Puzzles Guides About Contact All puzzles →
All guides
Method Memory Match 6 min read

Memory Match without trying to remember everything

Nobody can hold twenty card positions in their head, and the good news is that nobody needs to. The trick is deciding in advance which small part of the grid you are going to be responsible for.

Level Method Puzzle Memory Match About 6 minutes

Why remembering everything fails

Working memory holds somewhere around four unrelated items at once. A twenty-card grid asks you to track twenty positions and twenty symbols, which is far outside that, so what actually happens is you remember the last three or four flips clearly and everything before that turns to mush.

That is not a personal failing and it is not fixed by concentrating harder. It is fixed by reducing how much there is to remember, which is what every technique below does.

Turns, not seconds

The number worth improving is the turn count, not the clock. A turn is a decision; a second is mostly how fast you move a mouse. Our board records both, and the clock can be switched off entirely if watching it makes you rush the decision.

Chunk the grid into quarters

Instead of one board of twenty cards, treat it as four blocks of four or five. Deal with the top-left block until you have paired what you can, then move on. Positions inside a small region are much easier to hold, because you are remembering “the star is in the corner of this block” rather than an absolute coordinate on a grid.

The same twenty cards, read as four blocks instead of one field. Every card belongs to exactly one block, so you only ever have to remember what is inside the block you are working on — plus anything you happened to see elsewhere.

The practical benefit is subtler than it sounds. When you turn over a card in one region and its partner is in another, you have learned something about a region you were not paying attention to — and because you were only tracking one small area, you have the spare capacity to notice.

The order to flip within a turn

Each turn gives you two flips, and how you spend them matters more than which cards you pick.

  1. If you know a pair, take it immediately. Knowledge decays. A pair you are sure of now is a pair you may be unsure of in three turns, and taking it costs one turn and reveals nothing to lose.
  2. If you know one card's partner is unseen, lead with the known card. Flip the one you remember first, then use the second flip to explore. If the explored card matches, you win the pair; if not, you have learned a new position for free.
  3. Never explore with both flips when you hold a known card. Two unknown flips give you two new facts but no chance of a pair. One known plus one unknown gives you a new fact and a chance.
  4. Explore with both flips only when you know nothing useful. Which is mostly the first two or three turns.
Two known pairs sitting far apart. Both should be collected before doing any more exploring — the distance between them is irrelevant, and holding them in memory while you poke around elsewhere is how they get lost.

Anchor a card to something, not to nowhere

“Third from the left, second row” is a coordinate, and coordinates are hard to keep. “The star next to the corner” is a relationship, and relationships stick.

So describe each card you see in terms of the board's landmarks: corners, edges, the middle, and cards you have already placed. A symbol touching a card you already know is nearly free to remember, because you get it back whenever you recall the first one.

The same works for pairs of symbols. If you see a star and, on the next turn, a moon directly beneath it, remember the two together as a small shape. Recalling one drags the other along with it.

Twelve, twenty and thirty cards

Twelve cards is six pairs and fits comfortably inside ordinary memory — most people clear it in nine or ten turns without any method at all. It is a good place to practise leading with a known card, because the consequences of not doing so are visible immediately.

Twenty is where chunking starts to pay. Thirty is where it becomes the only thing that works: without regions, the middle third of the game is spent re-finding cards you had already seen twice.

One last thing worth knowing: the theoretical best on a deck of n pairs is n turns, which requires knowing every position before you take anything. That is not a realistic target. A good run is closer to 1.5n, and the difference between a good and a poor run is almost entirely whether you spent flips exploring while holding pairs you had not collected.

Try twenty cards in quarters

Pick one corner and work it before looking anywhere else. Take every pair the moment you are sure of it. The turn counter is the score to beat.

Open a deck