Deploying is the action that enables to split a combined piece into its individual components. It is played by taking the combined piece into a sequence of simple (one square) moves, after each of which it leaves one of its simple pieces on the destination square. This process is repeated until the combined piece has left all its simple pieces (or reached past the last rank, thereby winning the game), and is carried out as a single non-divisible action (which requires only one turn).

If this sounds complicated, don't worry! In the following sections you'll be explained how to deploy any combined piece in the game, and what rules are to be observed when deploying.

Double round pieces are deployed forward on diagonals.

Example: the double round piece on d7 may be deployed in two ways: either (a) on c6 and b5, or (b) on e6 and f5.

Double square pieces are deployed either forward or sideways.

Example: the double square piece on e2 may be deployed in three different ways: either (a) on d2 and c2, or (b) on e3 and e4, or (c) on f2 and g2.

Deployment of triple round and triple square pieces is carried out exactly as that of double round and double square pieces, excepts that there is one more simple piece to deploy.

Examples:

• The triple round piece on d7 may be deployed in two ways: either (a) on c6, b5, and a4, or (b) on e6, f5, and g4.
• The triple square piece on e1 may be deployed in three different ways: either (a) on d1, c1, and b1, or (b) on e2, e3, and e4, or (c) on f1, g1, and h1.

One cannot deploy over one's opponent's pieces. Hence, one cannot capture when deploying.

One cannot change direction when deploying (*). For instance, on the above figure, the triple round piece on d7 may not be deployed on e6, d5 and c4.

(*) Excepts if bouncing on a boundary of the board.

One may deploy over one's own pieces, thereby creating other combined pieces, provided that no more than three simple pieces are ever clustered on a single square.

One may win the game in a deployment, if any part of the deployed combined piece reaches past the last rank. Winning deployments are very common, so players should always care a lot about how their opponent's most advanced combined pieces may be deployed.

A combined piece, when deployed, must be fully deployed (*).

(*) That is to say, in the course of the deployment all the simple pieces initially contained in the deployed piece must be placed, one at a time, on a destination square from which the deployment then proceeds. One cannot stop a deployment before it's complete, excepts if the undeployed part of the combined piece reaches past the last rank (thereby winning): obviously this part doesn't need to be deployed further.

Bouncing on the boundaries of the board, which we already discussed in the case of regular moves, also pertains to deploying.

Example: on the opposite figure, the double round piece on g2 may be deployed using two different paths: either (a) f3, e4, or (b) h3, g4 with a bounce on file h (the normal course of the deployment would be: h3, "i4", but there's no "i4" square, so the deployment bounces and the second simple round piece is deployed on g4).

Round-square combined pieces are deployed either:

• by starting with the round piece: its destination is chosen one square forward on a diagonal (as if moving it); then the square piece is placed from there, one square forward or sideways.
• by starting with the square piece: its destination is chosen one square forward or sideways (as if moving it); then the round piece is placed from there, one square forward on a diagonal.

Examples: the opposite figure shows a few possible deployments, starting either with the round piece (Black), or with the square piece (White). In either case, only 3 deployment paths are shown (a, b, and c), whereas round-square pieces may have up to 12 distinct deployment paths!

Round-round-square combined pieces are deployed as round-square pieces, excepts that two round pieces are to be deployed instead of only one.

This deployment may be played either:

• by starting with the round pieces: their destination squares are chosen as when deploying a double round piece; then the square piece is placed, one square forward or sideways from the second round piece.
• by starting with the square piece: its destination is chosen one square forward or sideways (as if moving it); then the round pieces are placed from there, as when deploying a double round piece.

Examples: the opposite figure shows a few possible deployments, starting either with the round pieces or with the square piece. Please note the bounce on the c path for White.

Note: round-round-square pieces may never be deployed in the following order: round piece 1, square piece, round piece 2; both round pieces must be deployed either before or after the simple square piece.

Square-square-round combined pieces are deployed as round-square pieces, excepts that two square pieces are to be deployed instead of only one.

This deployment may be played either:

• by starting with the round piece: its destination is chosen one square forward on a diagonal (as if moving it); then the square pieces are placed from there, as when deploying a double square piece.
• by starting with the square pieces: their destination squares are chosen as when deploying a double square piece; then the round piece is placed, one square forward diagonally from the second square piece.

Note: as before, only a few deployment paths are shown (mixed combined pieces may have up to 12 distinct deployment paths).

Note: square-square-round pieces may never be deployed in the following order: square piece 1, round piece, square piece 2; both square pieces must be deployed either before or after the simple round piece.

With bouncing on boundaries allowed not only when moving but also when deploying, a few odd cases deserve to be mentioned.

In the example shown on the opposite figure, the square-square-round piece on c7 is deployed as follows: first the round piece on b6, then the square pieces, on a6, then... b6 again, after bouncing on file a! This deployment (*) thus results in a new round-square piece being formed on b6.

(*) Since the square-square-round piece is deployed in three separate steps, it is considered to be "fully deployed".

Note: the same deployment (not less surprising for one's opponent) may be performed starting from a7.

The example shown opposite is analoguous. The triple square piece on f2 may be deployed on g2, h2, then g2 after bouncing on the h file. This deployment thus results in a new double square piece being formed on g2.

The opposite figure shows how a triple square piece on g2 may be deployed with a bounce on the h file.

Its destination squares are, in this order: h2, g2 (after bouncing on the h file), then f2 (the resulting position is shown on the next figure).

Game are recorded using a common notation, which principles are best summed up by the examples in the following table.

Notation	Meaning
c2 - c4	Move from square c2 to square c4.
b7 x c6	Move from b7 to c6, with capture (x symbol) of the opponent's piece which was previously on c6.
c1 - c2+	Move from c1 to c2, with clustering (+ symbol) with the piece already present on c2.
c8 - OUT	The piece previously on c8 moves out of the board! (White wins)
c2 * c3, d4, e5	Deployment of a combined piece (* symbol) from c2 to c3, d4, and e5 (in this order). Important: destination squares are always listed in the order of the deployment.
f3 * e4+, d4	Deployment from f3 to e4, d4 (in this order), with clustering (+ symbol) on e4.
b2 * c1, OUT	Black wins by deploying a combined piece from b2 to c1, then out of the board.

You've just finished reading the rules of Gounki (and you shouldn't regret it)! If this sounded a bit too abstract, everything should come a lot clearer with an example.