The Chess Variant Pages

Check out Marseillais Chess, our featured variant for February, 2024.

This page is written by the game's inventor, Hans Bodlaender.

Go with chess pieces

In the end of 1997, Alfred Pfeiffer invented Chego, a chess variant that borrows some ideas from Go. For a competition to design a chess variant on a board with 38 squares, he designed four boards with 38 squares, and with Eric Greenwood, I played a game on an H-shaped board with the pieces of Hans37 chess (a variant invented by Ralph Betza on the occasion of my 37th birthday). Thinking about the game, I noted that while Chego is a crossover between chess and go, but not a generalization of go. Then, I thought up to play go with chess pieces. For doing this, I first give a different description of the game of go than usual.

Go with rooks

`Go with rooks' is a two player board game, played on a board of size 19 by 19. (One can also play the game on differently sized, or even differently shaped boards. The game is usually played on an uncheckered board, but one can of course use a checkered board as well.)

Each player has an unlimited reserve of rooks. (These are often called `stones' and look like white or black M&M's. However, here, I will call them rooks.)

A move of a player consists of dropping one of his rooks on an empty square, or of passing. Once placed, pieces do not move.

Every piece that can move to an empty square is alive. Every piece that defends a friendly piece that is alive is also alive. So, in general, a piece is alive when there is a chain of pieces, each can move to the position of the next, such that the last of the chain can move to an empty square.

After a drop of a piece, one looks whether this drop makes some pieces of the opponent no longer alive. All such pieces are removed simultaneously. It is not allowed to drop a piece to a position where it takes no pieces of the opponent and where it is not alive.

Ko rule

Making a move such that the position before the opponents last turn comes up again is illegal. (Example: white has rooks on c2, b3, c4. black has rooks on d2, c3, d4, e3. When white drops a rook on d3, it takes the black rook on c3. Black cannot directly take back by dropping a rook on c3.)


The game ends when both players pass. A square is controlled by a player if at least one of his pieces can move to the square (in one or more chess moves), while none of the opponents pieces can move to the square. Each player receives one point for each piece taken from the opponent, and one point for each controlled square.

Comment on `go with rooks'

Go with rooks is, if I did not make a mistake, just go with its rules formulated differently. Instead of rooks, one could use wazirs: pieces that move one square orthogonally, and one still has go. But, the way the rules of go are now formulated make it possible to define variants of go where different pieces are used. Just replace `rooks' in the description above by a different chess piece, and we have a different game. While this website is on chess variants, I'm not sure the following games would count as chess variants or go variants, but here they are:

Go with bishops

In a certain sense, go with bishops is just `go' on a differently shaped board. As bishops on white and black squares do not influence each other, there are actually two different games played simultaneously, and black has a rather easy draw strategy: just mirror the moves of white on the other colored squares. So, when playing go with bishops, one should stipulate that only moves on white squares are legal (or, if one wants, only moves on black squares are legal.) Play go with bishops on a checkered board.

Go with kings or queens

If we play with pieces that can move to all eight directions (there is actually no difference in whether we use kings or queens), then we get a go variant that might be interesting. Go experts probably will know what is wrong. (In go terms: stones that are diagonally adjacent are considered to belong to the same group.) Probably, it is too easy to make groups live, as eyes are formed about as easily as in `go with rooks', but groups can much easier connect to other groups.

Go with knights

Using knights might make a game that is rather confusing, but may be interesting to try out. Knights have large mobility, and their ability to jump makes that they connect easily. However, forming `eyes' might also be harder in this game.

Joao Pedro Neto tried a game of Go with knights. He wrote:

I read your article about go with diferent pieces, and I tried a game of knight go.

curiously, after the 1st surprise about the confusion of the board, we start to see a pattern.

Since a piece has usually 8 liberties and each connection has a 1-2 square range (knight jump) it is almost impossible to delete a group until the *end of the game*.

So you just construct at least two connected eyes, and then try to connect all pieces to that eyes. The one with most eyes win. A simple strategy that somehow demolishes the interest of the game.

well this is just a opinion, of course :)

Go with pawns

Go with pawns looks like an interesting game. As movement directions, one should take the `diagonally forward', i.e., capturing directions (also for counting dominated squares), as when we use the non-capturing direction, the game is boring as one can always take the last piece dropped by the opponent, unless prevented by the ko rule. People that demand to use the non-capturing direction should use Berolina pawns instead (these move without capturing diagonally forward and take straight forward.)

It seems there are nice little tactical possibilities here. The game is not as deep as go (or chess, for that matter), but could still be somewhat interesting.

Go with other fairy chess pieces

Are there other fairy chess pieces for which this gives an interesting game? For instance, one could consider dabbaba's (pieces that jump two squares orthogonally), or the elephants from Chaturanga and Shatranj (pieces that jump two squares diagonally.) Or consider the Chinese knights (knights that do not jump.)

Go with different pieces

The way the game is formulated, one can also play with different pieces. A player has an unlimited supply of a number of different pieces. It makes actually no sense to use pieces that have all movement possibilities of another piece also used, e.g., to use both rooks and queens, as in such a case the weaker piece (rook in the example) never should be used.

But, one could try combinations like rooks and bishops, i.e., some pieces connect diagonally to the group, and other pieces connect orthogonally to the group.

Moving pieces

A final idea is to drop the rule that pieces may not move. Now, a move of a player consists of either dropping a piece, making a chess move with one of his pieces (with the piece going to an empty square), or passing. Capturing is still by having the pieces of the opponent not having a `movement chain' to an empty square.

Final comments

These are probably just a bunch of silly ideas. However, some of these variants might give somewhat nice strategic games. I hope I did not offend go players with my unusual description of their game.
Written by Hans Bodlaender.
WWW page created: March 16, 1998. Last modified: April 3, 1998.