Learn · the Megaminx
Twelve faces, one new idea
The biggest puzzle in the sport by piece count, and most of it is the 3×3 method with more sides to walk round. The last face is the hard part.
A Megaminx is a dodecahedron: twelve pentagonal faces, each turning by a fifth instead of a quarter. It carries 20 corners, 30 edges and 12 fixed centres, 132 stickers in all against a 3×3's fifty-four.
Almost none of that extra size costs you anything. The centres never move, exactly as on a 3×3, so the colour scheme is fixed and there is nothing to establish. Corners still have three stickers and edges two. Every instinct you have for building a layer transfers directly. The new part is that there are more places to look.
A first Megaminx solve takes a long time, and almost nothing surprises you until the very end.
20 pieces
Corners
Three stickers each, one at every vertex where three faces meet. They behave as they do on a cube.
30 pieces
Edges
Two stickers each, one along every join between two faces.
12 pieces
Fixed centres
One a face and none of them can travel, so white is opposite the same colour on every Megaminx ever made.
Reading the diagrams
A dodecahedron cannot be drawn flat without distorting it, so the diagrams here show one face and the five around it, laid out the way a net folds. The middle pentagon is the face you are working on; the five around it are its neighbours, and their far sides are cut off because nothing you do to the last face touches them.
Each face is eleven stickers: a pentagon in the middle, five edges against its sides, and five corners at its points. Only the centre sticker is a pentagon. The other ten are four-sided.
Getting to the last face
One. The star, then the first face. Pick a colour, put its five edges around its centre, then bring the five corners in underneath them. This is the cross and first layer of a 3×3 with one extra piece in each set, and you do it by looking, without algorithms.
Two. Work down the sides. Turn the puzzle over so the finished face is at the bottom and fill in the ring of faces around it, then the next ring. Each piece goes in the way a 3×3's middle layer does: bring it above its slot, and use a pair of turns that drops it in without disturbing what is already built. There are more slots than on a cube. The method is the one you already know.
Three. Stop when one face is left. Nothing up to this point needs memorising. What follows does, and it comes to four algorithms, each doing exactly one job.
The last face: four algorithms
each found by searching the puzzleThe last face is five corners and five edges, and each of them can be in the wrong place or turned the wrong way round. That is four separate problems and four algorithms. Do them in this order: place the corners, place the edges, then turn whatever faces the wrong way.
Every one of these leaves the rest of the puzzle exactly as it was. That is what makes them safe to use in any order.
Three corners round. R U R' U2 R' U2' R2 U' R2' U2' R U2
Cycles three corners and touches nothing else. No piece is turned, so if a corner arrives in the right place it also arrives the right way round. Play it three times and you are back where you started, which makes it safe to try.
Any arrangement of the five corners that can occur is reachable by repeating this and turning the top face between goes, and it never takes more than 3 plays. Only half the arrangements can occur at all, because every face turn moves five pieces round in one cycle, and a five-cycle is an even rearrangement.
Three edges round. U2 R' U' R U' R U R2' U R U' R U' R'
The same job for the edges: three of them move, nothing turns, nothing else on the puzzle shifts. Again three plays returns it, and again 3 plays is the worst any arrangement needs.
Two corners turned. R U R' U R U2' R' U2 R' U' R U' R' U2 R U2'
Once the corners are in the right places, some may still be facing the wrong way. This turns two of them where they stand, one a third of a turn one way and the other a third back, and moves nothing at all. It has to work in opposite directions, because the turns of the five corners always add up to nothing.
At most 4 plays settles every case, and there are eighty-one of them.
Two edges turned. R U F U' F' U' F' U F U R' F' U2' F U
The last one, and the only one that needs a third face. Two edges turn round on the spot and nothing else moves. It undoes itself, so a wrong guess costs nothing but the time.
Sixteen cases, and at most 2 plays.
Those four finish the puzzle from any position it can reach. This is not the quickest way round it. Real Megaminx solvers know far more algorithms and use them to avoid repeating themselves, in the same way a 3×3 solver graduates from one last-layer algorithm to fifty-seven.
Getting faster is a matter of pausing less, here as everywhere else. The coaching pages apply unchanged.
The engine behind this page is built from the solid rather than from a move table: the dodecahedron is derived first, and a face turn is a seventy-two degree rotation about that face's normal. It is then modelled twice, points rotated in space against compiled permutations, and the two must agree or the build fails. The four algorithms came out of a search: sequences were enumerated by machine and only those leaving every piece below the top face untouched were kept, and each one's effect is re-measured before publishing. The players replay those same permutations, so the animation cannot contradict the proof. What is NOT proved is anything about the earlier layers, which is why the page calls them intuitive and promises nothing about them.