N-Back Test
Letters arrive one at a time. Press Match whenever the letter on screen is the same as the one two letters ago — which means holding two letters and replacing them both on every step. This is the n-back task in its single 2-back form: free, no account, forty seconds in your browser.
Take the test
Press Match when the letter is the same as two letters ago.
READY
—
targets caught
N-back loads working memory: you hold the last two letters and update them on every step. Missing some is normal — the task is built to sit near the limit.
- Hits
- —
- Missed
- —
- False alarms
- —
Needs sight — the letters are the only signal. Nothing is uploaded; your result stays in this tab.
That was a single run. In the app it becomes a short daily session, with a Focus Index to show for it.
Point your camera at the code — the App Store opens.
What the n-back task measures
Most memory tests give you something to hold and then ask for it back. This one never lets the contents settle: every new letter pushes the oldest one out, and the answer you owe changes with it.
That makes it a measure of working-memory updating rather than of storage. The load is not "how much can you keep" so much as "how cleanly can you keep replacing it" — and the failure you will feel is not forgetting so much as confusion between the letter two back and the one three back.
The task is also unusually unforgiving by design. There is no way to pause, no way to review, and the window for each answer closes on its own. One glance away costs a target, which is why attention and working memory are hard to separate here at all.
Where it comes from
- 1958
Kirchner needs a moving target
Wayne Kirchner wants to measure memory for information that keeps changing, not memory for a list that sits still. His answer is to keep a stream running and ask whether the current item matches the one n steps back — so the answer must be updated on every single step.
- 2008
The claim that made it famous
A study by Jaeggi and colleagues reports that training on a harder dual version raises scores on tests of fluid reasoning. The finding spreads far beyond the lab and becomes the foundation of an industry.
- Since
And the claim that did not hold up
Larger reviews and replications found little or no transfer beyond the trained task itself. People reliably get better at n-back. Whether that carries over into general reasoning is, at best, unsettled — and mostly looks small or absent.
How to read your result
Three numbers, and the third one decides whether the first means anything.
-
Hits
Targets you caught, out of the 5 that appeared. The headline — but on its own it is not enough, which is what the next two are for.
-
Missed
Targets that went by without a press. Together with hits this always adds up to the number of targets, so nothing hides between them.
-
False alarms
Presses on a letter that was not a match. Read this first: press on everything and you catch every target, so hits only mean something once you know what they cost.
The reason false alarms come first is simple arithmetic: pressing Match on every letter catches all 5 targets. A count of hits with no count of wrong presses is not a score, it is half of one — which is why both are on the card whether you want them or not.
What this version cannot tell you
-
One short run
14 scored letters is enough to feel the task and get a rough count, not enough to pin down a capacity. Research versions run many blocks and average them.
-
This is single, not dual
The famous training studies used a dual version — letters and positions at once — which is far harder. Do not compare a number from here with anything reported for dual n-back.
-
Distraction is the whole task
A pace you cannot control plus a fixed window means one glance away costs you a target. That is not a flaw in the measurement; it is what the measurement is about.
-
It is not a diagnosis
N-back appears in research batteries alongside other measures, interpreted by people trained to do it. A rough run here means you had a rough run here.
Questions about the n-back test
Does n-back training raise your IQ?
That is the claim the task became famous for, and it did not survive scrutiny. A 2008 study reported transfer from a harder dual version to tests of fluid reasoning; larger reviews and replication attempts since then have found little or no transfer beyond the trained task. You will get better at n-back if you practice it — that part is solid. Treat anything beyond that as unsettled, including when we would be the ones benefiting from the claim.
What is a good n-back score?
On a 2-back this short, catching most of the targets without a pile of false presses is a solid run. We deliberately do not publish a threshold: with only five targets, one lapse moves the count a long way, and a number from a browser tab on an unknown device is not comparable to a lab figure.
Why does it start with two letters I cannot answer?
Because on an n-back there is nothing two steps back yet. The first two letters are a warm-up and can never be a match — they exist to fill the memory you are about to be tested on. Pressing Match on them counts as a false alarm.
Why letters instead of positions or shapes?
Letters are unambiguous and need no explanation, which keeps the n-back about holding and updating rather than about decoding what you are looking at. The pool also leaves out easily confused pairs, so a misread glyph cannot end up in your score.
What is dual n-back, and is this it?
This is not dual n-back. Here one stream runs — letters — and you answer about letters. In the dual version two streams run at once, usually a letter and a position on a grid, and you have to answer about each of them separately on every step. That is a much heavier load than doubling the work sounds like, because both streams have to be updated in the same couple of seconds. The difference matters beyond difficulty: the training claims the task became famous for were made about the dual version, so a single-stream score is not evidence for or against them.
N-back vs digit span: what is the difference?
Digit span asks how much you can hold: a list is read out, you repeat it back, and it grows until you cannot. N-back asks how well you can keep replacing what you hold: the list never stops and the answer you owe changes with every item. Span is a capacity, n-back is an updating rate, and the two come apart in the same person — which is why batteries that care about working memory tend to include both rather than pick one.
How long should you train n-back?
Honestly, we cannot give you a number, and neither can anyone else with evidence behind it. The training studies used schedules of roughly twenty minutes a day over a few weeks, but since their transfer claims did not replicate, those schedules are not a dose for anything except getting better at n-back itself. If you enjoy the task, do it for as long as you enjoy it; if you are doing it because you were promised a smarter brain, read the section above first.
Do I need an account?
No. The n-back test runs entirely in your browser, nothing is submitted, and your result is shown to you and then gone when you close the tab.
One run is a snapshot. The trend is the point.
In NEURA the pace adapts to you and the result joins the rest of the catalog — 19 games — in a single Focus Index, so what you watch is your own line over weeks rather than one count from one afternoon. That is the part a single page cannot do, and the only part we will promise.
