Meowdoku Playbook
Everything this site knows about the cat logic puzzle, arranged in the order you would learn it — what the puzzle is, then the tactics, then the measurements underneath them. 9 articles so far.
Last updated
What the puzzle is, and the deduction that solves any board.
What a cat logic puzzle actually is, the five tactics that solve real boards, and one 5×5 board taken from an empty grid to a finished solution — with the row, column or region that forces every single move. No step in it is a guess, and the build refuses to publish the page if one ever becomes one.
See the shared rules, the important play-mode differences, and the solving techniques that transfer between Meowdoku and Queens.
One technique per article, in the order each becomes useful.
Two or three cats are decided before you touch a new board. After that, an open region is eighteen times likelier than an open row to be sitting on its last legal square — measured across four hundred boards per difficulty, with the arithmetic for why.
Regions produce most single-candidate deductions from start to finish, and rows and columns are not just rarer — they cluster hard at the endgame. Measured across four hundred boards per difficulty, tracking every forced cat from the opening to the point elimination stalls.
The rules call it an eight-neighbor check. Because a solved board has one cat per row, only neighboring rows can ever touch — so the whole rule reduces to one number: a column gap of at least 2, measured across four hundred boards per difficulty.
A move can satisfy every stated rule and still be wrong if it leaves the board unsolvable. A wrong ✕ mark costs nothing on the spot and quietly corrupts everything after it. Ten real mistakes, each grounded in something the game actually checks.
The validator checks whether the whole board can still be solved, not just whether your move was legal — which means an accepted move is guaranteed correct. Measuring how narrow a real stall actually is, and why two moves is the right amount of guessing.
What hundreds of generated boards actually show, measured rather than asserted.
A cell's fill color is the only thing marking its region — no border backs it up. The palette runs real graph coloring, but simulating red-green color blindness collapses one recurring pair on 95%+ of boards at every difficulty.
Locating a square board in a cluttered phone screenshot, counting the grid from its edges, sampling each cell on a ring so the cat in the middle cannot spoil the color, and clustering into exactly N regions — because the rules already tell us how many there are.
These are the long pieces: the ones that explain why a move is forced rather than telling you where the cats go. They are separate from How to Play, which answers "what are the rules". An article lands here when the question it answers takes more than a page of rules to cover.
Every board printed in an article is checked by the build before the page ships. The solution has to obey all three rules, the layout has to have exactly one solution, and each step of a walkthrough has to be genuinely forced at the point it is claimed. If any of that stops being true, the build fails instead of publishing a guide that quietly contains a guess.
Nothing here is sponsored, and the articles carry no ads — what pays for the site is a single advert on the pages where you play. No advertiser has ever seen an article before you did, or had any say in one. This is an independent fan site — the about page explains who runs it and how it is funded.
Spotted an error, or want a topic covered? The contact page reaches the person who writes these.
