How to play memory match
All cards start face down. Flip one, then another. If they match, they stay revealed. If they do not, both flip back after a moment and you try again. The game ends when every pair has been found.
Four grid sizes are available, from six pairs for a thirty-second round to eighteen pairs for something genuinely demanding. The stats track moves, elapsed time and accuracy — accuracy being the percentage of your flips that produced a match.
The strategy is real, even though the game looks like luck
Memory match is often dismissed as a children's game of pure chance. It is not. There is a clear optimal policy, and the gap between playing well and playing carelessly is large.
Always flip an unknown card first. This is the single most important rule and the one most people break. If you already know where a 🍎 is and you flip a new card that turns out to be a 🍎, you win the pair immediately. If instead you flip the known 🍎 first and then a random card, you only match by luck. Leading with the unknown card converts knowledge into pairs.
Every flip is information, even a failure. A pair that does not match still teaches you two card positions. The players who lose badly are the ones who flip quickly without registering what they saw. Take the extra half-second to place each card in memory before it turns back.
Use spatial anchoring. Do not try to memorise "there is a strawberry somewhere on the right". Anchor to grid positions: "strawberry, second row, third column". Human spatial memory is substantially stronger than free recall, which is the principle behind the method of loci that memory competitors have used since antiquity.
Work in regions. On the larger grids, explore one quadrant thoroughly before moving on. Learning four cards in one corner is more useful than four scattered across the board, because adjacent knowledge compounds — the more you know about one region, the higher the chance any new card there completes a pair.
What a good score looks like
With perfect memory, an eight-pair game takes at most twelve moves: at worst you reveal every card once, learning the whole board, then clear all eight pairs. The theoretical floor is eight moves, which requires every single flip to hit — achievable only with luck.
Realistic benchmarks: on the 4×4 grid, under 14 moves is excellent, 15–20 is solid, and above 25 means you are flipping faster than you are remembering. On the 6×6 expert grid, under 30 moves is a strong result. Accuracy above 60% indicates you are genuinely tracking the board rather than guessing.
What it actually exercises
The game loads visual-spatial working memory — the system that holds "what was where" for the few seconds to minutes you need it. That capacity is famously limited, around four items for most adults, which is why the game gets hard so quickly as the grid grows.
A caveat worth stating plainly, since brain-training claims are everywhere: the evidence that practising working-memory games improves general cognitive ability is weak. What does reliably improve is your performance at this specific task, through better strategy and better encoding habits. That is still worth something — the encoding habits transfer to remembering where you put things — but it will not raise your IQ, and anyone selling you a game that promises otherwise is overstating the research.
Origins
The game is usually called Concentration, Pelmanism or Pairs. Pelmanism takes its name from the Pelman Institute, a British memory-training organisation founded in the 1890s that popularised the card game as an exercise. Concentration became the American name after a long-running television game show of that name, which first aired in 1958 and ran in various forms into the 1990s.
It has survived because it needs nothing: a deck of cards, no reading ability, no shared language, and it works for two players or one. That is also why it is a fixture of language classrooms — pairing a word with its picture turns vocabulary drilling into a game with no instruction overhead.