Getting better, honestly
You can get much better at memory tasks in a few weeks. You almost certainly cannot make your underlying capacity much bigger, and nobody has convincingly shown that practising memory games makes you better at reasoning, reading or maths. Both statements are true at once, and holding them together is the key to spending your practice time well.
What the training research actually found
Around 2008 a study reported that training on a demanding n-back task improved fluid intelligence, with bigger gains for more training. It set off a decade of studies and a commercial boom in "brain training". The larger, better-controlled work that followed tells a more sober story.
- Melby-Lervåg and Hulme pooled 23 studies. Training produced reliable short-term gains on working memory tasks, but no convincing evidence of improvement in reasoning, attention control, reading or arithmetic.
- Melby-Lervåg, Redick and Hulme repeated the exercise with 87 publications. Straight after training there were reliable gains on untrained working memory tasks, verbal and visuospatial. On far-transfer measures such as nonverbal ability, reading comprehension and arithmetic, there was no convincing evidence of reliable improvement against a treated control group, and the size of the working memory gain did not predict any far-transfer gain.
- Redick and colleagues gave young adults 20 sessions of adaptive dual n-back, the same task as the 2008 study, against an active placebo group. Both groups improved on the task they practised. Neither improved on any of the ability tests.
- Au and colleagues, a group that includes the authors of the original 2008 study, meta-analysed 20 n-back training studies and reported a small but significant positive effect on fluid intelligence. A Bayesian re-analysis of the same data put the overall effect at g = 0.24 but found that the evidence for transfer came from studies with passive, no-contact control groups; studies with active control groups modestly favoured no effect, which points to expectation or placebo effects.
- Simons and colleagues reviewed the evidence cited by brain-training companies and found extensive evidence that training improves the trained tasks, less evidence for closely related tasks, and little evidence that it transfers more broadly.
What does improve, and by how much
If capacity barely moves, where do real gains come from? Mostly from three places.
1. Strategy
The most dramatic memory improvement ever recorded in a laboratory was almost entirely strategic. In 1980 Ericsson, Chase and Faloon reported a student, known as SF, who after more than 230 hours of practice raised his digit span from 7 to 79 digits. He did it by recoding groups of digits as running times, which as a keen long-distance runner he already knew well, so each group became one meaningful chunk. His short-term store had not grown. He had built a way to use long-term memory to pack more into it.
You will not spend 230 hours on this, and you do not need to. If you currently just try to hold the items as they come, a deliberate grouping strategy is the cheapest improvement available to you. Reviewers of the training literature draw exactly this line between strategy training, which teaches you to use capacity better, and core training, which tries to enlarge capacity by repetition.
2. Familiarity with the format
On an unfamiliar game you spend early trials decoding the screen, and an adaptive task that starts you in the middle reads those trials as failures and steps you down. Familiarity removes that cost completely. It is not a trick: it simply lets the test measure you rather than your first encounter with its interface.
3. Your state on the day
Working memory is unusually sensitive to sleep and stress, which the next section covers. Gains here are free: they cost you nothing but planning.
Strategies for each task family
A strategy is only useful if you can run it automatically under time pressure, so pick one per family and rehearse it until it is boring.
| Task family | Strategy | Why it works |
|---|---|---|
| Forward span | Group into threes (or 3 + 2, 3 + 3 + 1) and hear them as a rhythm | Turns seven items into three chunks |
| Backward span | Chunk forwards as usual, then reverse chunk by chunk | Reversing three small chunks is easier than reversing seven items |
| Ascending order | Note the smallest and largest as they appear; sort at recall | Anchors the ends, leaves only the middle to sort |
| n-back | Keep a rolling rehearsal of exactly n items, spoken as short labels | Updating becomes "say, compare, drop the oldest" |
| Complex span | Answer the interruption fast, then rehearse the stored items once before the next | Shortens the time the interruption can overwrite them |
| Visual memory | Name regions and anomalies in words ("top left dark, centre stripe") | Recruits the verbal store alongside the visual one |
Try the updating strategy now. Letters appear one at a time; press Match when the letter is the same as the one two back. Say the last two letters under your breath as a pair, and replace the older one each time.
Practice drillSign in to try it
How most people practise
- Grind one favourite game for an hour
- Hope capacity grows by itself
- Change strategy every few trials
- Practise tired, late at night
- Judge progress by percentage correct
How to practise
- Rotate across the memory games in short blocks
- Pick one named strategy per task family and drill it
- Keep a strategy for a whole session, then review it
- Practise when you would sit the real test
- Judge progress by the level you hold, since adaptive tasks keep accuracy roughly steady
Spacing: why short and often beats long and rare
Everything you learn fades on a predictable curve. Ebbinghaus measured it on himself in the 1880s, and a careful modern replication by Murre and Dros in 2015 found the same shape: fast loss in the first hours, then a slower decline.
The remedy is spacing. A review of 317 experiments by Cepeda and colleagues found that spreading practice out beats massing it together, and that the best gap between sessions grows with how long you need to remember. In a follow-up with more than 1,350 people, the best gap was roughly 20 to 40% of the retention interval for a test a week away, falling to about 5 to 10% for a test a year away. A large review of study techniques rated only two as high utility: distributed practice and practice testing.
- Crammed once
- Spaced reviews
For memory tasks, spacing matters for a specific reason. What you are learning is a strategy, and a strategy only helps under test pressure once it runs without conscious effort. Four 20-minute sessions across a week build that automaticity better than one 80-minute session, and they avoid the fatigue that makes long sessions teach you to perform badly. Our session quality weighting reflects this: very fast guesses, long idle gaps and a tab left in the background all lower a session's weight.
Your state on test day
Sleep and pressure affect memory tasks more than most other tests.
Working memory accuracy dropped by a medium-sized amount after a night without sleep, far more than reasoning accuracy did. If a game-based round is scheduled for the morning after a late night of revision, the revision probably cost you more than it gained.
Pressure works through the same channel. Worry takes up working memory, and in one well-known study it hurt people with the most capacity on the problems that needed the most. Practical countermeasures:
- Familiarise yourself with the format beforehand, so there is less to worry about.
- Use a fixed pre-task routine: read the instructions fully, choose your strategy, take one slow breath, start.
- Let a bad trial go. Adaptive tasks drop the next item a step when you miss, which is not a verdict. Carrying the last miss into the next trial is how one error becomes three.
A four-week plan with our games
This plan assumes four or five short sessions a week, 20 to 30 minutes each. Rotate games inside each session rather than doing one game for the whole time.
- Week 1: learn the formats
- Two short blocks each of Digit Span, Flashback and GridRecall; read each game lesson first
- Choose one named strategy per game and write it down
- Note the level you settle at in each game, not your percentage
- Week 2: drill strategies
- Digit Span: chunk every sequence into threes, forwards and backwards
- Flashback: rolling rehearsal of the last n items, spoken as labels
- Add DiscoNumbers and Puzzle Reconstruction, one block each
- Week 3: add interference
- GridRecall twice per session: fast symmetry judgements, rehearse cells between them
- Flashback at your top level: watch for near-miss lures
- One session in test conditions: quiet room, phone away, no pauses
- Week 4: consolidate
- Mixed sessions of all six memory games, including Sequence Transformer
- Review your trial reports: are your errors order slips or items lost
- Last two days: short sessions only, sleep well, keep the same routine
Throughout, judge progress by the level you can hold rather than by your percentage correct. Adaptive tasks push the difficulty up until you miss often enough to hold steady, so your accuracy stays in a narrow band by design. The level you work at, and the score that weighs it, is what moves.
Key points
- Working memory training improves the trained tasks and close relatives; large meta-analyses find no reliable transfer to reasoning, reading or arithmetic.
- Real gains come from strategy (chunking, rolling rehearsal, fast dispatch of interruptions), format familiarity and your state on the day.
- One named strategy per task family, drilled until automatic, beats improvising.
- Short, spaced sessions build automatic strategies better than long ones; distributed practice and practice testing are the best-supported learning techniques.
- A night without sleep cuts working memory accuracy by a medium-sized amount; pressure takes capacity through worry.
Sources
- Jaeggi, S. M., Buschkuehl, M., Jonides, J., and Perrig, W. J. (2008). Improving fluid intelligence with training on working memory. PNAS, 105(19), 6829–6833. doi:10.1073/pnas.0801268105
- Melby-Lervåg, M., and Hulme, C. (2013). Is working memory training effective? A meta-analytic review. Developmental Psychology, 49(2), 270–291. doi:10.1037/a0028228
- Melby-Lervåg, M., Redick, T. S., and Hulme, C. (2016). Working memory training does not improve performance on measures of intelligence or other measures of far transfer: evidence from a meta-analytic review. Perspectives on Psychological Science, 11(4), 512–534. doi:10.1177/1745691616635612
- Redick, T. S., Shipstead, Z., Harrison, T. L., et al. (2013). No evidence of intelligence improvement after working memory training: a randomized, placebo-controlled study. Journal of Experimental Psychology: General, 142(2), 359–379. doi:10.1037/a0029082
- Au, J., Sheehan, E., Tsai, N., Duncan, G. J., Buschkuehl, M., and Jaeggi, S. M. (2015). Improving fluid intelligence with training on working memory: a meta-analysis. Psychonomic Bulletin and Review, 22(2), 366–377. doi:10.3758/s13423-014-0699-x
- Dougherty, M. R., Hamovitz, T., and Tidwell, J. W. (2016). Reevaluating the effectiveness of n-back training on transfer through the Bayesian lens: support for the null. Psychonomic Bulletin and Review, 23(1), 306–316. doi:10.3758/s13423-015-0865-9
- Simons, D. J., Boot, W. R., Charness, N., Gathercole, S. E., Chabris, C. F., Hambrick, D. Z., and Stine-Morrow, E. A. L. (2016). Do "brain-training" programs work? Psychological Science in the Public Interest, 17(3), 103–186. doi:10.1177/1529100616661983
- Federal Trade Commission (2016). Lumosity to pay $2 million to settle FTC deceptive advertising charges for its "brain training" program. ftc.gov
- Ericsson, K. A., Chase, W. G., and Faloon, S. (1980). Acquisition of a memory skill. Science, 208(4448), 1181–1182. doi:10.1126/science.7375930
- Ericsson, K. A., and Kintsch, W. (1995). Long-term working memory. Psychological Review, 102(2), 211–245. doi:10.1037/0033-295X.102.2.211
- Cowan, N. (2001). The magical number 4 in short-term memory. Behavioral and Brain Sciences, 24(1), 87–114. doi:10.1017/S0140525X01003922
- Morrison, A. B., and Chein, J. M. (2011). Does working memory training work? The promise and challenges of enhancing cognition by training working memory. Psychonomic Bulletin and Review, 18(1), 46–60. doi:10.3758/s13423-010-0034-0
- Murre, J. M. J., and Dros, J. (2015). Replication and analysis of Ebbinghaus' forgetting curve. PLOS ONE, 10(7), e0120644. doi:10.1371/journal.pone.0120644
- Cepeda, N. J., Pashler, H., Vul, E., Wixted, J. T., and Rohrer, D. (2006). Distributed practice in verbal recall tasks: a review and quantitative synthesis. Psychological Bulletin, 132(3), 354–380. doi:10.1037/0033-2909.132.3.354
- Cepeda, N. J., Vul, E., Rohrer, D., Wixted, J. T., and Pashler, H. (2008). Spacing effects in learning: a temporal ridgeline of optimal retention. Psychological Science, 19(11), 1095–1102. doi:10.1111/j.1467-9280.2008.02209.x
- Dunlosky, J., Rawson, K. A., Marsh, E. J., Nathan, M. J., and Willingham, D. T. (2013). Improving students' learning with effective learning techniques. Psychological Science in the Public Interest, 14(1), 4–58. doi:10.1177/1529100612453266
- Lim, J., and Dinges, D. F. (2010). A meta-analysis of the impact of short-term sleep deprivation on cognitive variables. Psychological Bulletin, 136(3), 375–389. doi:10.1037/a0018883
- Beilock, S. L., and Carr, T. H. (2005). When high-powered people fail: working memory and choking under pressure in math. Psychological Science, 16(2), 101–105. doi:10.1111/j.0956-7976.2005.00789.x