Direct answer: what counts as a good number memory score
A good number memory score is 9 or more digits. The average run ends at 7, so reaching 9 already puts you ahead of most people who take the test, and 10 or more is uncommon without a deliberate memorization strategy.
The reason the bar sits there is well-established. The average adult's short-term memory span for digits is about seven, give or take two — the finding George Miller summarized in 1956 as the "magical number seven, plus or minus two." A typical run dies at 7 because that is roughly where raw, un-rehearsed digit span runs out. Scores in the double digits almost always reflect chunking, not a larger raw buffer.
Number memory score ranges, explained
Use the table below to read your result. The "interpretation" reflects published memory-span norms and the typical shape of online number-memory results; the percentile column is approximate and will be replaced with Deadline's own measured distribution as our run database grows.
| Digits recalled | Tier | What it means | Approx. percentile |
|---|---|---|---|
| 1–5 | Below average | Below typical digit span; often a slip, distraction, or unfamiliarity with the test format | Bottom ~15% |
| 6–8 | Average | The normal short-term span for digits (7 ± 2); where most runs end | ~25th–70th |
| 9 | Good | Above average; beats the majority of players | ~Top 10–20% |
| 10–11 | Strong | Uncommon on a single attempt; usually involves grouping digits | ~Top 5% |
| 12+ | Exceptional | Rare without trained chunking or mnemonic strategy | ~Top 1–2% |
| 20+ | Elite / trained | Effectively always a learned memory technique, not raw span | Far edge |
Why most people stop at 7 digits
Miller's 1956 paper proposed that immediate memory holds about seven "chunks" of information. He was explicit that there was nothing magical about the number — he used the phrase rhetorically, and noted the limit depends on how information is grouped, not on a hard slot count.
Later work tightened the estimate. Nelson Cowan's 2001 review argued that when you prevent people from grouping items, the true capacity is closer to four (roughly three to five) chunks. Both figures describe the same underlying fact: there is a small, fixed limit on what you can hold in mind at once. On a number memory test, you experience that limit as the level where the digits suddenly won't stick.
How number memory relates to digit span
Number memory is a forward digit span task: you see a string of digits and repeat it back in the same order. In clinical testing, the average adult forward digit span is also about 7 ± 2 (roughly 5–9 digits), which is why scores on the two tests line up.
Backward digit span — repeating the digits in reverse — is harder, averaging about two digits fewer, because it requires actively manipulating the sequence rather than just holding it. That distinction matters: forward span is closer to a measure of attention and storage, while backward span leans on working memory. If you want the full breakdown, see the digit span test.
- Forward span (number memory): hold and repeat — average ~7
- Backward span: hold and reverse — average ~5
- A large forward–backward gap is what clinicians watch for, not the raw number
Can you improve your number memory score?
Yes, but be honest about what's improving. You are not expanding the raw buffer — you are learning to compress digits into fewer, larger chunks. Grouping a string into pairs or triples (like reading a phone number) lets you carry the same load using fewer mental slots.
The ceiling for trained chunking is high. In a landmark 1980 study, Ericsson, Chase, and Faloon trained a single participant who used his knowledge of running times to recode digit strings; over about 20 months of practice he raised his span from 7 to roughly 79 digits. His span collapsed back toward normal the moment researchers fed him sequences that couldn't be mapped to running times — proof that the gain was strategy, not a bigger buffer.
For a single honest attempt, the practical takeaway is simpler: a quiet room, full attention, and light grouping will move most people from average to good. Anything past ~10 on demand means you've built a system.
Deadline first-party data (coming soon)
Every percentile above is drawn from published memory-span research and the general shape of online number-memory results. We are building something the textbook figures can't give you: a live distribution of real runs from this test.
This section will be replaced with Deadline's own median, percentile bands (p10 / p25 / p50 / p75 / p90), and age-banded breakdowns once we have a representative sample — original first-party data, clearly labeled with the number of runs behind it. Take the test now and your result becomes part of that distribution.
Frequently asked questions
Is 9 digits a good number memory score? Yes. The average run ends at about 7 digits, so reaching 9 puts you ahead of most players — roughly the top 10–20%. It's a clearly above-average result on a single honest attempt.
What is the average number memory score? About 7 digits. This matches the classic short-term memory span of 7 ± 2 (Miller, 1956) and the average adult forward digit span of roughly 5–9 digits.
Is 10 digits good on the number memory test? Yes — 10 or more is uncommon on a single attempt and usually means you grouped the digits into chunks rather than holding them raw. It places you in roughly the top 5%.
What is the highest possible number memory score? There is no fixed ceiling. Raw span tops out near 7–9, but with trained chunking the limit is enormous: in a 1980 study one trained participant reached about 79 digits. Memory-sport competitors recall far longer strings using mnemonic systems.
Does a high number memory score mean a high IQ? Not directly. Digit span contributes to the working memory part of IQ tests, so it correlates loosely, but a single number-memory run is not an IQ score. A very high score usually reflects a learned chunking strategy, not raw intelligence.
Related on Deadline
number memory test · digit span test · memory tests · what is a good reaction time · cognitive benchmark
Sources and notes
- 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://en.wikipedia.org/wiki/The_Magical_Number_Seven,_Plus_or_Minus_Two
- 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://www.cambridge.org/core/journals/behavioral-and-brain-sciences/article/magical-number-4-in-shortterm-memory-a-reconsideration-of-mental-storage-capacity/44023F1147D4A1D44BDC0AD226838496
- Ericsson, K. A., Chase, W. G., & Faloon, S. (1980). Acquisition of a memory skill. Science, 208(4448), 1181–1182. https://en.wikipedia.org/wiki/Chunking_(psychology)
- Memory span — Wikipedia. https://en.wikipedia.org/wiki/Memory_span
- Digit Span — ScienceDirect Topics. https://www.sciencedirect.com/topics/psychology/digit-span