Game Guide
A level by level walkthrough of the free Memory Match browser demo, from the 4 pair grid to the 12 pair board, plus where to watch it played.
Memory Match is the game in MyRin that looks like a warm up and is not one. Four pairs is trivial. Twelve pairs, on a six column grid, asks you to hold twenty four positions in working memory while the board keeps changing state under you. The jump between those two boards is the whole game, and it happens inside the five levels you can play for free in a browser tab.
This walkthrough goes level by level through that free demo, with the actual card counts, the actual grid widths and a move budget for each board. If you would rather see a round play out before you read about it, start with the clips.
Here is the MyRin short for Memory Match. A twenty second clip shows the flip timing and the grid growth faster than any paragraph can.
The motifs are not generic icons. Memory Match in MyRin descends from kai-awase, the Heian court shell matching game: clam shells were split in half, both halves painted with the same scene, and only the true pair fit back together. The cards here carry the motif family those shells actually wore, sakura, tanabata bamboo, tea, swan, wave, mountain, wind chime, maple, and on the deeper boards a bell, a torii gate and a moon.
Mechanically the browser demo is strict and simple. Every card starts face down. You tap one, you tap a second, and the board locks for a moment while it resolves. A pair stays open. A non-pair turns back over. Each two card attempt counts as one move, and the move counter is the only score that matters.
Eight cards in a four wide grid, so two rows. There is no reveal phase, which means your first two flips are pure information gathering and cannot be anything else. Expect to spend three to five moves and treat this level as calibration: it tells you how the flip delay feels and how long the board stays locked.
The one habit worth installing here, because it costs nothing on a board this small and pays on every later board, is naming the position out loud or in your head as the card turns: not "there is a wave" but "wave, top row, second from the left". Position first, motif second.
Twelve cards, three rows of four. This is the first board where a bad first move genuinely costs you, because you can now reveal a card, fail to match it, and lose track of where it was.
Use the first flip of every move as reconnaissance. Flip a card you have never seen. If it does not match your second flip, you have still bought a new coordinate for the price of a move you were going to spend anyway. Six or seven moves is a good result here.
Sixteen cards, four rows of four, and the point at which most people stop being able to hold the board as a flat list. Sixteen positions is above the working memory span of almost everyone if you try to remember them as items in a row.
Switch to quadrants. The grid is four by four, so split it into four two by two blocks and treat each as a slot: top left, top right, bottom left, bottom right. When a card turns over you file it in a block rather than at a coordinate. Four blocks holding two to four motifs each is well inside span, and when you need a match you search one block instead of the whole board.
Twenty cards, and the grid widens to five columns for the first time, which quietly breaks the quadrant trick because five does not divide evenly in two. Two new motifs also enter the deck here, a herb sprig and a bell, so the set you are discriminating between has grown as well.
The fix is to stop splitting by geometry and split by column instead: five columns, four cards deep, five bins. Column bins survive a change in grid width, which matters because the installed game varies board shape far more than the demo does. Ten to fourteen moves is strong on this board.
Twenty four cards in a six wide, four deep grid, with the torii and the moon added to the deck. This is the board the demo exists for. Twelve moves is the floor, and it is unreachable in practice. Under twenty two means you were genuinely mapping rather than probing.
Three things change at this size. First, the cost of a wasted move rises, because each failed attempt is a smaller fraction of the information you need but the same fraction of your patience. Second, partial knowledge becomes the dominant state: you will spend most of the board knowing where one half of a pair is and not the other, and the right move is almost always to lead with the known card and probe with the unknown one. Third, order starts to matter: clear the pairs you are certain about early, because every cleared pair removes two cards from the field you have to keep scanning.
The reason the "position first" habit works is not a game trick. Spatial memory and object memory are supported differently, and location information tends to be retained more reliably than the identity of an item at that location. When you encode "wave, top row, second from the left" you are attaching the fragile part, the motif, to the durable part, the place. When you encode "there is a wave somewhere" you have stored the fragile half on its own.
This is also why saying both positions aloud after a failed attempt helps so much more than it should. You are adding a second encoding channel, verbal alongside visual, to information that is about to disappear when the cards turn back over.
The honest claim is narrow. Memory Match loads visuospatial working memory and it loads it hard, especially on the last two boards, where the number of positions to maintain exceeds what you can hold without a chunking strategy. What it trains most directly is the strategy layer: chunking, rehearsal and deciding which flip buys the most information. Those transfer to anything with the same shape, which is most of the rest of the MyRin memory category.
What it will not do is raise your memory span in general. Do not expect the board to make you better at names or at where you left your keys. Expect it to make you noticeably better at Memory Match, and to make the chunking habit automatic enough that you reach for it elsewhere.
If the demo goes quickly, the same five boards exist in the installed game as the opening of a 600 level run, which shifts the difficulty curve from grid size to grid size plus time pressure. The demo is the honest version of the first ten minutes.