How to Solve a Sliding Puzzle (the 15-Puzzle Method)
A scrambled sliding puzzle looks hopeless at first - tap a tile, three others shuffle, and the picture seems to scatter every time you touch it. The trick is that you never have to solve the whole board at once. There is a clean, repeatable method, the same one speed-solvers use on the classic 15-puzzle, that lets you lock pieces into place one section at a time and never undo your own work. This guide walks you through the whole method, including the one move that trips everyone up: the last two tiles of a row.
How a sliding puzzle actually works
A sliding puzzle is a square grid of tiles with exactly one empty space. A classic 15-puzzle is 4×4 - fifteen numbered tiles plus the gap; a picture version cuts an illustration into the same grid. There is only one kind of move, and the whole game flows from it:
You can only slide a tile that sits next to the empty space, and it slides into that space - leaving a new gap behind it.
That single constraint is why the puzzle feels slippery: the only way to move any tile is to walk the empty space around the board and let it pull the tile you want along the way. Everything below is really one skill - steering the blank - applied with a plan instead of at random.
Learn on a small board first. A 3×3 grid (only eight tiles) teaches the exact same instincts as a big one but forgives mistakes far faster, because the final stretch is tiny. Get comfortable solving 3×3 by feel before you take on 4×4 and up. Mochi Slide opens on gentle 3×3 boards and only steps up to 4×4 and 5×5 once the moves feel natural, so the jump never catches you out.
The big idea: shrink the board layer by layer
A full board feels impossible because you are trying to hold every tile in your head at once. The method fixes that by carving the puzzle into pieces you finish and then never touch again:
- Solve the entire top row, left to right. Once it is correct, you will never need to disturb it.
- Solve the entire left column, top to bottom. With the top row and left column done, you have effectively peeled an L-shape off the board.
- What is left is a smaller sub-puzzle. A 4×4 has become a 3×3; a 5×5 has become a 4×4. Treat it as a brand-new, smaller board and repeat steps 1 and 2.
- Keep shrinking until only a 2×2 corner remains. That final little square is the one part you solve differently - by rotation, covered below.
That is the whole strategy. Every row and column you complete makes the remaining puzzle genuinely smaller, so it gets easier as you go. The discipline is simple: once a row or column is solved, keep the empty space out of it.
Placing a single tile
Most tiles are placed the same easy way. To get a tile to its home square:
- Find the tile you want and the square it belongs in.
- Steer the empty space around - sliding other tiles, not yet the one you care about - until the blank sits next to your target tile, on the side you want to push it from.
- Nudge the tile one step toward its home, then walk the blank around to the next side and nudge again.
- Repeat until the tile sits in its home square, then leave it and move on.
The mental shift that makes this click: you are not "moving tiles," you are driving the empty space in loops around the tile you want to nudge. Once that is automatic, placing the first several tiles of any row is straightforward.
The tricky part: the last two tiles of a row
Here is where almost everyone gets stuck, and where the method earns its keep. Place the last two tiles of a row the obvious way and putting the second-to-last in its corner forces the final tile out, with no room left to slip it in. The fix is a small setup trick:
- Do not place the last two tiles in their final squares. Instead, stack them out of order in the corner. Put the tile that belongs at the end of the row into the corner square itself, and the second-to-last tile directly beneath it.
- Now rotate them in. With both tiles parked like that, a short cycle of moves rolls the corner tile to the right and drops the one beneath it up into place - and both land correctly at once.
- Think of it as a hinge. The two tiles pivot into their slots together, so neither one knocks the other out.
The same rule applies to the bottom two tiles of a left column: stack them out of order against the edge, then rotate them home together. Learn this one move and the whole layered method becomes reliable, because the end of every row and column is the only place it ever gets fiddly.
Why "out of order"? A sliding puzzle has tight corners with no spare room, so two finished tiles can't both slide into adjacent edge squares one after another - the first one blocks the second. Parking them as a stacked pair gives the empty space the loop it needs to swing both in on a single rotation. If a corner ever feels like it is "fighting back," it almost always means you tried to place those last two straight instead of as a pair.
Finishing with the final 2×2
Once the layers have peeled away and only a 2×2 block in one corner is left, you are nearly done. Three of those four squares are tiles and one is the gap, and there is only one thing to do:
- Just rotate. Slide the tiles around the little 2×2 loop - clockwise or counter-clockwise - and watch each one pass through its correct square.
- Stop when the picture lines up. Within a couple of turns, all three tiles click into their homes and the gap settles into its corner. The board is solved.
- If it won't resolve, look up. On a true picture puzzle, every board is solvable, so a 2×2 that won't finish means a tile above or to the left is in the wrong place - back up one layer and re-check it.
The full routine, start to finish
Put it together and you have a method that solves any sliding picture puzzle the same way every time:
- Top row, left to right. Place each tile; handle the last two as a stacked, out-of-order pair and rotate them in together.
- Left column, top to bottom. Same idea; the bottom two go in as a pair against the edge.
- Repeat on the smaller board. The remaining sub-grid is one size smaller - solve its top row and left column the same way.
- Keep shrinking until a single 2×2 corner is left.
- Rotate the final 2×2 until the whole picture snaps together.
You only ever think about one small region at a time, and once a region is done it stays done. That is what turns a wall of scrambled tiles into a calm, step-by-step solve.
Common beginner mistakes
- Solving the whole board at random. Tapping tiles toward "looking more correct" undoes itself constantly. Commit to finishing the top row first and the rest follows.
- Disturbing a solved row. Once a row or column is right, keep the empty space out of it. If you must route the blank past, go around, not through.
- Forcing the last two tiles straight in. This is the number-one cause of a "stuck" corner. Always stack them out of order and rotate the pair home.
- Skipping the small boards. Jumping straight to 5×5 before 3×3 feels automatic is a fast way to get frustrated. The small grids teach the exact same moves with a far shorter endgame.
- Losing track of the target. On a picture puzzle the slices can look alike. Glance at the finished image before you commit to where a tile belongs.
Where to practise
The method only sinks in by doing it, and a picture puzzle is the friendliest place to learn, because the image itself tells you whether a tile is home. Mochi Slide is a sliding picture puzzle built exactly for this: tap a tile in line with the gap and the whole run slides over as one move, or swipe and use the arrow keys to walk the empty space yourself. It starts on small 3×3 boards so the layered method and the last-two-tiles rotation become second nature, then steps up to 4×4 and 5×5 across five hand-drawn worlds, and keeps generating endless larger boards once you have cleared the set. Every board is scrambled with legal slides only, so it is always solvable, and a hold-to-peek button shows the finished picture whenever you lose your bearings.