Check out Chess with Different Armies, our featured variant for July, 2024.

This page is written by the game's inventor, Derick Peterson.

Peterson's 3-d Space Chess

Derick R. Peterson invented in 1997/1998 rules to make use of his 8x8x3 Space Chess board. He wrote: These rules are much more 3-dimensional (and chess-like) than the ones which came with my board, I think. As I (HB) also own a Space Chess board, I think he is right: while the set has great ornamental value, the rules that come with the set are poor, and Derick's rules are a good alternative and correct many of the shortcomings of the standard Space Chess rules. Below, you read Derick's description of his 3d-chess variant.



White starts on level 1, and Black on level 3, in the usual initial setup with the standard set of 16 pieces. In addition, each army has 8 more Pawns which begin on the back rank of level 2, directly above White's and below Black's pieces.


Knights are the only pieces which jump directly to their destination square, unobstructed by pieces in between. The Knight always moves one square in one orthogonal direction followed by 2 squares in a second orthogonal direction for a total of up to 16 destination squares, 8 on their level and 4 on each other level. The Knight always changes color when it moves. When a Knight moves up (or down) a single level, it lands on a square exactly 2 steps away from the square above (or below) it along a rank or file on the new level. When the Knight moves up (or down) 2 levels, it lands on a square directly adjacent to the one directly above (or below) it.


Rooks move any number of unobstructed squares orthogonally in any of the 6 directions: up, down, right, left, forward, backward. Thus Rooks can only change levels if the square above or below them is vacant. The number of squares attacked by a Rook is only increased by 2 compared to 2-D chess since there are only 3 levels. However, Rooks can be easily developed by changing levels after only a single Pawn move, and they can quickly reach any square on the board.


Bishops move any number of unobstructed squares diagonally in planes parallel to the 3 levels or the edges of the levels only. Thus they have 12 directions to choose from. Bishops never change color; one controls the light squares and the other controls the dark squares. Like Knights, Bishops can easily change levels, and they can be developed on the first move of the game. When a Bishop moves up (or down) 1 level, it lands on a square of the same color directly adjacent to the square above (or below) its original position. When it moves up (or down) 2 levels, it lands exactly 2 squares away from the square above (or below) its original position. Compare this to the Knight's move, which is the exact opposite when changing levels.


The Queen combines the powers of the Rook and the Bishop, allowing her to move in 18 directions. Note that even she may not move in the long 3-dimensional diagonals which are defined by moving one unit in all 3 dimensions. She moves in only 18 of the 26 possible directions.


The King can capture to the 18 adjacent squares in the directions of the Queen's move, but his non-capturing move is limited to 9 of these squares. On a given level, the King has his usual 8 moves as in 2-D chess, but on a non-capturing move, the King may only change levels by moving straight up (or down) toward the enemy's starting level. Thus, once the King has left a level, he may only return to it via a capture move or by reaching his elevator square which is the enemy King's Bishop's starting square. As soon as the King reaches his elevator square, he is entitled to move to either of the squares in the same vertical column before the enemy's next turn, provided the square he moves to is vacant and not attacked by an enemy piece; the King cannot capture a piece in getting off the elevator, though he may capture one to get onto it. For example, once the White King has been forced to leave level 1, he may only return to it if he can eventually reach his elevator on level 3 or position himself over an unprotected piece on level 1 and capture it by moving down on top of it. Once the White King moves to level 3, checkmating can take place almost as in 2-D chess since the White King cannot move back down to escape unless Black fails to guard White's elevator or Black's men on level 2; in fact, Black should be able to enforce checkmate even more easily at this point since his pieces can still attack from below as well as from level 3! Restricting the King's movement in 3 dimensions is really necessary in order to enforce checkmate in a reasonable time frame; otherwise, an incredible material advantage (or very poor defense) is required, and the endgame can seem like an endless game. One could limit the King's movement in another way; I chose to allow the King (as well as all other pieces) to function normally on each level. The King's elevator was provided mostly as a means to ensure that a player with a material advantage could get his King to any level to aid in checkmating the enemy King.


Castling is allowed when it is the first move of the King and the Rook, no intervening squares are occupied or attacked (including attacks from above or below!), and the King is not in Check. Only the usual Castling move is permitted; this involves no level change.


Pawns move one square toward the enemy camp, either horizontally or vertically. Thus Pawns move normally on a level, and they move straight up (or down) like the King to change levels. Once a Pawn leaves a level, it may never return.

Pawns capture to diagonally adjacent squares which move them only one step closer to promotion. On a level, they capture as usual. When changing levels, they may capture to either of the opposite colored squares directly to the left and right of the square above (or below). No Pawn may capture to the square which it could reach in two of its non-capturing moves, despite the fact that this would also be a one-step Bishop move toward the enemy camp. Thus Pawns attack only 4 (not 5) squares.

On its first move, a Pawn may move 1, 2, or 3 squares. This move can be any combination of horizontal and vertical moves. Thus a Pawn's first move could look like a Knight move, though the intervening squares cannot be jumped. A Pawn which runs by an enemy Pawn by way of a double or triple first move is subject to En Passant capture. For this purpose, the attacker can interpret the double or triple 1st move in any fashion he chooses; you may not specify that you moved up one and THEN forward two.

A Pawn may promote to any piece when it gets within one step of the enemy camp (including the back rank of level 2 and the 7th rank of the enemy's level), and it must promote upon reaching the enemy camp since otherwise it would have no legal move on the next turn. As in FIDE Chess, it is legal to have more than 1 Queen etc.

Written by Derick R. Peterson, introduction by Hans Bodlaender.
WWW page created: February 10, 1998.