Minesweeper opening patterns

Written and play-tested by the Budi Games team · Updated August 2026

Most Minesweeper advice starts in the middle of a board, with a wall of numbers already on screen. This one starts at the beginning: the first click, the shape of the region it opens, and the handful of readings that turn that region's edge into a list of squares you can open without risk. The opening is where boards are won and lost, and it is the part almost nobody explains.

Everything below is drawn out square by square. If you want to follow along on a live board, the Minesweeper page deals a new one in a click.

The first click cannot lose

On our board the mines do not exist when the page loads. Nothing is decided until you click, and at that moment the game removes the square you clicked and its eight neighbors from the pool before scattering the mines through everything that is left.

Two consequences follow, and the second is the useful one. First, the opening click is always safe — losing on move one is impossible here. Second, because none of the eight surrounding squares can be a mine either, your first square always shows a blank rather than a number, and a blank always opens outward until it runs into numbers. Your first click therefore never gives you one square. It gives you a region with a numbered border, which is exactly the raw material the rest of this guide works on.

That also settles a question beginners ask: where should the first click go? Anywhere, as far as safety is concerned. In terms of information, a click well inside the board tends to open a larger area than one wedged into a corner, simply because the cascade has room to spread in every direction. It is a mild preference, not a rule.

One thing worth knowing before you settle in: the board itself is not saved. Only your best time per difficulty is kept, in your own browser, so reloading the page mid-game means starting a new board rather than resuming the old one.

The three boards

There are three sizes, and the difference between them is not only area — it is mine density, which is what decides how often the numbers on your border are readable.

At Beginner density most borders resolve completely; at Expert density they frequently do not, and part of playing the big board well is recognizing quickly which sections are worth reading and which are not.

Reading the border of an opened region

An opened region has a frontier: opened numbered squares on the inside, covered squares on the outside. Every number on that frontier is a statement about the covered squares touching it, and only about those. A 3 says "exactly three of the covered squares I touch are mines." It says nothing about anywhere else on the board, ever.

Two readings come straight out of that, and between them they clear most of a Beginner board.

• The number is already used up. If a 2 touches two flags and three other covered squares, both of its mines are accounted for. All three of the others are safe and can be opened.
• The number has no room. If a 3 touches exactly three covered squares, all three are mines. There is nowhere else for them to be.

Neither reading is clever, and that is the point: they are the two things to check on every number before you look for anything more interesting. The habit that makes them pay is going back. Every time you open a square or place a flag, re-read the two or three numbers next to it, because a single new fact often satisfies several numbers at once and each of those hands you more squares.

In the diagrams that follow we use a plain notation:

   #   covered square, unknown
   ·   opened square with no mines touching it
   1   opened square touching exactly one mine
   F   your flag
   ║   the edge of the board

The 1-1 along an edge

Here is the most common readable shape in an opening. The left wall is the board's edge, the top two rows are already open, and a row of covered squares sits underneath the numbers. Columns are labeled a through d for reference; everything to the right of d is off-frame and plays no part in the argument.

        a   b   c   d
   ║    ·   ·   ·   ·    …
   ║    1   1   2   1    …
   ║    #   #   #   #    …

Work with the leftmost 1 first. Because the wall is on its left, the only covered squares it touches are a and b. So exactly one of a and b is a mine — we have no idea which.

Now the second 1, above b. It touches a, b and c. It also needs exactly one mine, and we already know one is sitting in a or b. Whichever way that resolves, the second 1's quota is filled by a square it shares with the first 1. Square c is therefore safe, and you can open it without thinking about it further.

Note what did the work here: the wall. Without it the leftmost 1 would also touch a covered square to its lower-left, its mine could be over there instead, and the whole deduction would collapse. This is why the pattern is described as happening along an edge — the same three numbers in the middle of the board, with covered squares on both sides, tell you nothing at all.

The frame above happens to resolve completely, and it is worth finishing because that is what a good border reading feels like. Look at the 2 above c: it touches b, c and d, and needs two mines. We just proved c is safe, so the mines must be b and d. Flag both. Now return to the leftmost 1: its one mine is b, so a is safe too. From two 1s and a 2 we got two flags and two free clicks.

The 1-2 version

Same wall, same corridor, but the second number is a 2 rather than a 1. One mine lives in a or b, and the 2 above b needs two mines among a, b and c — so the shared pair supplies one and c must supply the other. Flag c. This is the mirror image of the 1-1 and it comes up just as often; the two together will carry you through most edges.

The 1-2-1 corridor

The 1-2-1 needs a stricter setup, so read the condition before the pattern: three covered squares in a row, with the numbers 1, 2 and 1 sitting above them, and those three covered squares are the only covered squares any of the three numbers touch. In practice that means the surrounding squares on both sides are already open.

        a   b   c
        1   2   1
        #   #   #
   ═══════════════   edge of the board

Rather than reasoning about it, list the possibilities. The 2 in the middle touches all three covered squares and needs exactly two mines, so the mines are one of three pairs: {a, b}, {a, c} or {b, c}.

• The 1 on the left touches a and b, and needs exactly one mine, so it kills {a, b} — that pair would give it two.
• The 1 on the right touches b and c, so by the same argument it kills {b, c}.
• One pair survives: {a, c}.

So a and c are mines, b is safe, and none of it required a leap. Flag the two ends, open the middle. The related 1-2-2-1 corridor works out the same way with four covered squares — the mines land under the two 2s and the squares under the 1s are safe — and you can confirm it by listing pairs exactly as above.

Do not memorize these as pictures. Memorize the method: when a small group of covered squares is surrounded by numbers, write out the arrangements that satisfy one number and cross off the ones that break another. It works when the corridor is vertical, when it is rotated, and when the numbers are 2-3-2 instead of 1-2-1, which no list of named patterns will ever cover.

Deducible squares and squares that need a choice

There is a hard line in Minesweeper that is worth naming, because beginners often blame themselves for crossing it. A square is deducible when every arrangement of mines consistent with the visible numbers agrees about it. A square requires a choice when two consistent arrangements disagree — one puts a mine there, the other does not — and in that case no amount of staring will help. The information is not on the board.

The classic version is a pocket of two. Two covered squares sit against the bottom edge, the squares on either side of them are already open, and both numbers above them read 1:

        p   q
        1   1
        #   #
   ═══════════════   edge of the board

Each 1 touches both p and q and nothing else that is still covered, so there is exactly one mine in the pair and both numbers say the same thing about it. Mine at p, or mine at q: both arrangements fit every number on the board. This is a coin flip, and recognizing one quickly is a skill in itself — the time worth spending on a genuine 50/50 is zero seconds.

Before accepting one, though, check two things. Is there really no other number touching p or q, further around the frontier? And what does the mine counter say? A local tie can sometimes be broken by the global count: if the counter shows one mine remaining and there are covered squares elsewhere on the board, the mine is not necessarily in this pocket at all, and one of the other regions may be the safer click.

Flags and the counter

Flagging is bookkeeping, and it is optional. You can place a flag by right-clicking a square, by pressing F while it is focused, by holding your finger on it for about half a second on a touch screen, or by switching the Flag mode button on so that ordinary taps place flags instead of opening squares.

The counter at the top of the board shows the total mine count minus the number of flags you have placed. Three things follow from that definition, and all three matter:

• It counts your flags, not correct flags. The board never tells you a flag is misplaced. If you flag the wrong square, the counter keeps going down and will happily reach zero while a real mine is still uncovered.
• Zero is a deduction, conditional on your flags. When the counter reads zero and you are confident every flag is right, every remaining covered square is safe. That is often the fastest way to finish a board — the last dozen squares open in a few seconds.
• A negative number is an alarm. The counter can go below zero, and if it does you have placed more flags than the board has mines. Somewhere behind you is a square you flagged on a hunch.

Two more details about how this board behaves. Flags are not required to win: the game ends the moment every square that is not a mine has been opened, so a board cleared without a single flag counts exactly the same, and any unflagged mines are marked for you at the end. And there is no chord click here — no double-click that opens all the neighbors of a satisfied number at once — so each safe square is opened individually. Flagging is worth the extra clicks when it stops you losing your place on a crowded border, and not worth it when you can already see the whole region resolving.

A workable opening routine

Putting it together, the first minute of a board looks like this. Click somewhere with space around it and let the cascade run. Walk the frontier of the opened region once, applying only the two basic readings, and open everything they hand you. When the easy readings dry up, look for edges — the board's own walls and the walls of your opened region — because that is where 1-1 and 1-2 live. Only then start listing arrangements for the awkward pockets.

And when a pocket turns out to be a genuine choice, take it and move on. A board that ends on a coin flip after twenty clean deductions was played well; the deductions are the game, and the flip is just the weather.

Questions players ask