Slitherlink (Loop the Loop)

Slitherlink asks you to draw one single continuous loop along the grid lines so that every numbered cell uses exactly that many of its four edges, with the path never crossing or branching. It is also known as Loop the Loop or Fences, and it is played on a grid of dots with numbers scattered in the cells between them. You connect dots only horizontally and vertically, so that the finished path forms one closed ring.

What does each number tell you?

The numbers are the whole puzzle. Each clue tells you exactly how many of its four surrounding edges the loop uses. A 3 means three of that cell’s four sides are part of the loop and one is empty; a 0 means none of its sides are used at all; a 1 means exactly one. Cells without a number can have any number of edges — they simply give you no direct information.

The single and closed loop

Two rules keep the path honest. First, the loop is single and closed: it must join up into one continuous ring with no loose ends, and you can’t finish with two separate smaller loops. Second, it never crosses or branches — the path can’t intersect itself, and no dot may sprout three or four edges. Put those together and you reach the single most useful fact in the whole game: every dot has either exactly 0 or exactly 2 edges meeting it. A dot the loop ignores has none; a dot the loop passes through has one edge coming in and one going out. A dot can never have just one edge (a dead end) or three (a branch).

Why every board has one solution

Because the loop is one unbroken ring, it always cleanly divides the grid into an inside and an outside. That is why every board is solvable by logic alone: each clue and each dot is a hard constraint, and a well-made puzzle has exactly one solution that those constraints force without any guessing. Every board on this page is generated and checked to guarantee that unique answer, so there is always a next move you can prove.

The fastest way to solve Slitherlink is to stack small, certain deductions — start with the 0s, work the 3-patterns, and apply the 0-or-2 rule at every dot — marking ruled-out edges with a small × as carefully as the edges you draw. A handful of patterns will carry you through almost every board.

The core edge patterns

• Zeros first. A 0 uses none of its four sides, so cross out all four straight away. Zeros are the best opening move on any grid because they instantly settle four edges.
• 3 next to 3. When two 3s sit side by side, the edge between them is always used, and so are the two outer edges on the far sides of the pair. That single pattern often hands you five edges at once.
• 3 in a corner. A 3 tucked into a corner of the grid forces the two edges that run along the grid’s border, because the loop has nowhere else to satisfy three sides.
• The 0–2 dot rule. Every dot must end with exactly 0 or 2 edges. If a dot already has one edge and three of its four directions are crossed off, the loop is forced down the last remaining direction to make two.
• Edges around a 1. A 1 beside a 0, or a 1 in a corner, often forbids edges immediately: if three of a 1’s sides are already crossed, the fourth has to be the single used edge — and the moment a 1 reaches its one edge, its other three sides are all crossed off.
• No dead ends, no branches. If drawing an edge would give a dot three edges, or strand a single edge with no way to continue, that edge is illegal — cross it out.
• Don’t close early. Never join the path into a small loop while clues elsewhere are still unsatisfied; the finished loop must be one ring that uses every required edge.

The winning habit: forced moves first

The winning habit is to do every forced move first — the edges and crosses you can prove with certainty — and only then look harder. Mark what you’ve ruled out as carefully as what you’ve drawn, because a crossed edge is just as much information as a used one. And never guess: on a proper Slitherlink there is always a logical next step waiting to be found.

A worked example shows the rhythm: cross out a 0’s four sides, let an adjacent 3 force its three remaining edges, then follow each forced edge from dot to dot as the loop assembles itself without a guess. Watching a few opening moves play out shows how the deductions chain together. Picture a board with a 0 somewhere in the grid. Cross out all four of its sides immediately — none of them can be used. That alone shapes its neighbours: any clue sitting next to that 0 now has one of its own sides already eliminated, so it has fewer ways to reach its number.

How a neighbouring 3 gets forced

Say a 3 sits directly beside that 0, sharing the edge you just crossed off. The 3 needs three used edges out of four, but one side is already gone — so the other three sides must all be used. Draw them in. Notice what that does to the corners of the 3: each dot where two of those new edges meet now has its two edges, which means every other direction at those dots must be crossed off to respect the 0-or-2 rule.

Following the loop forward

Follow the loop from there. One of the edges you just drew reaches a dot that previously had nothing. It now has a single edge, so it needs exactly one more — and if two of its remaining directions are already crossed, the loop is forced down the last one. Each forced edge feeds the next dot, and the path grows on its own. That is the whole rhythm of Slitherlink: settle the certainties, let each one constrain its neighbours, and the single loop assembles itself without a single guess.

This page offers Slitherlink at two sizes — 5×5 and 6×6 cells (the dot grid is one larger in each direction) — but the real difficulty lever is how many clues a board gives and how generous they are, not raw size. Both are quick to play on a phone, yet they scale in difficulty in a way that has little to do with raw size.

5×5 versus 6×6

• 5×5 is the friendlier board — a compact loop with plenty of clues, ideal for learning the patterns and for a short, satisfying solve.
• 6×6 opens up more room for the loop to wander, which means longer deduction chains and fewer easy openings. The same techniques apply; you just hold more of the grid in mind at once.

What really sets the difficulty?

The real difficulty lever is how many clues a board gives and how generous they are. A grid sprinkled with 0s and 3s offers strong footholds and solves quickly; a sparser board with more middling 1s and 2s forces you to lean on the 0-or-2 dot rule and longer chains. Tap New for a fresh puzzle at either size whenever you like, or play the shared Daily to take on the same board as everyone else that day.

Slitherlink is published in English as Loop the Loop and as Fences, and is known in Japan as Takegaki — all the same puzzle of one loop with numbered edges, so any tip carries across every name. Slitherlink travels under several names. They are all the same puzzle, so a tip you learn under one name works under every other.

It belongs to the wider family of loop puzzles, where the goal is to draw a single closed path under different rules. If you enjoy drawing one continuous loop, try Masyu, where the loop must pass through white and black pearls obeying straight-and-turn rules, or Hashi, where you connect numbered islands with bridges into one connected network. Each trains the same instinct — building a single, rule-obedient path — from a different angle, and Slitherlink’s focus on counting edges makes it a natural companion to both.

The key Slitherlink terms are clue, edge, dot, loop, cross (×) and inside/outside — together they describe a single closed ring drawn along grid lines where every dot ends with 0 or 2 edges. A few words turn up in every Slitherlink guide:

• Clue (number) — a digit in a cell saying exactly how many of that cell’s four edges the loop uses.
• Edge — one segment of the loop drawn between two neighbouring dots, along a grid line.
• Dot — a grid point where edges meet; every dot ends with 0 or 2 edges.
• Loop — the single closed ring you are building; it never crosses or branches.
• Cross (×) — a small mark on an edge you’ve ruled out, showing the loop definitely doesn’t go there.
• Inside / outside — because the loop is one closed ring, every cell ends up enclosed by the loop or outside it; thinking in terms of inside-versus-outside is a formal way to solve.

You don’t need the vocabulary to play — just tap between two dots — but it makes strategy guides far easier to follow.

Most Slitherlink mistakes come from guessing, forgetting to cross out ruled-out edges, breaking the 0-or-2 dot rule, closing the loop too soon, or ignoring the 0s — all avoidable habits. Most Slitherlink trouble comes from a few avoidable habits:

• Guessing. A drawn edge you can’t justify can lead you a long way down a false path before it contradicts a clue. There is always a forced move — find it instead.
• Forgetting to cross out. Edges you’ve ruled out are as informative as edges you draw. Mark them with × so you can spot when a dot or clue is forced.
• Breaking the 0-or-2 rule. Accidentally giving a dot one edge (a dead end) or three (a branch) is the most common slip; check each dot as you go.
• Closing the loop too soon. Joining up a small ring while clues elsewhere are unsatisfied means starting over — the finished loop must be a single ring that meets every number.
• Ignoring the 0s. A 0 settles four edges for free; skipping past one leaves easy deductions on the table.

Slitherlink is a classic Japanese pencil puzzle created and popularised by the publisher Nikoli, first appearing around 1989 and quickly becoming one of its signature titles alongside Sudoku. It is one of the classic Japanese pencil puzzles, and it took its place among the other grid puzzles in this collection.

As the puzzle spread to English-speaking solvers it picked up new names — Loop the Loop and Fences — while keeping its Japanese name Takegaki at home. What never changed is the elegant core that made it popular in the first place: a single closed loop, a handful of numbered clues, and an answer reachable by pure logic. A common formal way to solve it treats each cell as inside or outside the loop, a tidy reflection of the fact that one closed ring always splits the grid cleanly in two.