Sequence memory test average, from Corsi blocks to Button Chain

September 26, 2026 · 6 min read · Tyamka team

How long a sequence adults usually repeat in the Corsi block test and in Simon-style games, where the numbers come from, and how Button Chain differs.

If you have just played a game where buttons light up and you repeat the order, and you want to know the average, the research gives a rough answer. In the Corsi block test, the classic lab version of this task, one often cited adult sample averaged about 6.2 positions. For the electronic game Simon, one study reports an adult span of about seven. Both numbers come with conditions, and neither is a norm for a particular web game. Below is where they come from and how our sequence memory test differs from the lab.

The Corsi block test

The task was developed by Philip Corsi for his doctoral thesis at McGill University in Canada, finished in 1972 and supervised by the neuropsychologist Brenda Milner. The setup is simple. Nine wooden blocks are fixed on a board in an irregular pattern. The examiner touches some of them one by one, and you touch the same blocks in the same order. The sequences get longer until you can no longer repeat them. Your block span is the longest sequence you got right.

The test was meant as a spatial counterpart to digit span, where you repeat a list of spoken digits. It became one of the standard tools in neuropsychology, and today there are many computer versions where squares on a screen light up instead of a hand touching blocks.

What the averages say

Two numbers come up most often.

A standard deviation of 1.3 means that spans in that sample varied quite a lot. The number also depends on how the test is run. How many trials you get at each length, how fast the blocks are shown and whether one mistake ends the test all change the result.

Why would Simon give a longer span than Corsi? We do not have a direct comparison, so this is only a guess. Simon has four large buttons with fixed colours, and in the original game each one also plays its own tone. You can hold the sequence as a string of colours or a short tune. Corsi has nine identical blocks and no sound, so you have to hold positions.

Backwards is not always harder

In digit span, repeating a list backwards is clearly harder than repeating it forwards. For Corsi this is less clear. A study by Roy Kessels and colleagues in 2008, on 246 healthy adults aged 50 to 92, found that backward Corsi was not more difficult than forward Corsi, while backward digit span was more difficult than forward. That was one sample of older adults, so it is a result for that group, not a rule. It is also the reason we are not sure the backward levels in our games are as hard as they look.

How Button Chain differs from the lab

Our game follows the idea of Simon more than Corsi. A sequence is shown, you repeat it, and each round the same sequence comes back one button longer. The levels work like this.

• Levels 1 to 5. Four buttons. The same button can light up twice in a row. Rounds start at 3 to 5 buttons and end at 4 to 6.
• Levels 6 to 10. Six buttons, and the same button never lights twice in a row. Level 10 ends with a round of 8, the longest in the game.
• Levels 11 to 13. Six buttons, and you repeat the sequence backwards, from the last button shown.
• Levels 14 and 15. Nine buttons, and after each show the buttons change places. Every button carries a number, so you have to follow the numbers, not the positions.
• Mistakes. A wrong tap costs one of three lives, the right button is outlined, and the round goes on. The sequence is not shown again.
• Speed. Buttons light for 600 milliseconds at first and 400 at level 10, and each level also has a speed bonus for quick taps.

So the length of the longest round you cleared is not a Corsi span. Four buttons with repeats are closer to a short repeating pattern than to nine blocks. The hint after a mistake is not part of the usual Corsi procedure either. And the swapping levels on 14 and 15 test something the Corsi board never asks, following a number while its position changes. We set the longest round at 8 because the adult spans reported for Corsi and Simon are around 6 to 7, but all our level settings are a guess until playtests. We have not measured how far players get.

Getting further in a sequence game

Nothing here comes from a study of Button Chain, so treat it as suggestions to try rather than tested advice.

• Split the sequence into threes. In a 1964 study of digit lists, Wayne Wickelgren found that people who rehearsed in groups of three remembered more than those who used pairs or groups of five. A sequence of seven buttons becomes three, three and one.
• Give it a rhythm. Tap along in your head as the buttons light, with a short pause after each group. The pause marks where one group ends.
• Name what you can. On the four-button levels each button sits in a corner, so a sequence can be read as a short string of words, like "top left, top right, top right, bottom left". On the swapping levels the numbers are the only thing that stays, so read them as a number, like 3 7 1.
• Build on the last round. Each round repeats the same sequence with one button added at the end. You only need to add one new item to what you already hold.

How to read your own result

The results screen shows points, the time, mistakes and your best on that level. It does not show an average for other players, because we do not have data for one.

The useful comparison is with yourself. Play the same level a few times, note the middle score, and look again a week later. Each attempt uses a new random sequence, and some come out easier than others, so single attempts jump around. A slow upward trend in the middle score means you have got better at this game. We do not claim it means more than that.

For the spatial side of the same topic, read about visual working memory. For numbers instead of positions, see the number memory test average. And for the idea that ties these games together, read what working memory is. All games are on the all games page.