The Chess Variant Pages

Check out Wildebeest Chess, our featured variant for May, 2023.

Shafran's Hexagonal Chess

Shafran's Hexagonal Chess was invented by Isaak Grigorevich Shafran, a Soviet geologist, in 1939, and it was registered in 1956.



Unlike the completely symmetrical board used for Glinski's Hexagonal Chess or McCooey's Hexagonal Chess, this board has five spaces on four sides and six spaces on the other two. Instead of containing 91 spaces, as that board does, it has only 70 spaces, which brings it closer in size to the 64 square board used for Chess. Note that its files are vertical, and its ranks are descending. By a descending rank, I mean that the a? space is at the highest position in the rank, at least as viewed on the computer screen, and each subsequent position in the rank appears lower. For example, White's left Rook starts on a1, and White's King appears at the lowest space on the board at e1. The spaces to its right are not on the same rank. The Bishop is on f1, the Knight on g3, the second Bishop on h4, and the Rook on i5. Although the rank and file keep changing from White's King to White's right hand Rook, this is only for notational purposes, and movement along this line still counts as orthogonal. Unlike the square board that Chess is played on, movement on the hexagonal board cannot be described in terms of uniform mathematical changes to the coordinates. Instead of that, it will be described in terms of geometric relations.


Except for details concerning castling, the distance a Pawn can move on its first move, and en passant capture, the pieces in this game move exactly like the pieces in McCooey's Hexagonal Chess.


The King moves ones space in any orthogonal or diagonal direction. Orthogonal directions go through the sides of spaces, and diagonal directions go through the corners of spaces. With six sides and six corners, the King can move in up to 12 different directions.


The King castles by moving three spaces toward the Rook with the Rook moving adjacent to the King on the other side. Apart from the longer distance for the King to move, castling works the same as it does in Chess.


The Rook may move any number of spaces in any orthogonal direction until it reaches an occupied space.



The Bishop may move any number of spaces in any diagonal direction until it reaches an occupied space. As in Chess, Bishops are colorbound. Unlike Chess, the board has three colors, and each player has three Bishops to have one for each color. Because each Bishop can reach only one third of the board, a Bishop is less powerful than it is on the Chess board, where it can reach half the spaces.



The Queen moves as a Rook or a Bishop. Note that it can reach any space in its third perimeter, which it cannot do on a square board.



The Knight can leap directly to any space in its fourth perimeter that a Queen cannot reach. As in Chess, each space it can leap to is one that could be reached by moving one space orthogonally, followed by one more space in an outward diagonal direction. Besides not including the inward diagonal directions, outward diagonal directions do not include the sideways diagonal directions that would turn the piece a full 90 degrees.



A Pawn normally moves ones space vertically forward without capturing or one space diagonally forward to capture.


On its first move, each of the three center Pawns may move up to three spaces forward in its own file, the two closest Pawns on each side of these three can move up to two spaces, and the lone Pawn on each edge can move no more than one space forward. The general rule is that each Pawn can move as far as it can on its first move without moving to the opponent's side of the board. But since there are some perfectly midway spaces that each side can reach with a Pawn's first move, this may increase the first-move advantage of White. As in Chess, the double and triple moves must be to an unoccupied space, and the Pawn may not leap over an occupied space. The diagram below shows the vertically forward moves available to each Pawn. Each circle shows a legal non-capturing move for the White Pawn in the same file.


When a Pawn does make a double or triple move, a Pawn that could have captured it if had moved a shorter distance may capture it by moving diagonally forward to the space it would have been able to capture it upon. In the position below, which is possible after three moves, Black's center Pawn has just moved three spaces forward, and the White Pawns on each adjacent file are in position to capture it by en passant. White's d Pawn can move directly behind it, and White's f Pawn can move two spaces directly behind it. Besides capturing the Pawn that just moved by en passant, the f Pawn may also capture Black's g Pawn.



With the exceptions noted above about piece movement, Shafran's Hexagonal Chess is played like Chess. Note that all changes to piece movement are adaptations to the hexagonal board, and none introduce new abilities that the pieces do not possess in Chess.


My sources are the entry on "Hexagonal Chess" in David Pritchard's Encyclopedia of Chess Variants, the description of the game written for Shafran's Hexagonal Chess on Jocly, and Wikipedia's Hexagonal Chess page. Pritchard and the Jocly developers both agree on how Pawns capture, which is different from how they capture in Glinski's, but Wikipedia says all the pieces move as in Glinski's. This may be from a misreading of Pritchard, who wrote:

Pawns capture diagonally ahead on next hex of own colour. Castling permitted: K moves three hexes toward R, thus after O-O, White position is Kh1, Rg1 and after O-O-O, Kb1, Rc1. Other moves as Glinski.

Pritchard did not describe the game's coordinate notation, but from the above, it is evident that he ignored it, as he used Glinski-style coordinates to describe castling. However, both the Jocly developers and Wikipedia are in agreement about the coordinates.

The Java applet for Shafran's Hexagonal Chess does not include a description of the rules, but some testing can reveal what rules it follows. It mostly follows the rules described here, but there are some differences. It handles en passant correctly when a White Pawn moves three spaces forward, but when a Black Pawn does, it will not let a White Pawn that could have captured it if it had moved one space capture it by en passant. However, it will let a Pawn that could have captured it on the space it moved from capture it by moving one space in a slanted orthogonal direction to a space it could not otherwise legally move to, as if Pawns captured as they do in Glinski's Hexagonal Chess. It allows the King to move two or three spaces when castling, but when White tries King-side castling, or Black tries Queen-side castling, which are both on the same physical side of the board, the Rook doesn't move. Inconsistencies in en passant and castling are most likely bugs. But the option to move two or three spaces when castling is consistent and might have been programmed intentionally. It says up front that promotion is allowed only to Queens but also indicates that this is supposed to have been a temporary measure until the game could be programmed better.

This 'user submitted' page is a collaboration between the posting user and the Chess Variant Pages. Registered contributors to the Chess Variant Pages have the ability to post their own works, subject to review and editing by the Chess Variant Pages Editorial Staff.

Author: Fergus Duniho. Inventor: Isaak Grigorevich Shafran.

Last revised by Fergus Duniho.

Web page created: 2023-04-05. Web page last updated: 2023-04-05