Minesweeper number patterns that solve themselves
Casual minesweeper players treat each number as a fresh math problem. Expert players don’t, they recognize shapes. Three or four numbers in a row form patterns with fixed solutions, and once you can spot them instantly, you stop calculating and start sweeping. Here’s the canonical set.
Quick legend
In the diagrams below:
- Numbered cells (
1,2, etc.) are revealed and show how many mines neighbor them. - Unrevealed cells (shown as dots) are unknown.
- Flags (⚑) mark cells you’ve identified as mines.
Every pattern assumes the cells abovethe row of numbers are already revealed (otherwise you wouldn’t see the numbers). The pattern resolves the cells below.
The 1-2-1 pattern
A row of 1 2 1 along the edge of an unexplored strip is the most common auto-solve pattern in minesweeper. The two flanking 1s force a unique mine placement.
Logic: the left 1 has three unrevealed neighbors below it (positions 1, 2, 3 of the bottom row). It says exactly one of those is a mine. The middle 2 has three unrevealed neighbors too, but it says two of them are mines. The right 1 same as the left.
Together: the only way all three constraints can hold is if the mines are at positions 1 and 3, the corners. Position 2 (under the 2) is safe.
Flag the corners, reveal the middle. You’ll see this pattern dozens of times per Intermediate board.
The 1-2-2-1 pattern
The four-cell extension of 1-2-1. A row of 1 2 2 1 bordering an unrevealed strip resolves uniquely.
Logic: the 1s force only one mine in their three-cell neighborhoods. The 2s force two mines in theirs. Working through the constraints, the mines must be at the two middle positions (under the 2s), and the cells under the 1s are safe.
Flag the middle two, reveal the ends.
The 1-1 pattern (with constraint)
By itself, two adjacent 1s don’t resolve. But when the unrevealed strip below them ends, i.e. one side hits a wall or a revealed cell, they do.
Suppose the strip ends to the right (only cells 1, 2, 3 are unrevealed below; cell 4 doesn’t exist). Then:
- The left 1 sees three neighbors: cells 1, 2 (below it and the right one). Exactly one is a mine.
- The right 1 sees three neighbors: cells 1, 2, 3. Also exactly one is a mine.
- Both 1s share cells 1 and 2. If the mine for the right 1 were at cell 3, the left 1 would still need a mine at cell 1 or 2, and that mine would also count for the right 1, making it 2 mines, contradicting its value.
Conclusion: the mine is in cell 1 or 2 (shared between both 1s), and cell 3 is safe. You can’t flag yet, but you can reveal cell 3 freely.
Corner patterns
Corners constrain more aggressively than open board cells because the corner cell only has three neighbors (instead of eight). A 1 in the corner with two unrevealed neighbors is solvable immediately.
The 1 in the top-left has three neighbors. If two are already revealed, the third is the mine. If one is revealed and two are unrevealed, the mine is one of those two, but if other numbers around bound the possibilities, you can often pin it.
Rule of thumb: always scan corners first when opening. Even a single 1 there leaks more information than an interior 1.
The forced 1 (chord)
This isn’t exactly a pattern, it’s a habit. When you flag a mine, every revealed number nearby loses one unknown neighbor. If a number had already-flagged neighbors equal to its value, you can chord: click the number (or middle-click on desktop) to auto-reveal all its remaining unflagged neighbors.
The 1 here has one neighbor flagged. Its constraint is satisfied, all other neighbors are safe. Chord on the 1 and every other neighboring cell reveals at once.
On Melio Minesweeper, click any satisfied number to chord. It’s how Expert speedrunners blast through sections, flag the obvious mines first, then chord on every satisfied number for free reveals.
The 1-2 pattern (open boundary)
A 1 adjacent to a 2, both on the edge of an unrevealed strip, with the unrevealed strip continuing past the numbers, you can’t flag any mines, but you can safely reveal one cell.
Compare the 1 and the 2: they share two unrevealed neighbors. The 2 has one extra unrevealed neighbor (the one to its far right, not adjacent to the 1). If that extra neighbor was NOT a mine, the 2 would still need 2 mines from the shared cells, but the 1 says there’s only 1 mine in those same shared cells. Contradiction.
So the extra unrevealed neighbor (the far-right cell) must be a mine. Flag it. Then the 2 is satisfied with one more mine in the shared cells, and the 1 is also satisfied with one mine in the shared cells, same mine.
This pattern alone saves several seconds on every Expert board.
50/50 guesses, when nothing else works
Sometimes the board legitimately doesn’t give you enough information to deduce. Two unrevealed cells, both equally likely to be the mine. This happens, usually late game when the board is mostly cleared but a small cluster remains.
When you hit a 50/50:
- Don’t panic.First check the global mine count. If the total remaining mines plus what you’ve flagged equals the global mine count, and one specific cluster contains exactly the right number of unrevealed cells matching remaining mines, you might be able to deduce by exclusion.
- If it’s truly 50/50, pick the cell farthest from open space.If you survive, you’ll learn more from the new number than if you’d clicked into open space.
Some Expert boards have unavoidable 50/50s. World-record speedruns mostly come from boards that don’t luck shapes the absolute ceiling.
Practice progression
Don’t try to learn all the patterns at once. Pick one, play 5-10 Beginner boards specifically looking for it. Once you’re spotting it instantly, add the next.
- 1-2-1, the most common, learn first
- 1-1 with boundary, the second most common
- 1-2-2-1, appears in long strips
- 1-2 with open boundary, the harder logic
- Chording, speed multiplier once flags are reliable
By the time you can see all five patterns at a glance, Intermediate is a 1-minute board and Expert is around 2-3 minutes consistently.