Mahjong Solitaire strategy: how to clear every board

In Mahjong Solitaire you remove tiles in matching pairs, but you can only touch a tile that is free: nothing stacked on top of it, and at least one long side (left or right) open. Every move you make either frees new tiles or it does not. That is the lens for the entire game. A good move opens up the board. A wasteful move just trades two tiles you could already reach for nothing new underneath.

So before you click any pair, ask one question: what does taking this pair uncover? If the answer is “a tile that was stuck” or “a side that was blocked,” it is probably a good move. If the answer is “nothing, both of these were on the edge anyway,” you may be spending a pair you will wish you still had later.

The tiles that lose you games are the ones buried deepest: under the peak of the turtle, pinned in the center rows, or wedged with tiles on both sides. Those are the tiles you must dig out, and the only way to dig them out is to remove whatever is sitting on top of and beside them.

So when you have a choice, prefer the matchable pair that does the most digging. Picture two legal matches in front of you. One is a pair of tiles sitting loose at the outer edge of the board. The other is a pair that, when removed, lifts the cover off a tile in the trapped center. Take the second one almost every time. The loose edge pair is not going anywhere; it will still be free in ten moves. The buried tile might not be.

A useful instinct: glance at where the board is densest, the stacked peak and the long inner rows, and treat that region as the work that has to get done. Edges are easy and can wait. Centers are hard and set the pace. If you spend your early moves clearing the easy outside while the hard inside stays locked, you are saving the hardest problem for the moment you have the fewest options left to solve it.

The classic turtle is layered, and the upper layers cover the lower ones. A tile on a high layer blocks every tile directly beneath it. That makes the top layer the lid on the whole board: while it sits there, a big chunk of what is underneath cannot even be looked at, let alone matched.

So treat peeling the top down as a steady background goal. You are not required to clear the top before anything else, that would be rigid and often impossible, but all else being equal, prefer removing tiles that bring the stack height down. Every high tile you take exposes the ones below it and usually frees several sides at once. Removing a single top-of-the-peak tile can unstick three or four tiles in one move, where clearing a flat edge pair unsticks none.

The practical version: when you scan for your next match, look up first. If one half of a pair is on a high layer and removing it drops the lid on a section you have not been able to reach, that is worth more than a match of two tiles you could already get to anyway.

This is the single habit that separates a clear from a dead end. Before you take a pair, check what that pair is currently holding in place. Removing two tiles can open the board, but it can also free a tile whose removal then buries something you still need, or it can be the move that leaves a needed tile with no partner you can ever reach.

Here is the trap in concrete terms. Suppose you need a particular bamboo tile that is sitting open near the edge, and its only reachable match is a second bamboo two layers up. If you casually clear an unrelated pair and that move drops a fresh tile right on top of the upper bamboo, you have just sealed away the one partner your edge tile had. Both bamboos are still on the board, but they can never meet. The board is now unwinnable, and nothing flashed to warn you.

The defense is cheap: before a removal, ask whether either tile you are about to take is the thing keeping a path open. If you are removing a tile that is the last clear route to a buried partner, slow down and find a different pair first. The board rewards the player who treats every removal as also a decision about what gets covered or uncovered next.

Most tile kinds come in fours (the suit numbers, the winds, the dragons), so a match is really a choice of which two of the four to pair off. When all four are visible and free, you have a small but real decision, and beginners throw it away by grabbing the first two their eye lands on.

The principle: pair the two that you gain the most from removing, and keep the two that cost you the least to leave behind. If two of the four are buried tiles that are blocking other things, take those two, because removing them does double duty: it scores the match and it frees the board. If two of the four are harmless loose tiles on an edge, leaving those two on the board costs you nothing, they will stay free and matchable.

The mistake to avoid is matching two free tiles together and leaving the two hard ones still stuck, because now the remaining pair of that kind is the buried pair, and you have thrown away the easy partners you could have used to clear them. Said simply: spend your easy tiles to rescue your hard tiles, not the other way around. When in doubt, match the pair that includes the most blocked tile.

It feels productive to clear one corner of the board completely, but tunneling into a single area is how you run out of moves with tiles still standing. The reason is matching: a tile can only leave the board with a partner of the same kind, and if you clear one region down to nothing, the kinds that lived there are now represented only by their partners elsewhere, often in places you cannot yet reach.

Spread your progress. Work the board so that free, matchable tiles stay available in several regions at once, rather than draining one and leaving the rest locked. The healthy feeling mid-game is “I have a few different matches I could make right now,” not “I have exactly one legal move.” Breadth is safety. The more independent matches you keep live, the less likely any single forced sequence paints you into a corner.

A concrete version of this: when two matches are equally good for digging, prefer the one that keeps tiles available in more parts of the board over the one that finishes off a region. You are trying to stay flexible right up until the end, because the last dozen tiles are where boards are actually lost.

Melio gives you three tools, and they are not equal. The hint flashes a pair you can legally match right now. Use it whenever you are stuck staring at the board, because it only tells you that a move exists, it does not change the board or cost you anything structural. It is a way to find a tile you missed, not a crutch that hurts you. The same goes for undo: if a removal opened up worse than you expected, take it back and try a different pair. That is the fastest way to learn which moves help.

Shuffle is different, and you should respect it. It is there for when no legal move remains, and it rearranges the remaining tiles. The catch is that a shuffle can turn a position that was still solvable into one that is not, because it does not promise to preserve a winning arrangement, it just gives you a fresh set of pairings. So do not reach for it the moment a board feels hard. First exhaust the real options: scan every layer for a match you overlooked, use the hint, and undo back to a branch point if you suspect an earlier move was the mistake.

Only shuffle when you have genuinely confirmed there is no available pair and undo cannot recover a better line. Treated that way, shuffle is a rescue from a true dead end rather than the thing that quietly costs you a winnable board.

Put the habits together and your decision process for each move becomes short and repeatable.

1. Scan all layers, top first. Look for matches that bring the stack height down before matches that just clear flat edges.
2. Prefer the pair that frees the most-blocked tile. Spend easy tiles to rescue hard ones, not the reverse.
3. Check what the removal covers or uncovers. Do not take a pair that buries a tile you still need or strands a partner you can no longer reach.
4. Keep matches alive in several regions. Stay flexible; do not strip one area bare and lock yourself in.
5. Lean on hint and undo, hold shuffle for true dead ends. The board was solvable when it dealt; play the order and it stays that way.

None of this is fancy. It is the same five questions every time, and the more you run them, the faster the right pair jumps out at you. That is the whole game: not finding clever tricks, but refusing to make the careless move that strands a tile.