Mind Dojo
Learn

What is working memory — and how much does it really hold?

It's the mental workspace you're using to read this sentence: a small, fast, ruthlessly limited buffer. Psychologists have measured its size, argued about the famous number seven, and spent two decades testing whether you can expand it. Here's what they found.

Published 2026-08-10 · Sources verified

The scratchpad model

The modern concept of working memory dates to 1974, when Alan Baddeley and Graham Hitch showed that short-term memory isn’t one passive store but a working system: a central executive directing limited attention, plus specialized buffers for verbal material and for visual-spatial material.[1] That last distinction matters for games: remembering a spoken phone number and remembering a path across a grid use different machinery — which is why a sequence game and a path-tracing game feel so different to play.

Seven was always a legend. The real number is four.

The most famous number in psychology — George Miller’s “seven, plus or minus two” — was coined in 1956 for how many chunks fit in immediate memory span.[2] Miller himself framed the number half-rhetorically, and later research tightened it considerably: when experiments prevent rehearsal and chunking tricks, the capacity of the focus of attention comes out around three to five chunks, with a central tendency of about four.[3] The trick your brain uses to beat that limit is chunking — recoding many items into fewer, richer units. It’s why “FBI CIA NASA” is easy and “AFC BIN AAS” is hard, and why long runs in a memory game suddenly collapse: the level where your chunking strategy stops working is the level where your run ends.

Why psychologists care so much

Working memory earned its central place in cognitive science because capacity predicts things. In a classic latent-variable study, working memory capacity — but not simple short-term storage — strongly predicted performance on fluid reasoning tests.[4] That’s a correlation across people, not a promise that raising one raises the other; people with more workspace tend to reason better. The distinction becomes important two sections from now.

How it’s measured

Verbal span is measured by repeating back sequences of digits or letters. Visuospatial span has its own classic: the Corsi block-tapping task, in use since the 1970s, where an examiner taps blocks in sequence and you tap them back — standardized with normative data in 2000.[5] If that description sounds familiar, it should: watch-then-reproduce sequence games are direct descendants of these span tasks, with the difficulty ramp automated. Your best level is, quite literally, a span measurement.

Can you expand it?

In 2008 a PNAS study made headlines with the claim that weeks of “dual n-back” training improved fluid intelligence itself.[6] The follow-up is a cautionary tale in how science self-corrects: a placebo-controlled replication with seventeen outcome measures found n-back scores improved — and nothing else did.[7] A 2016 meta-analysis of 87 publications settled the question for most researchers: training reliably improves performance on tasks similar to the trained one, and does not raise intelligence.[8]

So train it for what it is

The honest pitch for memory training isn’t “get smarter” — it’s the same as the honest pitch for a grip strengthener: a specific capacity, trained directly, with visible progress. Trained spans genuinely grow. Your chunking strategies genuinely sharpen. And a level counter tells you exactly where your edge is today. That’s what the memory discipline is built around — and if you want the full story on what training does and doesn’t do, read our evidence summary.

References

  1. [1]Baddeley, A. D., & Hitch, G. (1974). Working memory. The Psychology of Learning and Motivation, Vol. 8, 47–89. https://doi.org/10.1016/S0079-7421(08)60452-1
  2. [2]Miller, G. A. (1956). The magical number seven, plus or minus two: Some limits on our capacity for processing information. Psychological Review, 63(2), 81–97. https://doi.org/10.1037/h0043158
  3. [3]Cowan, N. (2001). The magical number 4 in short-term memory: A reconsideration of mental storage capacity. Behavioral and Brain Sciences, 24(1), 87–114. https://doi.org/10.1017/S0140525X01003922
  4. [4]Engle, R. W., Tuholski, S. W., Laughlin, J. E., & Conway, A. R. A. (1999). Working memory, short-term memory, and general fluid intelligence: A latent-variable approach. Journal of Experimental Psychology: General, 128(3), 309–331. https://doi.org/10.1037/0096-3445.128.3.309
  5. [5]Kessels, R. P. C., et al. (2000). The Corsi Block-Tapping Task: Standardization and normative data. Applied Neuropsychology, 7(4), 252–258. https://doi.org/10.1207/S15324826AN0704_8
  6. [6]Jaeggi, S. M., Buschkuehl, M., Jonides, J., & Perrig, W. J. (2008). Improving fluid intelligence with training on working memory. Proceedings of the National Academy of Sciences, 105(19), 6829–6833. https://doi.org/10.1073/pnas.0801268105
  7. [7]Redick, T. S., 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. https://doi.org/10.1037/a0029082
  8. [8]Melby-Lervåg, M., Redick, T. S., & Hulme, C. (2016). Working memory training does not improve performance on measures of intelligence or other measures of “far transfer”. Perspectives on Psychological Science, 11(4), 512–534. https://doi.org/10.1177/1745691616635612

Put it into practice

Every claim above is about trained skills — here’s where you train them.

Play free — no signupBrowse the games