Four coloured buttons. A growing sequence. A limit that seems frustratingly close. Simon Says has been revealing the boundaries of human memory since 1978 — and the neuroscience behind it explains far more than why you keep forgetting that third button.
Simon was invented by Ralph Baer and Howard Morrison and released by Milton Bradley in 1978. Its mechanics are deceptively elegant: the device plays a sequence of coloured light-and-sound pairs — red, blue, green, yellow — and the player must reproduce the sequence exactly. Each successful round adds one more element; failure resets the game.
What makes Simon an extraordinary cognitive tool is what it measures: sequential working memory — the ability to hold an ordered list of items in conscious awareness and reproduce them accurately. Unlike free-recall memory tests, Simon is unforgiving about order. Red, blue, green and blue, red, green are two completely different answers.
Working memory is the cognitive system that holds information in mind while simultaneously manipulating or using it. It is distinct from long-term memory — it is temporary, capacity-limited, and actively maintained rather than passively stored.
The most influential model of working memory is Alan Baddeley's Multi-Component Model (1974, revised 2000), which identifies four interconnected systems:
In Simon Says, all three slave systems engage simultaneously: the phonological loop rehearses the colour names, the visuospatial sketchpad tracks button positions, and the episodic buffer binds both into an integrated sequence. Players whose phonological loop and visuospatial sketchpad work well in coordination consistently outperform those relying on only one.
In his landmark 1956 paper "The Magical Number Seven, Plus or Minus Two," psychologist George Miller established that working memory can hold approximately 7 ± 2 items — a range of 5 to 9. This became one of the most cited findings in cognitive psychology.
However, more recent work by Nelson Cowan (2000) and others challenged this, finding that the true capacity — when rehearsal strategies are controlled for — is closer to 3–4 chunks. The larger apparent capacity (5–9) arises because experienced humans automatically chunk individual items into larger meaningful units.
Consider a 12-item Simon sequence: R-B-G-Y-R-R-B-G-Y-B-G-R. As raw items, this is 12 — far beyond working memory. But if you notice "RBGY" appears twice, and track shorter sub-patterns, you compress 12 items into 4–5 chunks. Your recall performance improves dramatically — not because your memory grew, but because you restructured the information.
Expert chess players, memorists, and musicians all heavily use chunking. Chase and Simon (1973) showed that expert chess players could reconstruct mid-game board positions from brief viewing, but performed no better than novices on random board positions — proving that expertise is stored as chunks (meaningful patterns), not raw capacity.
One of the most robust findings in memory research is the serial position effect: items at the beginning and end of a list are remembered better than items in the middle.
Approximate recall probability by list position (primacy effect = first items; recency effect = last item)
The primacy effect (better recall for early items) arises because early items receive more rehearsal time — they were in working memory longer and had more opportunity to consolidate. The recency effect (better recall for the most recent item) arises because the last item is still in the phonological buffer when recall begins, not yet displaced by subsequent items.
In Simon Says, this predicts where you will make errors: not at the very start or very end of a sequence, but in the middle. A round-10 Simon sequence of 10 items will typically see the most errors at items 4–8. Knowing this, players should concentrate extra rehearsal effort on middle items, not the beginning or end.
The phonological loop has two components: the phonological store (a short-term acoustic buffer that can hold about 1.5–2 seconds of speech sound before it decays) and the articulatory rehearsal process (the subvocal "inner speech" mechanism that refreshes items in the store by cycling through them mentally).
Memory span for words is inversely related to the time it takes to say them — you can hold more short words than long words in the phonological loop because the articulatory rehearsal loop can recycle short words faster before they decay. This is called the word length effect, and it has direct implications for Simon strategies: using abbreviated colour codes (R/B/G/Y vs red/blue/green/yellow) allows faster rehearsal cycling.
Items that sound similar are harder to maintain separately in the phonological store. This is why "green" and "keen" are harder to distinguish in a list than "green" and "red." For Simon players, this suggests using maximally distinct verbal labels — some experts use number codes (1=red, 2=blue, 3=green, 4=yellow) which are acoustically distinct and brief, offering a slight phonological advantage.
Memory errors in Simon Says rarely come from absolute forgetting — more often, items are present in memory but become confused or reordered. Two types of interference drive this:
Earlier sequences interfere with recall of later ones. After playing 10 rounds, the sequences from rounds 3, 5, and 7 compete with the current round for recall. This is why performance often dips after extended play even when the current sequence is not particularly long — the accumulated interference burden grows with each round.
Counter-strategy: Between rounds, briefly clear working memory by looking away and thinking about something unrelated for 3–5 seconds. This allows the phonological buffer to reset.
New information overwrites or distorts older memories. When Simon adds a new element to the end of a sequence, the encoding of that new item can disrupt the imprecisely-held middle items from the previous round. Research by Waugh and Norman (1965) showed that the rate of information presentation dramatically affects interference — faster sequences leave less time for consolidation, increasing error rates for middle items.
Counter-strategy: Encode each new item immediately and firmly before the next one appears, rather than letting them accumulate passively.
Allan Paivio's Dual Coding Theory (1971) proposes that the brain has two distinct symbolic systems: a verbal system and an imagery system. Information encoded in both systems simultaneously is recalled more reliably than information encoded in only one.
Applied to Simon Says: simply watching the lights flash is single-code visual encoding. Simultaneously sub-vocalising the colour names engages the verbal system as well. Research consistently shows that dual-code encoding produces retention rates 30–50% higher than single-code encoding for the same material.
One of the most interesting Simon phenomena is the "late-game choke" — sequences that players have recalled successfully many times suddenly fail at high levels, often in the 12–18 item range. Research on choking under pressure offers a compelling explanation.
Sian Beilock and Thomas Carr (2001) showed that performance pressure causes experts to devote explicit attentional resources to processes that normally run automatically — essentially re-engaging the slow, capacity-limited conscious mind in tasks that had become fast and automatic. In Simon, this manifests as players suddenly consciously counting steps in the sequence that they had been reproducing fluidly.
Anxious thoughts ("what if I lose?", "I can't remember that middle section") are not free — they consume the same limited working memory resources needed to hold the sequence. Eysenck's Attentional Control Theory estimates that performance anxiety can reduce effective working memory capacity by the equivalent of 1–2 items. This is often the difference between success and failure at high levels.
Research suggests that high-pressure performers do better when they focus on the process (each individual button as it comes) rather than the outcome (winning or maintaining a streak). Process focus keeps attention on the task-relevant information — the current item — rather than diverting it to task-irrelevant worry. This is directly applicable to advanced Simon play.
The cognitive demands of Simon Says map directly onto skills needed for language acquisition. Learning vocabulary in a new language requires holding phonological sequences in working memory during pronunciation, maintaining grammatical sequence (subject-verb-object order), and resisting interference from the native language. Children with higher Simon-task performance scores show faster foreign language vocabulary acquisition in research studies, suggesting the game genuinely exercises foundational language-learning circuitry.
Similarly, music and Simon share structural demands. Reading music while playing requires maintaining a sequence in working memory (upcoming notes), executing the current note, and anticipating the transition — a triple-task situation directly analogous to advanced Simon play. Musicians, on average, outperform non-musicians on auditory sequential memory tasks, and regular Simon-style practice produces measurable improvements in musical sight-reading performance.
Simon Says engages a distributed neural network including the prefrontal cortex (executive control and working memory maintenance), the hippocampus (sequence binding and episodic memory formation), the basal ganglia (sequence learning and procedural memory), and the cerebellum (timing and sequential prediction).
Neuroimaging studies using fMRI show that as sequences become longer, prefrontal cortex activity increases substantially — reflecting the growing executive demand. Crucially, with practice, some of this activity shifts toward the basal ganglia, indicating that well-learned sequences become more automated and require less conscious effort. This shift is the neural signature of moving from the cognitive to the autonomous stage of motor learning described by Fitts and Posner.
For practical purposes, this means that the 15-item expert who seems effortless is not using a better working memory than the struggling beginner — they have simply offloaded more of the sequence to procedural memory systems that operate in parallel with conscious thought, effectively expanding their usable capacity.