What is the probability of hitting a mine in Minesweeper?

At the start of a game, the baseline probability roughly equals the mine density — the percentage of cells that contain mines. For example, if 14% of cells are mines, any random unknown cell has about a 14% chance of being a mine. However, neighbouring number clues dramatically change these odds for specific cells.

How do you calculate mine probability from clues in Minesweeper?

When a number is revealed, it tells you exactly how many mines are in its neighbouring cells. If a "1" has three unrevealed neighbours and you know none of them are already flagged, each neighbour has roughly a 1-in-3 chance of being the mine. When two numbers share neighbours, you can subtract their constraints to narrow down further.

Why are corner and edge cells different in Minesweeper?

Corner cells have only three neighbours and edge cells have five, compared to eight for interior cells. This means they are touched by fewer number clues, which can make the probability of a mine in those positions harder to pin down from the surrounding information. On infinite boards, the concept shifts slightly since every cell is effectively in the interior.

The Math Behind Minesweeper

You do not need to be a mathematician to play Minesweeper well, but understanding a little of the underlying logic makes the game feel much more readable. The numbers on the board are not just labels — they are probability constraints, and knowing how to interpret them turns guesswork into reasoned decisions.

This page covers the key mathematical ideas in plain language, using concrete examples rather than formulas. No advanced math required.

Mine density and baseline probability

Every Minesweeper board has a mine density — the proportion of cells that contain mines. On a Normal difficulty board where roughly 14% of cells are mines, a completely unknown cell picked at random has about a 14% chance of being a mine. That is the baseline: before any clues are revealed, the probability is simply the density.

Different difficulty levels have different densities. Basic is around 10%, while Master and Hell push toward 20-25%. Higher density boards are harder not just because there are more mines, but because the probabilities shift — even cells that appear safe carry more risk, and ambiguous situations become more common.

The baseline probability matters when you are forced to guess among cells that no visible number directly constrains. If none of the surrounding clues say anything useful about a group of hidden cells, your best estimate of each one is still just the mine density. That is useful information: on a 14% density board, a guess among completely unconstrained cells has roughly a 6-in-7 chance of going well.

How clues change the odds

Once numbers start appearing, the baseline probability stops applying to cells near them. A "1" with exactly three unrevealed neighbours tells you that one of those three cells is a mine. That shifts the local probability for each of those cells from the baseline (say 14%) to about 33%. The mine is definitely nearby — the question is just which one of the three cells it is hiding in.

This is why revealed numbers are so valuable: they give you local, precise information that overrides the global density estimate. A "3" with four unrevealed neighbours tells you that three of those four are mines — a 75% probability for each one. Flagging those mines and revealing the one safe cell becomes the obvious move.

The flip side is also true. When a number has already had some of its mines accounted for by adjacent flags, the remaining constraint applies only to the remaining unrevealed neighbours. A "2" with one flagged mine already identified has only one mine left among its remaining hidden neighbours — and if there is only one remaining hidden neighbour, that cell must be a mine with certainty.

The subtraction technique

One of the most practical tools in Minesweeper is what players sometimes call the subtraction technique: comparing two adjacent numbers that share some of their hidden neighbours. When two numbers overlap in their constraints, you can subtract one from the other to derive new information that neither clue reveals on its own.

Here is a simple example. Suppose a "2" and a "1" are adjacent, and the "2" has three unrevealed neighbours while the "1" shares two of those neighbours and also has one unique neighbour. The "2" says two mines are hiding among its three neighbours. The "1" says one mine is hiding among its two shared neighbours plus the one unique neighbour. If you reason about the overlap, you can often deduce exactly where the mine in the unique position must or must not be.

This technique does not always give you a certain answer — sometimes the constraints are still compatible with multiple arrangements. But it frequently resolves situations that looked ambiguous, and it is one of the main ways experienced players avoid unnecessary guesses. The strategy guide goes into more depth on when and how to apply this approach.

Edges, corners, and neighbour counts

In a standard finite grid, cells at the edges and corners of the board have fewer neighbours than interior cells. An interior cell has up to eight neighbours. An edge cell has five. A corner cell has only three. This affects both how informative a clue is and how many clues bear on an unknown cell.

A "1" in a corner with two unrevealed neighbours is quite informative — the mine is one of two cells, giving each a 50% probability. The same "1" in the interior with seven unrevealed neighbours leaves each of those cells at only a 14% chance of being the mine — much harder to act on directly. This is why corner and edge areas of finite boards often generate easier deductions: fewer candidates means clearer constraints.

On this site's infinite board, every cell is technically in the "interior" as the board expands around you. The relevant version of this principle is about how many clues are touching an unknown cell — a cell that is adjacent to several revealed numbers is more constrained than one sitting at the current frontier with only one visible clue nearby.

Why exact calculation is usually impractical

In theory, you could treat the entire board as a system of constraints and calculate the exact probability for every unknown cell. In practice, this becomes computationally explosive very quickly. Even on a medium-sized board, the number of possible mine arrangements consistent with all the visible clues can run into millions or more. Humans cannot do this kind of exhaustive enumeration mentally.

This is related to why the general Minesweeper problem is NP-complete — the computational difficulty of exact solutions grows rapidly with board size. Even programs designed to solve Minesweeper optimally use approximations and heuristics on larger boards rather than brute-force enumeration.

The practical takeaway is that you do not need exact probabilities to play well. Local reasoning — reading the constraints near each number, applying the subtraction technique, and choosing forced moves before guessing — gets you most of the way there. For the moments when you genuinely must guess, the density estimate and the question of which cell opens the most new information are the two most useful guides. More on that in the guide to solvability and guessing.

To put these ideas into practice, read the Minesweeper Strategy guide. To understand when logic runs out and guessing becomes unavoidable, see Is Minesweeper Always Solvable Without Guessing?