Martian Go is an elegant little strategy game for two or four players. It is a turn-based game, unlike Icehouse or IceTowers; and there is no luck involved. You need a chessboard, a set of Icehouse pieces or Zarcana pieces, and four other tokens (which can be anything -- coins, pebbles, Brazil nuts, what have you).

As of July 2002, Martian Go is more or less obsolete. I have posted a new version of this game, called Branches and Twigs and Thorns. The basic concept is the same, but the scoring rules are slightly refined -- this should help fix the persistent tied-game problem that plagued the original Martian Go rules.

So go play that.

Each player gets a stash of 15 pyramids.

If you have four players, lay out the chessboard, and place the four
tokens in the center four squares. These squares are the
**root**. Note that this leaves sixty empty squares,
which is exactly the number of pyramids in the game.

*Four player setup*

If you have two players, fold the chessboard in half (or cover half of it), producing a rectangular board. Place two root tokens in the center of one long side of the rectangle. Again, this leaves exactly as many empty squares as pyramids.

*Two player setup*

Players take turns. On your turn, you place one pyramid in an empty
square. The pyramid must be adjacent to a square which is occupied
(either by the root or another pyramid.) You must place it lying down,
and **pointing at the adjacent occupied square**.

If there is more than one adjacent occupied square, you decide which one to point at. Diagonals don't count.

At the beginning, the only occupied squares are the root squares, so the first player may make any of the following moves:

*Possible first moves*

If the first player (blue) chooses the following move, the second player may make any of these moves:

*Possible second moves*

Note that in one space, the second player has a choice of which way to point.

When you put down a piece pointing at the root, that's safe.

When you put down a piece pointing at another of your own pieces, that's safe.

When you put down a piece pointing at an enemy piece, you take a
**penalty**.
You are penalized according to **the size of the enemy piece you are
pointing at** -- one, two, or three points. The size of the piece
you played yourself isn't important.

At the end of the game, whoever has the fewest penalty points wins.

*A completed two-player game, with penalties marked*

In this game, Blue has taken a penalty of 1 point, and Gold has taken 4. Blue wins.

Obviously, **pointing at the enemy is bad.**
You want to spread a tree of your own pieces across the board, so that
you can always add new pieces to your own tree, and never be forced to
grow off your opponents' trees.

If you can surround an open area, so that you're covering more than your fair share of the board, then eventually an opponent will be forced to play inside that area -- growing from one of your pieces.

Every time an opponent plays a small pyramid, he's making himself vulnerable. Do you want to jump on it, choking off his tree and giving yourself more space, for a mere one-point penalty? If you wait, you may have to do it later -- on your opponent's terms -- perhaps taking two or three points.

Of course, you have small pyramids too. You have five vulnerable points, five average points, and five strong points. Eventually you'll have to play them all. The strategy is deciding when and where.

Aggressive play is not always rewarded, but neither is very defensive play. The last few moves are critical, as the last few open areas are filled in, and players are forced to play their last small pyramids.

For a three-player version, you could block off 16 squares of the chessboard with null (unpointable-at) tokens. Then put out three root tokens. It's hard to make this work, though; there are asymmetries that favor the first player.

(Footnote: Elliott Evans has invented a Three Player Martian Chess Board which works like a charm. It's essentially the above suggestion, three-quarters of a chessboard, but stitched together along the seam to form a half-cube. And then flattened. It looks nice. And you can use the same board segments for four-, five-, and six-player Martian Go.) (Or Martian Chess. Or Zagami, RamBots, ...)

If the standard board setups are getting dull, try varying the location of the root. One player puts a root token anywhere on the board. The remaining players then each put down a token; each token must be placed next to an existing token. The original player then goes first.

Another possibility for four-player games: The first and second players each put a null token on the board. (No piece can point at a null token; it's like the edge of the board, a missing square. No fair isolating a corner square behind two nulls.) The third and fourth players each place a root token; but they don't have to be adjacent. In most cases, this gives eight possible opening moves, just like the standard central root arrangement. (We tested this at Origins 2000, and it worked fine.)

Last updated July 7, 2002.