Minesweeper patterns, explained with interactive boards

Read the shape, not just the digits

A sequence of numbers is not enough to identify a usable pattern. You must know which covered squares each clue touches. The classic straight-edge patterns below assume that the covered neighbors are exactly the squares in the single row shown above the clues, with no extra covered squares beyond the ends or on the other side.

The interactive diagrams are complete two-row mini-boards. Their outside edges close the constraints, and every displayed number is calculated from a consistent mine arrangement. On a full board, surrounding revealed cells can provide the same closure. Check those conditions before copying a solution.

Rotating or reflecting a valid pattern preserves its logic because adjacency does not change. Moving it into a wider covered region does not. An extra diagonal covered square can add another possible mine location and remove the guarantee.

The diagrams use A, B, C, and D in the explanations to name covered squares from left to right. A value of 1 means that square contains a mine; 0 means it is safe. Each clue creates an equation about its covered neighbors. You can use the equations even if you do not remember the pattern’s name.

The 1-2-1 pattern

With three covered squares directly above a closed 1-2-1 edge, the outer squares are mines and the middle square is safe. The first 1 sees A and B; the 2 sees A, B, and C; the last 1 sees B and C.

1. The left clue gives A + B = 1.
2. The middle gives A + B + C = 2. Subtract the first equation: C = 1.
3. The right clue gives B + C = 1. Since C is a mine, B = 0.
4. The left clue now requires A = 1.

The resulting top row is mine, safe, mine. Flag the outer squares and reveal the middle. Either outer flag can come first: they are both forced by the same complete set of constraints.

When the shortcut does not apply

If the first 1 also touches an extra covered square to its left, its equation becomes something like X + A + B = 1. The proof above no longer follows as written. Extra information elsewhere may still solve the position, but the three visible digits alone do not establish this pattern.

The 1-2-2-1 pattern

Four covered squares along a closed 1-2-2-1 edge have the arrangement safe, mine, mine, safe. The two middle squares are mines; the outer two are safe. This differs from 1-2-1, where the outer squares contain the mines.

1. The left 1 says A + B = 1.
2. The next 2 says A + B + C = 2, so C = 1.
3. The right 1 says C + D = 1.
4. The adjacent 2 says B + C + D = 2, so B = 1.
5. Substitute those mines into the outer 1s: A = 0 and D = 0.

Notice that the two subtraction steps prove the middle mines without any guess. The safe outer squares then follow from the satisfied 1s. If you forget the visual arrangement, reconstructing those two comparisons is enough to recover it.

Do not assume every longer sequence containing 1-2-2-1 is closed. Additional covered neighbors past an endpoint, or on the opposite side of the clues, change the equations. Count the actual covered neighbors of all four clues.

The underlying rule: subsets and differences

Pattern recognition is a convenient way to recognize repeated subset deductions. If one clue’s covered-neighbor set is contained in another’s, compare the mines required by those two sets.

When the required mine counts are equal, the extra squares in the larger set are safe. When the difference equals the number of extra squares, every extra square is a mine. When the difference is between zero and that number, you know only the total mines in the extra group.

In the four-square example below, the first 1 sees A and B. The next sees A, B, and C. Both require one mine, so C is safe. Applying the mirrored comparison at the right proves B safe. The remaining mines are at A and D.

This technique extends beyond straight rows. Neighbors can form an L shape, a corner, or an irregular frontier. What matters is set inclusion, not the appearance of the digits. Trace exactly which squares each clue sees if the geometry feels unfamiliar.

Reduce clues using proven flags

Patterns can be hidden inside larger numbers. A displayed 3 touching one proven mine still needs two mines among its other covered neighbors. For the remaining unknowns, it behaves like a 2. A displayed 2 touching one proven mine behaves like a 1.

For example, a displayed 2-3-2 sequence can reduce to 1-2-1 if each clue touches exactly one already proven mine. Those flags may be different mines or a mine shared by several clues; subtract the actual adjacent flags for each clue individually. Then check that the remaining covered-neighbor sets match the closed 1-2-1 shape.

Do not subtract the same number from every clue automatically. A nearby flag may touch the middle clue but not the outer clues. A displayed 2-3-2 is not inherently a reduced 1-2-1. The deduction depends on adjacency and correct flags, not a numeric resemblance.

A reliable checklist is: identify proven mines, subtract them from each clue, list each clue’s unflagged covered neighbors, and compare those lists with the pattern’s equations. If any flag is speculative, the reduced pattern is speculative too.

Recognize patterns that do not force a move

A 1 beside another 1 does not automatically tell you where a mine is. If both clues touch the same two unknown squares and no other information distinguishes them, all you know is that the pair contains exactly one mine.

Likewise, a clue saying C + D = 1 proves a group count but identifies neither individual square. Marking one as a mine because it “looks more likely” turns a pattern deduction into a guess. Keep those categories separate.

A position may become solvable after another part of the board opens. Treat an unresolved pair as a constraint to remember and keep scanning. The global mine count can also rule out some arrangements, so a locally ambiguous shape does not prove that the whole board requires guessing.

A compact pattern reference

This table summarizes the complete mini-boards above. It is not a rule for every occurrence of these digits on an unrestricted board. All extra neighboring covered squares must be excluded or accounted for before the reference applies.

How to practice recognition

Reset each exercise and explain the equations before clicking. Then find similar positions on Beginner or Intermediate boards. Rotate the shape mentally, check both endpoints, and subtract proven neighboring flags where necessary. Aim to recognize the proof, not merely memorize where the flags go.

What should I learn next?

Once these deductions are comfortable, practice scanning the full frontier and using the remaining mine count. For difficult positions, study valid arrangements and probability rather than accumulating a large list of unproved visual shortcuts.