Alak

    Keywords: Variant

Alak is a 1-dimensional Go variant.

A.K. Dewdney, in his book "The Planiverse" (about a 2D universe) mentioned a 2 player game called Alak, which was described as 'go in 1D', but only a few rules were described.

Known rules:

  • Black and white alternate placing stones on a line divided into N sections. (In Planiverse, they are seen using an 11-point board; see diagram below.)
  • When placing a stone surrounds stones of the opposite color, those stones are captured.

Unknown rules: Whether you can you surround at an edge, whether you can fill the last liberty of your own group, ko, passing, win & draw conditions, scoring.[1]

References:

  • Article by Neto and Taylor in Abstract Games #13.
  • Proposal by Alan Baljeu on re.games.abstract, 2001-12-30.

[1] See also: Alak0, Alak Game 1

Original description of Alak

From page 202 of Planiverse:

It is quite an interesting problem to watch a game with which one is totally unfamiliar and try to guess what the rules are. The two astronomers took turns placing pieces of jebb (a type of rock - s) on a board which had been marked off into eleven segments, each having room for only one piece. A long period of thought preceded each move, although sometimes a player would put down his stone and pick up some others.

Later that day Yendred informed us that this was a very popular Punizlan board game called "Alak." One side has white pieces and the other black. If a contiguous set of one color is surrounded (at both ends) by two pieces of the other color, then that player gets to capture all the pieces thus surrounded.

original Alak diagram

From the end of the chapter:

After the contact ended, Ffennell (one of the Earth scientists) took a pad of paper and drew a row of eleven squares on it. Then he collected pennies and nickels from everyone in the room.

"The game of Alak," he said, "is basically just one-dimensional Go, a well-known board game here on Earth. Let's try a few games!" I was too tired to stay on, but Lambert and Chan remained to play. The other two followed me out of the laboratory. Dawn was streaking the early Friday sky.

Baljeu's version

Alan Baljeu proposed a ruleset on rec.games.abstract on 2001-12-30. From that posting:

''Alak was described as go in 1D, but that doesn't work as a serious game. So I fiddled with the rules a bit and came up with something with surprisingly complex strategy. Rules:

  1. It is illegal to place a stone where one was just captured.
  2. The game ends when the player having the move cannot move.
  3. The winner is the player with the most pieces on the board.''

Implied:

  • A player may fill in the final liberty of her own chain, making it temporarily safe.

Questions and examples

[Diagram]
Legal or illegal?  
[Diagram]
White lives ... for now  

jsha: I haven't read the original source, but from the blurb in Planiverse it sounds like filling the last liberty of your own chain is not like making two eyes. IIf you fill the last liberty of your own chain, you are alive until you capture one of the adjacent (enemy) groups. After that, your opponent may immediately play under one of the captured stones to capture your group in a snapback. (W asks: Doesn't that break rule 3?) Note there is no ko here; in order to fill your last liberty, you must make a chain of at least 2 stones.

As far as I can tell, that is entirely correct. Watch the Alak Game 1 -- Migeru
[Diagram]
White captures  
[Diagram]
White is vulnerable again  
[Diagram]
Snapback  
It seems to me that the moves above and below are a violation of rule 3 - pieces may not be played where one was just captured.Bezman
[Diagram]
White grabs one last Black stone  
[Diagram]
White makes a solid, living group  

... and then this allows White to make a chain that is not in atari, filling in the last liberty, thereby giving it life. Yes?

[Diagram]
Possible continuation. B+1  

Modulo the legality of W4, of course. -- Migeru


Scartol: Okay, I'm lost. As far as I can tell, both W4 here and on the page for AlakGame1 are suicide, and therefore illegal. What gives?

You're probably right. But apparently suicide of more than one stone is not suicide but "making life". What gives? -- Migeru

HadouKen: It looks to me like W4 is not suicide because it both creates a chain (of one stone) and deprives it of its last liberty at the same time. So it's life.

Scartol: I hope this doesn't sound crass, but I'm lost and no longer care about all this intricacy. I've got my hands full with a new teaching job. If someone else wants to take my place in AlakGame1, feel free.


Sebastian -- Scartol has a point. The justifications and patches to the rules make the game quite unelegant - much unlike Go. Maybe we can preserve simplicity when we instead add one rule that says: Instead of a normal move, a player can choose to declare one eye (or maybe one field) as safe. If it is safe, nobody can place a stone there. (This mimicks the extra effort in building a second eye in 2D (regular) Go. We may define a marker for this - similar to a special stone. However, by contrast to special stones, it is not limited in number and can not used as an advantage of one player of the other. I like the idea of a "double eye" picture. -- 2003-09-20

Traditional Alak (as close to Go as possible)

The original Go rules work fine for Alak; it remains interesting and even elegant, with familiar considerations of how close to get to a stable formation, and from which side to approach.

To be clear about how traditional rules work in 1 dimension:

  1. A player cannot fill in the last liberty of their own group.
  2. Ko: If a group of more than 1 stones was just captured, you can snapback on the next move.
  3. Passing: you may pass at any time, a pass may be forced if there are no legal moves.
  4. Threefold repetition: results in a draw.
  5. Territory: empty spaces between your own pieces (& the two edges)
  6. Winning: the player with the highest score = territory + pieces on the board + pieces captured.

The one rule that may need to be modified is how the game ends. The simplest mod I can see is: ending the game on 2 passes rather than 3. On small boards at least, this reduces the number of boards that end in back-and-forth draws, allowing a win if you end on a ko and your opponent has no valid moves.

  • The game ends when both sides pass (or a threefold repetition occurs).

SiouxDenim Rule 6 isn't a normal way to score a go position. Luckily, it doesn't change the analysis that follows very much.

Example: 6-Alak

To save vertical space, we try a board format that shows moves one line after another. For other detailed examples, see Traditional Alak.

6 is the smallest non-trivial board size. Black can ensure a win by opening on point 2; White can win if Black opens on point 3.

[Diagram]
If White is careless  

Opening on point 3. After Both pass, black wins by 1

[Diagram]
White to win  

After move 6, a is Black's best move, but she still loses by 2 points.

[Diagram]
after playing a  

Black can also pass after move 6, leading to this final position:

[Diagram]
 

Unlike the board after move 4, ko prevents Black from moving, so all pass & White wins.


If Black opens on point 2, however, she can win:

[Diagram]
Black to win  

It's not in white's interest to pass until move 8, at which point further moves only hurt their score.


Other Examples: 5-Alak to 11-Alak

see Traditional Alak.


Alak last edited by PJTraill on December 12, 2018 - 12:01
RecentChanges · StartingPoints · About
Edit page ·Search · Related · Page info · Latest diff
[Welcome to Sensei's Library!]
RecentChanges
StartingPoints
About
RandomPage
Search position
Page history
Latest page diff
Partner sites:
Go Teaching Ladder
Goproblems.com
Login / Prefs
Tools
Sensei's Library