How much can you hold in mind at once?
Digits appear one at a time. When they stop, type them back in the same order. That is the whole task — no strategy, no writing it down, no second look.
It takes about two minutes. There is no account, no email, and nothing to sign up for. Your answers are not stored or sent anywhere — the test runs entirely in your browser and your result exists only on your screen.
Sequences start short and get longer. You get two attempts at each length.
Works with a keyboard, a numeric keypad, or one thumb.
Before you start
- It measures one narrow thing — how much you can hold at once, right now, in the conditions you are in.
- Conditions matter more than you might think. Tired, distracted, or on a train are all part of the result rather than exceptions to it.
- This is the one test here that may work without sight. The digits are put on screen as text inside a live region, so assistive software is given something to read out, and the answer is typed on a keypad that is a set of ordinary buttons. We have not tested it with a screen reader and will not claim it works until we have — we are telling you what the page does, not what your software will make of it. If you try it, we would like to know what happened. Odd one out and paper folding cannot be done without sight, and say so.
- Doing it twice tends to raise the number without anything about you having changed. Across 68 studies of working memory tasks the average rise from a first attempt to a second was small — about a quarter of a standard deviation — and an average is not a prediction about you.
Digit span test in progress
Watch the digits.
Type the sequence back, in order.
Your result
What this task tested
This task asks you to hold a sequence in mind for a few seconds and give it back exactly — no strategy, no writing it down, no second look. It is one task in the Working Memory domain.
What that means
What this did not measure
This says nothing about long-term memory, about whether you remember faces or names, or about your memory as a whole. It measures one narrow thing: how much you can hold at once, right now.
This is a domain score, not a BrainScore. A single task in a single domain does not make a composite, and we are not going to invent the rest of it.
Your seven scores make one shape. Take the other tests and this number becomes one arm of your Cognitive Fingerprint — seven arms and a single number between 0 and 100. That number says how far you got through what these four tasks can measure. It is not a percentile and not a comparison with other people, and every constant behind it is published, including the one we chose rather than derived.
An online test cannot tell you whether something is medically wrong. This one cannot, and neither can we. That is not caution — it is the limit of the task. If your memory is worrying you, a doctor is the right person to ask. And you do not need a low score here to justify the appointment.
How much should you trust this number? What moved it that is not you, what it legitimately supports, and what it cannot.
Things that moved this number that are not your memory
- Repeat attempts. Scores on working memory tasks tend to rise on a second administration without any underlying change — a meta-analysis of 68 studies puts the average rise at about a quarter of a standard deviation, and finds it still growing up to a fourth attempt. A second attempt is not progress, and an average across studies does not tell you what your own second attempt will do.
- Your day. Sleep, stress, illness, caffeine, time of day and interruptions are conditions we cannot hold constant and did not measure.
- Where you were. This taken on a train is not comparable to this taken at a desk.
Should you believe this number?
Some people take a result like this and move on. Others immediately want to know whether the test is any good — and if that is you, the answer is that you are right to ask, and here is what this number is actually worth.
What it legitimately supports: that on this task, today, under whatever conditions you were in, you held this many digits in order. That is a real measurement of a real thing, and it is narrow on purpose. It also gives you something to compare against — yourself, on another day, under conditions you can name and tell apart.
What it cannot support: any statement about how you compare with other people. We have measured nobody, our reference sample is zero, and we would rather say that than imply a comparison we have not earned.
Why your digit span will not match another site's →
If another test gave you a different number, both can be right. "Digit span" names a family of procedures, and the procedure sets the figure — in one published study the same people produced three different averages from a single sitting, depending only on which scoring rule was applied. What is a normal digit span? sets out why, with the sources.
The rule we used is published in full further down this page, including the exclusions and the band boundaries. If you disagree with the number, you can check the working and find out where we differ — that is the point of publishing it.
Two questions, if you have twenty seconds
We are trying to find out whether explaining a result is worth doing at all. Nothing is recorded except which buttons you press — no result, no identifier, nothing that could be traced to you. There are no wrong answers here and nothing follows from your choice.
1 · Did the explanation above make sense to you?
2 · Did it tell you anything about yourself?
Recorded. Thank you — that is genuinely useful.
Want the harder version?
There is a second phase where the digits come the same way but you type them back in reverse order. It is meaningfully harder, because holding a sequence and turning it round are not the same job.
It is optional and it does not change the result above. Your forward span is your result whether you do this or not. Nothing above is being withheld.
Reverse span
Your result is still your forward span. The reverse number is reported next to it, not folded into it — combining two different demands into one figure would make the number harder to check, not easier.
What next
Read what this actually measures
Working memory is the thing this task is a narrow sample of. What it is, what genuinely moves it from day to day, and what the research does and does not support about training it — including the part most sites leave out.
Working memory: what it is, what moves it, and what the evidence actually shows
Wondering whether your number is normal
It is the question almost everyone has at this point, and it has a real answer — just not a single number. Published figures for digit span disagree with each other because the scoring rule, the sample and the way the test is given all change the result, and the studies you will find were run under conditions this test does not reproduce.
Find out something about your conditions
Your span today is one measurement taken under one set of conditions. If you want that to be more informative, the useful thing is not to repeat it now — it is to repeat it once under conditions you can name and tell apart: rested against tired, morning against late evening, quiet room against a noisy one.
A difference between two such attempts tells you something about the conditions. It does not tell you that anything about you has changed, and we would be inventing a finding if we said otherwise.
You do not have to write it down. Name the conditions for the attempt you just did and we will hold it on this screen while you do a second one.
We are not asking you to come back, and nothing here is keeping count. There is no streak, no reminder, no record of your last attempt — nothing is stored, so there is nothing to return to. This is a suggestion about how to read your own number, not a request for your time.
— worth knowing that scores on this kind of task tend to rise on a second attempt because the task has become familiar, not because anything about you has changed — and that the measured average rise is small, which is a fact about a group of studies rather than a forecast for you.
Check our working
- How BrainScore works — what we measure, how a task becomes a score, and what the score cannot tell you
- How this test is scored — the rule, the exclusions and the band boundaries, in full, further down this page
- Claims standards — the language we ruled out in advance, including the comparisons we will not make
How this is scored
Published in full, because a score you cannot check is not worth much.
The task
- Digits appear one at a time, each visible for 1000ms with a 200ms gap. The rate is fixed and does not adapt to how you are doing — adaptive timing would make results incomparable between attempts and between people.
- Sequences start at length 3. You get two attempts at each length.
- You advance a length if at least one of the two attempts is correct. The test stops when both attempts at a length are wrong.
The scoring rule
Your forward span is the longest length at which at least one of your two attempts was correct. That is the whole rule. There is no adjustment, no weighting, and no hidden modifier. What is shown is what was computed.
The reverse phase
- It is offered after your result, never before, and it is optional. Your result is complete without it and nothing is held back pending it.
- It runs the same procedure with the same rule — same fixed rate, same two attempts per length, same start at 3 — and the only difference is that a correct answer is the sequence reversed.
- Your reverse span is reported next to your forward span, not combined with it. They measure different demands, and averaging them would produce a number that is harder to check rather than easier.
- The reverse span carries no band. Bands exist for the forward task, and inventing a second set for a phase we have no basis to band would be exactly the thing this page exists to avoid.
Exclusion rules
- An attempt where entry begins before the sequence has finished is void, and that length is presented once more.
- An attempt abandoned by navigating away is not scored; the result stands at the last completed length.
- An attempt where this tab is hidden while the sequence is showing is void, and that length is presented once more. Browsers slow down timers in background tabs, so switching away leaves each digit on screen longer than the timing above — an easier task than the one described here. Added 5 September 2026, after we measured it happening: with the tab hidden, each digit stayed up for about 2000ms instead of 1000ms.
The bands
| Forward span | What the result says about the task |
|---|---|
| 3 or fewer | A short sequence, and conditions weigh heavily here |
| 4–5 | A sequence held and reproduced exactly |
| 6–7 | More held at once than the task invites |
| 8–9 | Beyond raw holding — structure is doing the work |
| 10 or more | At the task’s ceiling, where it stops resolving |
These boundaries are a design decision, not an empirical finding. They are set from how the task is built, because we have no reference sample to set them from. When real data exists they will be revisited, and the methodology version will increment when they are. The norming ladder, and where we are on it, is published in full.
What we do with your answers
Nothing. The test runs entirely in your browser. Your digits, your attempts and your result are never sent to us and never stored. Closing the tab destroys them. The privacy notice covers this in full.