You are on the backup site for Any posts, moves, or other changes you make here will not be permanent, because the pages and database from the main site will be backed up here every midnight EST. Additionally, things may not be working right, because this site is also a testbed for newer system software. So, if you are not here to test, develop, or merely read this site, you may want to change .org to .com in the navigation bar and go to the main site.

The Chess Variant Pages

This page is written by the game's inventor, Charles Gilman.

Switchback Xiang Qi

Xiang Qi is notoriously hard to redesign for different geometries compared to Shogi. Note how hex Shogis tend to be alternative satisfactory approaches whereas hex Xiang Qis tend to be "works in progress" toward a satisfactory solution. The problem is the weakness of so many pieces. That is why "upgrading" to Yang Qi first and adapting that to the new geometry often works better, hence my Liu Yang. Yang Qi also addresses Xiang Qi pieces' short range, which even hinders use of larger square-cell boards. Thus my Nearlydouble series includes Long Yang and Strong Yang but no direct version of Xiang Qi. Even Nearlydouble variants based on Shogi require some piece modification to overcome the short range.

By 3d standards, my 3d Minishogi was well received. That game has the same number of files as standard 2d Shogi, and it occurred to me that preserving the number of files was the key to a 3d Xiang Qi much closer to the 2d original than, say, my Maolike-piece-heavy Triaxial Qi. Unlike 3d Minishogi I do not reduce the number of ranks, as Elephants still require each player's half of the board to have 5 ranks. Therefore I simply folded up the standard Xiang Qi board into three lots of 3 files, affecting the 2d moves in the intuitive way without 3d-specific moves. This approach raises the question of how this variant compares with Corrugated Chess (also without 3d-specific moves) for title of most modestly-3d variant. I also decided to enlarge the Fortress by interpreting the 2d definition as the back three cells of "the General files and all orthogonally adjacent files" - adding Knight files once the board is folded.


Note that the middle level is taken to be reversed, so that each pair of originally-adjacent Ferz and Elephant files remain adjacent, albeit vertically. This allows the statement that any move within two originally-adjacent files (as well as within a single file or one of the three thirds) remains from 2d Xiang Qi.

top level

middle level

bottom level


Pieces are as in Xiang Qi, but affected as follows by the new geometry:
The GENERAL can now move directly up or down from its own file (to Knight ones) and back as well as directly left and right (to Ferz ones). It cannot also move within Knight file exactly as it cannot within its own and Ferz ones. It cannot move directly between a Ferz file and a Knight file.
The FERZ can now move diagonally up or down between the General file and either Knight file as well as diagonally left and right - though never beyond the third rank. It can also move directly between a Ferz file and a Knight file, within a rank, enabling triangulation.
The ELEPHANT can no longer reach the third cell of the General's rank. It can however move directly from the third cell of either Rook file to the first or fifth cell of either Elephant file or vice versa, contingent on an empty halfway cell in the Ferz or Knight file. It can also move direct between opposite corners of one rank, contingent on an empty halfway cell in the General's file, enabling triangulation.
The KNIGHT retains its two-rank moves, contingent on an empty cell where the halfway file interects the starting file. It also retains its one-rank moves between a Rook file and an Elephant file and between the two Ferz files, and new one-rank moves between a Rook file and the "other" Elephant file and between the two Knight files. These are contingent on an empty cell where the starting rank intersects respectively the Knight, General, Ferz, and General file. It can however no longer make a one-rank move to or from the General file, or between a Ferz file and a Knight file.
The ROOK and CANNON cannot "turn a corner" within a rank, but can move directly up or down - cubic geometry's usual effect on orthogonal pieces.
The POINT is unaffected until it reaches the enemy half of the board and is promoted. It then gains moves one step up or down as well as sideways. This gives it the same range of moves within the final rank as the General in 2d Xiang Qi has within its Fortress.


As in 2d Xiang Qi there is no initial double-step move, En Passant, or Castling. Xiang Qi restrictions (including to a larger Fortress), Check, Checkmate, and Stalemate are as in 2d Xiang Qi.


I briefly considered applying it to 12-file variants, as Bishops would not be quite so hampered by a maximum 3 steps than a maximum 2. In Courier and Xiang Courier the effect on Elephants is so severe - taking out access to the middle level reduces their accessible cells by a third - that I rejected it for these. It might however work for Courier-Spiel, Courier de la Dama, and my spinoff thereof Courier de los Combinados, which have no piece bound to less than half the board. Once you reach 16 files, as in the likes of Doublewide, this is too far removed from the simplicity being sought here and encroaches on the likes of Tunnelchess and Quadlevel.

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.

By Charles Gilman.
Web page created: 2009-04-11. Web page last updated: 2016-03-18