🎮 How to Play
- A letter appears every 2 seconds.
- Click MATCH (or press Space) if the current letter is the same as the one N positions back.
- Higher N = harder. Tests working memory.
About this tool
N-Back: a working-memory test. A letter appears every 2 seconds. Click MATCH if it's the same as the one N letters ago. Higher N = harder. Studied as a possible IQ booster (mixed evidence). Brain training apps build on this.
About the N-back test
The N-back test is the most studied working-memory paradigm in cognitive psychology. It was introduced by Wayne Kirchner in 1958 as a way to measure short-term updating capacity, and rebuilt as a dual N-back task by Susanne Jaeggi and Martin Buschkuehl in 2008 in a much-cited paper that claimed transfer to fluid intelligence (Gf). Decades of follow-up have softened that claim. What remains uncontested is that N-back reliably measures working-memory updating: the ability to hold a sequence in mind and decide, item by item, whether the current item matches one N positions back.
This implementation shows a single letter every 2.5 seconds for 20 trials. Your job is to press the match button whenever the current letter is the same as the letter that appeared N positions earlier. At N equals 2, you compare to two letters ago. At N equals 3, you compare to three letters ago. The harder the N, the more letters you must keep in active rotation.
How to play and how it is scored
Trial sequence: letter shown for 1.5 s, then 1 s blank N-back match: current letter == letter shown N trials earlier Score per session: hits out of approximately 20 trials Difficulty levels: N = 2 (entry), N = 3 (intermediate), N = 4 (advanced) Adult benchmark: 1-back near-perfect, 2-back 70-90%, 3-back 40-70%
Score one point each time you correctly press during a real N-back match. Misses (a real match you did not press) and false alarms (pressing when there was no match) both lower your effective hit rate. Honest self-scoring requires both: the on-screen score here counts hits, but you should also track how often you pressed when no match existed.
Worked example (2-back)
Suppose the sequence is: B, K, T, K, P, T, L, T. At N equals 2:
- Letter 1 (B): no comparison yet.
- Letter 2 (K): no comparison yet (only one prior letter).
- Letter 3 (T): compare to letter 1 (B). No match. Do not press.
- Letter 4 (K): compare to letter 2 (K). Match. Press.
- Letter 5 (P): compare to letter 3 (T). No match. Do not press.
- Letter 6 (T): compare to letter 4 (K). No match. Do not press.
- Letter 7 (L): compare to letter 5 (P). No match. Do not press.
- Letter 8 (T): compare to letter 6 (T). Match. Press.
Typical adult benchmarks
Population studies and Cambridge Brain Sciences norms place adults in roughly this band. Genuinely above-average performance starts around 80 percent at 3-back.
| Difficulty | Typical adult hit rate | Trained adult | What it measures |
|---|---|---|---|
| 1-back | 95 to 100% | 100% | Vigilance, attention |
| 2-back | 70 to 90% | 90 to 100% | Working memory updating |
| 3-back | 40 to 70% | 70 to 90% | Updating with interference |
| 4-back | 20 to 45% | 50 to 75% | Capacity ceiling for most adults |
Common pitfalls
- Reading instead of rehearsing. If you say each letter silently in your head as it appears (subvocal rehearsal) you load only auditory working memory. Some people score better by visualising letter shape, others by speaking aloud. Try both.
- Chasing transfer to IQ. The 2008 Jaeggi paper claimed N-back training raised Gf. A 2014 meta-analysis (Au et al.) and Redick 2013 failed to fully replicate. Use N-back to measure capacity, not as a shortcut to a higher IQ.
- Doing dual N-back without training the single first. Dual N-back layers spatial and auditory streams. Most people benefit from mastering single 2-back and 3-back before attempting dual.
- Practising tired or after caffeine peaks. Working memory tracks alertness. Test at the same time of day to compare scores honestly.
- Comparing to a child or elderly relative. Working memory peaks in the early 20s and declines slowly with age. Compare to your own previous sessions, not someone else's score.
Related brain tools on 3Tej
Frequently asked questions
What is a good N-back score for an adult?
At 2-back, a healthy adult typically scores 70 to 90 percent hits. At 3-back, the same adult drops to 40 to 70 percent. Trained users (a few weeks of daily practice) can reach 80 to 95 percent at 3-back and start to make real progress at 4-back. Below 50 percent at 2-back without explanation (no fatigue, no distraction) is worth retesting before drawing conclusions.
Does N-back training raise IQ?
The original 2008 Jaeggi study claimed yes; subsequent replication attempts have been mixed. The current consensus: N-back training reliably improves N-back scores and modestly transfers to closely related working-memory tasks, but does not produce reliable gains on fluid intelligence (Gf) tests. Treat it as a working-memory benchmark, not an IQ booster.
What is the difference between single and dual N-back?
Single N-back streams one modality (usually visual letters or positions). Dual N-back streams two modalities at once (a spatial position and an auditory letter, scored independently). Dual N-back is significantly harder and was the version Jaeggi tested. Single is sufficient for capacity measurement and easier to train.
How often should I practise N-back?
10 to 15 minutes daily for 3 to 4 weeks is the typical training block in the literature. Beyond that, gains plateau on the trained level and you should raise N. Sessions longer than 20 minutes degrade due to fatigue.
Why do I score better in the morning than the evening?
Working memory is one of the cognitive functions most sensitive to circadian timing, sleep, hydration, and caffeine. Most adults peak between 10 a.m. and noon, dip in early afternoon, and recover slightly in the early evening. To compare scores fairly, run sessions at the same time of day.