Block Fit

Five or six block shapes, a tray four squares across and five down, and exactly one way they all go in. No gaps, no overlaps, and nothing left over. The pieces never turn: the shape you see in the rack is the shape that fits, so this is a thinking puzzle rather than a fiddling one. Every tray is built by cutting a solved tray into pieces and then checking, before you ever see it, that no second arrangement works — so a perfect fit definitely exists, and there is only the one.

It runs entirely on your device: 0 network requests, works offline. No ads, no account, nothing to buy, and no betting, money or prizes of any kind.

1. Focus the board — click it, or press Tab until it is focused.
2. Look at the rack. Each piece is drawn as a miniature of its own outline, with a glyph beside it and its square count underneath. The one you are holding is outlined.
3. Press Shift to hold the next piece in the rack. Picking a piece up is free.
4. Move around the tray with the arrow keys. The dashed outline shows exactly where the held piece would land, and it carries a ✕ when that landing is illegal — so you can check before you commit.
5. Press Enter to put it down. The square under the cursor is where the piece's own first square goes — its top-left-most one — and the rest of the piece follows from there.
6. Press Enter on a piece already in the tray to take it back out, or press Esc, or click its entry in the rack. A pointer does the same job: click a square to drop, click a filled square to lift.
7. Watch the moves. Every press of Enter costs exactly one move, whether it places a piece, takes one back or bounces off. A flawless run spends one move per piece; the rest of the budget is your room to be wrong.
8. Fill all twenty squares before the moves run out.

How do you fit a set of block shapes into a tray with no gaps?

Work from the most awkward square, never from the biggest piece. Find the first empty square reading left to right and top to bottom: whichever piece ends up covering it must have its own first square sitting exactly there, because a piece's first square is its top-left-most one and there is nothing above or to the left of that gap left to cover it. That usually leaves one or two candidates instead of thirty, and once you place one the next gap is decided the same way. Keep repeating and the tray unwinds almost without a choice. The mistake that costs people the round is the opposite habit — dropping the big five-square piece somewhere roomy because it feels productive, then discovering the three-square piece has nowhere legal to go.

Is every tray guaranteed to have exactly one answer?

Yes, and both halves are settled before it is dealt. The tray is cut up first: the twenty squares are grown into connected regions and those regions become the pieces, so a perfect fit exists by construction — it is the cut we just made. Then a search fills the tray every way the pieces allow, counting to two and stopping. If it finds a second arrangement the tray is thrown away and another is cut. Two identical pieces swapping places is not counted as a second answer, because on the tray you could not tell the difference.

Why can't the pieces be rotated?

Because rotating is dexterity, not deduction. Once a piece can appear in eight orientations, the game becomes trying them all by hand, and any square you are stuck on has too many candidates to reason about. Fixing the orientation means each piece has only a handful of legal seats, which is what makes "which piece covers this gap?" a question with an answer you can work out. It also means the whole game plays from the keyboard, with no rotate-and-flip gesture to invent, and it keeps the tray small enough to verify exhaustively before you see it.

Where does the move budget come from?

From the verifier, not from a guess. The floor is one move per piece — what a run with no mistakes spends. On top of that comes a fixed allowance plus extra moves in proportion to how much work the uniqueness search had to do: a tray whose squares are nearly all forced gets less slack, a tray the search had to dig through gets more. So the budget is always achievable and always tight enough that dropping pieces at random loses.

Why does taking a piece back cost a move?

Because it is a real move, and metering it is what makes planning worth anything. Undo being free would turn the game into brute force — put everything down anywhere, take back whatever failed, repeat. Charging for it keeps the cheapest route the same one the puzzle is trying to teach: work out where a piece goes before you commit it. The landing preview is there so you never have to spend a move to find out whether something fits.

Is there any money, betting or in-app purchase in this?

None, and there never will be. No currency, nothing to buy, no lives to refill, no prize of any kind at the end.

Does it work offline?

Yes. Once the page has loaded once, the whole game is on your device — the rules, the cutter, the verifier, all of it. No tray is ever fetched and no server is involved in play at any point.

Can I play it with a keyboard or a screen reader?

Both. Arrows move around the tray, Shift changes which piece you hold, Enter commits and Esc takes back. Nothing is carried by colour: a placed piece is drawn as one joined outline, it repeats its own glyph in every square it owns, and every piece is announced in words — "4 squares in a 3 by 2 shape" — along with whether it is in the rack or on the tray. The board publishes a full parallel description, so the game is playable without seeing it at all.

Twenty squares, cut into five or six pieces of three, four or five squares each. That is deliberately small: the uniqueness check is exhaustive rather than sampled, and it has to finish instantly on a phone. Pieces are fixed in the orientation they are dealt in, and a shape made of squares joined edge to edge is called a polyomino if you want to read further. Trays that could be finished two different ways are discarded rather than shipped, so you will never meet one. Losing reveals the answer on the tray.

Colour Sort is the other tile puzzle here where the state you can see is the whole problem — no hidden information, just an ordering you have to find. Fence Loop and Light Grid are the planning puzzles: both derive their move budgets from their own solvers the way this one does, so tight is always achievable.

Bookmark this page (Ctrl+D, or ⌘D on Mac) or install the app — it works offline the next time you need it.