Number Memory

This test measures digit span — the number of digits you can hold in short-term memory and reproduce in the correct order. A number is displayed for a few seconds, it disappears, and you type it back. Get it right and the next number is one digit longer. Get it wrong and the run ends.

Digit span is deliberately content-free. The digits mean nothing, follow no pattern, and connect to nothing you already know, so vocabulary, education and general knowledge give no advantage. That is what makes it a comparatively clean measure of raw short-term capacity.

Your score is the longest number you recalled exactly. A single wrong or transposed digit ends the attempt, so the score reflects perfect recall rather than partial credit.

Short-term memory is the workspace everything else runs in. Following directions, holding a phone number between hearing it and dialling, keeping track of a sentence long enough to understand its end — all of it draws on the same limited buffer this test probes.

Digit span also has an unusually long research history. It appears in the Wechsler scales and in decades of cognitive studies, which means the task itself is well understood even when a browser version of it is not a clinical measurement.

For most people the practical value is the strategy it exposes. Discovering that you can hold eleven digits by grouping them, when you could only hold seven by reading them individually, is a demonstration of chunking you can apply well beyond the test.

The measurement itself is simple and reliable: either you reproduced the number exactly or you did not. There is no timing, no hardware latency, and no display dependency, which makes this one of the more device-neutral tests on the site.

What varies is the strategy you bring to it. A first attempt usually measures unrehearsed capacity, and later attempts measure how quickly you learned to chunk. Both are real, but they are not the same thing, so compare like with like.

Digit span responds to technique far more than to practice volume. These are the changes that actually move the number:

• •Group the digits into pairs or triplets as they appear rather than reading them one by one
• •Look for structure — a repeated pair, an ascending run, a year — and use it as one chunk
• •Rehearse silently in the gap between the number disappearing and the field appearing
• •Type the answer promptly; hesitating gives the earliest digits time to decay
• •Take the test somewhere quiet, since speech competes directly with verbal rehearsal
• •Try again after sleep rather than late at night, when short-term memory degrades first

Be realistic about what improves. You are getting better at encoding digits efficiently, not enlarging your memory. That skill transfers to other strings of numbers and very little else.

Anyone curious where their short-term memory sits against the average

Students and professionals who hold numbers briefly — codes, references, readings

People who want a clean baseline to track across sleep, caffeine or time of day

Anyone who wants to see chunking work on themselves in about two minutes