The ability of chess grandmasters to memorize complex board positions after just a brief glance is one of the most famous phenomena in cognitive psychology. For decades, it was assumed that these masters simply possessed superhuman, "photographic" memories.
However, psychological studies—most notably by Adriaan de Groot in the 1940s, and later by William Chase and Herbert Simon in the 1970s—revealed a fascinating caveat: grandmasters can only remember board positions that could logically occur in a real game. If the pieces are placed randomly, the grandmaster’s memory is barely better than that of a complete beginner.
This paradox is explained by a cognitive mechanism known as chunked pattern recognition, combined with the use of long-term working memory. Here is a detailed breakdown of how this cognitive process works.
1. The Limits of Short-Term Memory
To understand the chess master's brain, we must first understand human memory limits. The average human short-term (or working) memory can hold roughly 7 (plus or minus 2) items at a time.
If a novice looks at a chessboard with 25 pieces on it, their brain tries to remember 25 distinct data points (e.g., "White pawn on e4," "Black knight on c6"). Because 25 far exceeds the capacity of short-term memory, the novice will only accurately recall about 4 or 5 pieces before their memory fails.
2. The Solution: "Chunking"
"Chunking" is a cognitive process where the brain groups individual, disjointed pieces of information into larger, meaningful wholes (chunks).
Think of reading: you do not read this sentence by consciously looking at every individual letter (T-H-I-S). Your brain recognizes the chunk "THIS" as a single concept.
In chess, grandmasters do the exact same thing. Over thousands of hours of study and play, they have built up a mental library of chess patterns. When a master looks at a board, they do not see 25 individual pieces. They see 3 or 4 meaningful "chunks." For example: * A "fianchettoed kingside defense" (which accounts for a king, a rook, a bishop, and three pawns). * A "minority attack pawn structure." * A specific grouping of attacking pieces aiming at a weak square.
Because the master's brain groups these 25 pieces into just 3 or 4 familiar chunks, the information fits perfectly within the limits of human short-term memory.
3. Long-Term Working Memory and Template Theory
Cognitive scientists Anders Ericsson and Walter Kintsch expanded on this by proposing the concept of Long-Term Working Memory (LTWM).
Through practice, experts develop "templates" in their long-term memory. It is estimated that a chess grandmaster has between 50,000 and 100,000 of these chess patterns stored in their long-term memory.
When a master glances at a board, their visual cortex rapidly scans the position and instantly matches it to a template stored in long-term memory. They are not actually memorizing the board in that five-second glance; rather, they are using the visual cue to retrieve a pre-existing memory. Once the overarching template is retrieved, the master only has to use short-term memory to note the slight deviations (e.g., "It's the standard Sicilian Dragon structure, but the rook is on c8 instead of c7").
4. Why the Mechanism Fails with Random Positions
This chunking mechanism perfectly explains why grandmasters fail miserably at recalling random piece arrangements.
When researchers place pieces on the board randomly—putting pawns on the back row, placing bishops in impossible clusters, and creating structures that defy the rules and logic of chess strategy—they completely bypass the grandmaster's mental library.
When the master looks at a random board: 1. No Patterns Exist: The visual input does not match any of the 100,000 templates stored in their long-term memory. 2. Chunking is Impossible: Because the pieces have no logical relationship to one another (no attacks, defenses, or familiar pawn chains), they cannot be grouped into meaningful chunks. 3. Reduction to Novice Processing: Forced to remember the pieces as individual, isolated units, the grandmaster must rely entirely on basic short-term memory.
Consequently, the grandmaster hits the exact same biological bottleneck as the novice: they can only remember about 4 to 7 random pieces.
Summary
The chess grandmaster’s memory is not a product of raw neurological hardware, but of highly structured software. Their "genius" memory is highly domain-specific. It relies on a vast, internalized dictionary of chess patterns (chunks) stored in long-term memory. When a position makes logical sense, they recall it by recognizing the pattern. When a position is random, the illusion of photographic memory vanishes, proving that expertise is built on the meaningful organization of information, not just the capacity to store it.