Game Guide
The five Simon Says demo sequences, why level four is harder than level five looks, and the chunking method that carries a six step pattern.
Simon Says in MyRin is dressed as a rank of taiko drums. A pattern sounds, the drum heads light in order, and you play it back. The drums are told apart by the colour of the mon on the head, red, blue, green and yellow, with the Japanese colour name on each, and the sequence lengthens every level.
The five free browser levels run two, three, four, five and six steps. That reads like an even ramp and is not one, because the fourth level introduces a structural problem rather than just another step.
Here is the MyRin short for Simon Says. This game genuinely needs video: the demonstration timing is the thing you are working against, and no description of it substitutes for seeing the rhythm.
Each drum lights for roughly seven tenths of a second, with about a third of a second of darkness between flashes, and a pause of under a second between the end of the demonstration and the moment your input is accepted. Those numbers matter for one reason: the gaps are long enough to name a colour in, and not long enough to do anything more elaborate.
Your replay, by contrast, is untimed. The game waits. This asymmetry is the most exploitable fact about the game and the one most players ignore, tapping their answer back at the speed the demonstration played rather than at the speed they can be sure of.
Red then blue. Two items is far inside anybody's span and there is nothing to strategise. Use it to check the interaction: how the drum responds to a tap, and how long the input window feels before you commit.
Green, red, blue. Still inside effortless span. This is the level on which to start sub-vocalising, while it is unnecessary, so that it is habitual by the time it is essential.
Sub-vocalising means quietly saying the colour names as they flash, either aloud or internally. It sounds trivial. What it does is add a verbal encoding alongside the visual one, so the sequence is stored twice in different formats. When the visual trace decays, which it does within a couple of seconds, the verbal trace is still there and can be rehearsed by repeating it.
Blue, yellow, red, green. Four distinct colours, no repeats, and each colour appearing exactly once, which makes this the easiest four step sequence that could exist. If you have been sub-vocalising, you are simply reciting a four word list.
Level three is also the last level where you can get away with remembering the sequence as a list of individual items. From here on, item by item storage runs out.
Red, red, blue, green, yellow. This is the difficult level in the demo, and the difficulty is not the length.
Two identical flashes back to back are perceptually ambiguous. A drum lighting, going dark, and lighting again reads as one longer event unless you are deliberately counting, and if you store "red, blue, green, yellow" you will fail a sequence you actually saw correctly. The error is in the encoding, not the memory.
The defence is to count the repeat explicitly as you watch: not "red" but "red twice", or "two reds". Converting the repeat into a count turns two fragile events into one robust item, and it has the pleasant side effect of shortening the list you have to hold from five items to four.
This is also where the untimed replay earns its keep. When you tap back a repeat, pause deliberately between the two taps. Two fast taps on the same drum risk registering as one input, and you will have lost a sequence you remembered perfectly.
Green, blue, red, yellow, blue, red. Six items, with blue and red each appearing twice but never adjacent, so the repeat problem from level four does not apply.
Six is above the comfortable span for most people as a flat list, so chunk it. Split the sequence into two groups of three: green, blue, red, then yellow, blue, red. Now you are holding two items instead of six, and each is a small verbal unit you can rehearse as a phrase.
The grouping is helped here by a coincidence worth exploiting: both chunks end in blue, red. Recognising internal structure like that is what expert performance on this kind of task actually consists of. Nobody holds six independent items; they hold two or three structured ones.
Rushing the replay is the biggest. The game is not timing you, and tapping at demonstration speed converts a memory task into a motor accuracy task you had no reason to accept.
Watching the drums instead of naming them is the second. Purely visual storage of a sequence decays quickly. If you are silent while the pattern plays, you are relying on the weakest available channel.
Trying to hold more than four items as a list is the third. Span limits are real and they are not a matter of effort. The way past them is always to restructure the material, by chunking or by counting repeats, not to concentrate harder.
Simon Says is a sequential memory span task with a motor reproduction component, and it is close kin to the digit span tests used in cognitive assessment. What improves with practice is chunking, rehearsal and repeat detection, all strategies, and those improvements are real and fairly quick to acquire.
What does not improve much is the underlying span itself. Expect to get better at the game and to notice that you reach for chunking automatically in other places where a sequence has to be held. Do not expect the number of raw items you can hold to grow. The installed game runs the same mechanic across 600 levels, which is where sequences get long enough that chunking stops being optional advice and becomes the only way through.