Slitherlink (also known as Fences, Takegaki, Loop the Loop, Loopy, Ouroboros, Suriza and Dotty Dilemma) is a pen and paper logic puzzle developed by Nikoli. For excellent examples, see krazydad’s Slitherlink pages. There are also a variety of other puzzles, including mazes for younger kids. Slitherlink is an original puzzle of Nikoli and was first published in 1989 in Puzzle Communication Nikoli #26 (June 1989). Many varieties of the concept exist and a search will turn up many examples of related puzzles.
The puzzle is played on a grid of dots. Some of the squares formed by the dots, are called clue cells, and contain numbers. The objective is to connect adjacent dots horisontally and vertically to form a continuous loop with no loose ends (a Slitherlink). The loop is not allowed to cross itself or have any branches. The numbers inside clue cells indicate how many of its sides are part of the loop. Empty cells are not specified and can have any number of sides as part of the loop. Each puzzle has one unique solutions, that can be found without having to guess.
Puzzles using non-square grids are also available. These include Penrose, Altair, Laves, Snowflake, Honeycomb, Cairo, Pentagon and Snub Square Slitherlink variations. In theory, any tiling of the plane that fills the plane with no overlaps can be used. These add an additional dimension and complexity to the game, but similar rules apply.
Rules and clues to solve the standard square grid is widely available. This can be found on Wikipedia, Hirofumi Fujiwara’s Homepage, the parity method by Mel-o-rama. He also describes the concept of Jordan curves, that can assist in getting out of tricky situations and provide an Excel spreadsheet to assist in solving existing puzzles. Software are available and computer science projects and publications has been done on the basic puzzle.
The variations, on the other hand, has not been specifically addressed. This is an attempt to formalise a general set of rules that can be applied to any type of tiling, as long as the core rule set are used.
In these descriptions, n will refer to the number in the primary clue cell that are being looked at; a is the clue cell adjacent to the primary clue cell; bn is the number of borders of the cell in which n is located and ba is the number of borders of the adjacent clue cell. z is an additional adjacent cell.
General rule 1
For n in bn = n+1 and a = n-1 in ba = n : communal border will be part of the loop.
General rule 2
For n in bn = n+1 and a in ba = n+1 : communal border is solid; opposite borders excluding the neigbours of the communal border is part of the loop.
General rule 3
For n in bn = n+2 and n-1 corners excluded : rest of the borders is part of the loop.
General rule 3
For n in bn = n+1 and one incoming segment, the segment opposite to the incoming segment will be part of the loop.
General rule 4
This could be a general rule, but still need some further investigation. It involves a special situation with three adjacent clue cells. I noticed it on the Snub Square Slitherlinks from Crazydad with clue cells containing the values 1, 2 and 3 with 3 in the middle and 1 and 2 on opposite sides.
For n in bn = n+1 and a = n-1 in ba = n and z = n-2 in bz = n : adjacent borders of n-a and n-z is part of the loop.
General rule 5
This one is more a note to self to investigate further. It seems to be rather obvious, but it could be possible to extend it to a general rule, depending on the type of tilings that are available.
When two corner connected clue cells contain a 1 and 0, the two sides of the 1 can be eliminated from the loop. Can this be extended to higher order numbers?
TODO:
- Add graphical examples of each
- Tabulate each example
- Investigate rules 4 and 5
© september 2011 marius loots