Memory Tests
Five tasks that find the point where your short-term memory gives out.
Short-term memory has a ceiling. It is lower than people expect, and it is easy to find. Every test in this hub does the same basic thing. It hands you more to hold than you can hold, then reports where you stopped coping.
They differ in what they ask you to hold. Digit span uses numbers, which most people repeat silently to themselves. Corsi block tapping uses positions on a board, which you cannot say to yourself at all. It usually comes out one or two units lower.
Visual memory shows a whole pattern at once rather than a sequence. Sequence memory adds order to position. The n-back task is the odd one out. It never lets you stop updating, so it measures juggling rather than holding.
If you take only one, take digit span. It has a century of published comparison data behind it. For the most honest picture of your own limits, take digit span and Corsi, then look at the gap.
The 5 tests in this domain
Numbers and reading in this domain
Span is not one number, and it is not seven
Miller's 1956 paper on the magical number seven is the most misquoted result in psychology. He was writing about chunks, not digits, and he said so plainly.
Seven applies to items with nothing in common. People routinely hold twelve digits by grouping them into four lots of three. Cowan's later work puts the real limit closer to four, once you block rehearsing and grouping.
That matters for reading your score. A digit span of 9 does not mean nine slots. It usually means you grouped digits into pairs and triples on the fly. That is a technique, and techniques can be trained in a way that raw capacity cannot.
Why the spatial version comes out lower
Words can be refreshed by an inner voice that runs as fast as you need. Positions cannot. There is no inner voice for 'third block, then first, then fifth'.
That is the main reason Corsi spans land around five to six blocks while digit spans land around seven digits in the same person.
It is also why the gap between your two scores is interesting. A large gap either way is common and normal. It mostly tells you which method comes naturally to you. It is the closest thing on this site to a real profile rather than a single ranking.
The training question, answered honestly
People arrive at memory tests wanting to know whether practice makes them better. Practice reliably makes you better at the test you practise. Whether it carries over to anything else is genuinely contested.
Two decades of trials give an honest summary. Gains on untrained tasks are small, inconsistent, and often absent once you rule out people simply expecting to improve.
The strongest single review is Simons and colleagues in 2016. They checked the brain-training industry's claims against the evidence and found the evidence far weaker than the marketing.
So we are not going to tell you that ten minutes of n-back will make you smarter. It will push your n-back score up. If that is satisfying the way beating a lap time is satisfying, that is reason enough to come back.
Getting a clean measurement
Do not write anything down. Do not say the items out loud unless the instructions tell you to. Take the test somewhere you will not be interrupted, because one interruption at the wrong moment costs a whole level.
Digit span and Corsi both end when you fail twice at the same length, so a single lapse of attention is expensive.
Expect your first attempt to be your worst. Nearly everyone gains a point or two over the first few runs purely by learning the interface. Compare your recent attempts with each other, not your first attempt with someone else's tenth.