is confined to cells of one shade in the chequering.

Fortunately when there is a zero coordinate or two coordinates are equal the number of moves to be considered reduces considerably. Any 2-D move can be made in 3-D space in each of the three coordinate planes passing through the cell initially occupied by the moving piece. In this way each 2-D (or 1-D) piece defines a corresponding Space Piece. Thus the Space Rook makes moves of type {0, 0, n} in 6 directions (up-down, left-right, and to-fro), where n can take any value except zero, the Space Bishop makes moves of type {0, n, n} in 12 directions and the Space Knight moves {0, 1, 2} in 24 directions (shown by the marks * in the figure).

Moves with all three coordinates non-zero are 'true' 3-dimensional moves, and pieces making them are 'essentially' 3-D pieces. The simplest is the Sprite, which is the {1, 1, 1}-leaper, or √3-leaper. Its corresponding line piece, the {n, n, n}-mover, or √3-rider, is called the Unicorn in Maack`s game (but 'Fool' in Kogbetliantz`s).

The Rook moves through the faces of the cubic cells, the Bishop moves through the edges of the cubes, and the Unicorn moves through the corners. Thus a Unicorn has a choice of 8 directions of movement. A Rook has access to all the cells of the board, in a series of moves, but it takes two Bishops to patrol the whole space, and four Unicorns.

The King in two-dimensions can be defined either as a Wazir + Fers or as a piece that moves to every 'adjacent' cell (by which we mean any cell that has a boundary point in common with the initial cell). In three dimensions these two definitions are not equivalent. From the definition of Space Piece given above it follows that a piece that has the usual King moves in any plane is properly called a Space King. A piece that moves to every adjacent cell in space is a Wazir + Fers + Sprite which I call a Cubi(c)-King. The royal pieces in the Maack and Kogbetliantz games are of this type. The Cubi-King moves to all 26 outer cells of the 3󫢫 cube around it.

The choice of the Cubi-King rather than the Space King as the 3-D monarch is supported by the fact that one Space King could stalemate another on a cubical board (e.g. on Aa1 and Bb2) which may be considered an undesirable property for the royal piece, or at least not analogous to the 2-D case.

In Cuboid Chess however I use Kings of both types.

The Space Queen, is of course Rook + Bishop (used in Kogbetliantz`s game under the name 'Favourite'). Its more powerful cousin, the Cubi-Queen, runs in the directions that the Cubi-King walks, i.e. Rook + Bishop + Unicorn. Both types of Queen are used in my new game.

In Cuboid Space Chess there is a Royal Family of four: Space King and Queen and Cubic King and Queen. All are Royal, but the Queens are permitted to pass through check. The aim is to capture any one of the four. Thus checkmate occurs if any one of the four pieces is in check and the check cannot be parried without leaving at least one of the four in check. The Space King, being the weakest of the four is the piece most likely to be attacked, but forks and skewers can also be fatal.

This weaker type of checkmate is to counterbalance the increased mobility of all pieces in space (which, as Schmittberger has pointed out, see EVC p.310, makes normal checkmate difficult if not impossible.)

It is convenient to call the R + U a Cubi-Rook and B + U Cu(bi)-Bishop. A pawn moves like a king but forwards only (i.e. towards the opponent`s side), so we can have Space Pawns and Cubi-Pawns, the latter having the added Sprite capture move.

Cuboid Space Chess

An impression of the Cuboidal Board in the round.


The opening position for 'Ordinary' Cuboid Chess is shown above. Two normal, but distinguishable sets are needed. The upright symbols are Space pieces and the inverted symbols Cubi-pieces. The Royals occupy the central cells of the 44 home base. The inverted Knight is the {1, 2, 3}, but other pieces can optionally be substituted.

Let us finally examine some other distinctive space- piece possibilities. The 3D-moves with coordinates non- zero and less than 3 are: {1, 1, 1} {1, 1, 2}, {1, 2, 2} and {2, 2, 2}. The first and last are Unicorn moves.

The {1, 1, 2} move is of length √6, and I call the corresponding rider a Sexton, it is confined to cells of one colour like a Bishop, and can reach all the cells of that colour (PUZZLE 1: Show how the Sexton gets to a diagonally adjacent cell in fewest moves.)

The {1, 2, 2} move is of length √9 = 3. Thus the Threeleaper, becomes of special interest in Space: besides its 1-D move {0, 0, 3} it has this 3-D move. Unlike the Threeleaper in two dimensions, which is confined to 1 cell in every 9, in three dimensions it can get to every cell of the board! (PUZZLE 2: Show how the Threeleaper gets to the next cell in the rank in fewest moves.)

In Space we also encounter two other double-pattern fixed-distance leapers before we reach the familiar 5-leaper, these are the √17-leaper or Space Giraffe which moves {0, 1, 4} and {2, 2, 3} and can reach any cell, and the 3√2-leaper or Space Tripper which moves {0, 3, 3} and {1,1,4}, and is confined to one colour.

In 'Advanced' Cuboid Chess the inverted R, B and N represent Threeleaper, Tripper and Giraffe, the {0, 0, 3} and {0, 3, 3} moves being blockable.

Solutions to Puzzles

Back to: Home Page