Expert Minesweeper

One Cell in Five Is a Mine

Expert hides 99 mines under 480 cells: 20.6 percent density, a third denser than intermediate. Every characteristic of this board follows from that one figure, and the clearest way to see it is to ask how often a cell is blank. A cell shows blank only when none of its eight neighbors is a mine, which at expert density happens roughly 16 percent of the time, against about 26 percent on intermediate and around 35 percent on beginner.

Blank cells are what produce cascades, so halving their frequency more than halves the free information you get per click. Expert openings are small, the board never really opens up, and you spend almost the entire game working a wall of numbers rather than watching territory fall away. That is the actual difficulty of this level - not that the deductions are harder, but that you have to make many more of them per cell revealed.

The Long Frontier

At 30 columns wide, the boundary between revealed and hidden cells stretches further than you can hold in your attention at once. On beginner the entire board is one glance; here the frontier can be a hundred cells long and folded into a dozen pockets, and the cell that resolves the pocket you are staring at is often twenty columns away.

Work it as a sweep rather than a search. Pick an end, move along the boundary taking every forced move, ignore anything you cannot prove immediately, and start over from the same end when you reach the other side. Two or three passes usually clear positions that felt stuck on the first one. Expert also punishes cramped screens: 30 cells across is comfortable on a desktop or a tablet in landscape, and genuinely awkward on a phone held upright.

Counting Mines Is Not Optional Here

On the smaller boards, global mine counting is a convenience. On expert it is a core technique, because the local numbers frequently run out of information before the board does. Two forms of the count do most of the work. If the mines remaining exactly equal the mines a single unresolved region must contain, every hidden cell outside that region is safe. If the mines remaining exactly equal the number of hidden cells left, all of them are mines and the rest of the board is already clear.

The refined version is to split the frontier into groups, work out the minimum and maximum mines each group can hold, and check those bounds against the counter. On a 99-mine board this resolves a large share of the positions that appear to be coin flips, and it is the single biggest difference between a player who loses expert boards to guesses and one who mostly does not.

Real 50/50s and How to Take Them

Expert is dense enough that genuinely unavoidable guesses are a normal part of the level rather than an anomaly. Estimates from solver studies vary with method, but they commonly put the share of expert boards containing at least one forced guess somewhere between a fifth and a third. No amount of skill removes those, which is why even the strongest players lose boards they played perfectly.

Two rules make guessing cheaper. First, take a forced guess the moment you have proved it is forced, not after clearing the rest of the board - a coin flip that ends the game at 20 seconds costs far less than the same coin flip at three minutes. Second, when several uncertain cells are available, compute rather than feel: divide the mines that a group must contain by the number of cells in it and click the lowest number. Cells far from the frontier are often safer than the ambiguous pair you were staring at, because the counter constrains them less tightly.

Efficiency: 3BV, Chording and No-Flag Play

Once you win expert reliably, speed stops being about clicking and starts being about counting clicks. The standard measure is 3BV, the minimum number of left clicks needed to clear a given board without chording: every blank region counts as one click, and every numbered cell not touching a blank region counts as one more. Expert boards typically land somewhere around 150, with plenty of variation either side, so a three-minute win is running at less than one useful click per second and there is a lot of room in that.

Two techniques buy the time back. Chording clears every remaining neighbor of a satisfied number in one double-click, turning several clicks into one. No-flag play skips flagging entirely and tracks mines mentally, trading the certainty flags give you for the right clicks they cost. Most strong expert players use a hybrid: flags where they enable a chord, nothing where they do not. The mechanics of both are covered in the how to play guide, and the shapes worth recognizing instantly are in the pattern atlas.

Records and Realistic Targets

A first expert win commonly takes ten minutes or more, and finishing at all is the achievement. Under five minutes means your solving is sound. Under three is good, under two is strong, and getting reliably under 100 seconds requires chording, mine counting and a frontier sweep that never doubles back. The fastest verified human expert times sit around the 30-second mark, which works out at roughly five useful clicks a second sustained for the length of the board.

Because forced guesses exist here, expert is a percentage game in a way the other levels are not. Judge yourself on your best time and on how many boards you lose to something that was actually provable, not on your win rate. A player who never loses to a real 50/50 is probably not playing expert.

Where to Go From Expert

There is no harder standard board, so the ways forward are size and density. A 50x50 custom board with 450 mines keeps density slightly below expert but runs for 2,500 cells, which turns the game into an endurance test of concentration rather than deduction. Pushing density instead - a 16x30 grid with 130 mines through custom settings - produces boards where forced guesses are the norm and mine counting is the only thing keeping you alive.

The daily challenge is a different kind of test at this level: one attempt on a board everyone else is playing too, so a forced guess on day nine of a streak actually stings. Track how the times move on the best times page, and if you want a gentler board to warm up on, the 16x16 intermediate grid at 15.6 percent density is the usual choice.

Expert Minesweeper FAQ

How big is the expert Minesweeper board?

Thirty columns wide by sixteen rows tall, usually written 16x30, with 99 mines among the 480 cells. That leaves 381 safe cells to clear at a mine density of 20.6 percent.

Why are there so few openings on expert?

A cell only shows blank when none of its eight neighbors holds a mine. At 20.6 percent density that happens roughly 16 percent of the time, compared with about 35 percent on beginner. Fewer blank cells means fewer and smaller cascades, so the board reveals itself a number at a time instead of a region at a time.

How often do expert boards require a guess?

Often enough to matter. Estimates depend on the solver and on whether global mine counting is used, but they commonly place the share of expert boards with at least one unavoidable guess somewhere between a fifth and a third. Counting mines against remaining cells eliminates many positions that only look ambiguous.

What is 3BV in Minesweeper?

The minimum number of left clicks needed to clear a board without chording. Each connected blank region counts once, plus one for every numbered cell that does not touch a blank region. Expert boards commonly sit around 150, and dividing 3BV by your time gives 3BV per second, the standard measure of solving efficiency.

What is a good expert Minesweeper time?

Under five minutes means your solving is sound, under three is good, and under two is strong. Getting below 100 seconds requires chording and mine counting rather than faster clicking. The fastest verified human times are around 30 seconds.

Can expert Minesweeper be played on a phone?

It works, but 30 columns across is tight on a phone held upright. Landscape orientation or a tablet keeps the cells large enough to tap accurately, which matters more here than on the smaller boards because a misclick on a 20.6 percent density grid usually ends the game.