Skip to content 
If you’ve ever crammed for an exam and then forgotten everything a week later, you’ve experienced the central problem that memory researchers have been studying for over a century. The uncomfortable truth is that most common study habits—rereading notes, highlighting, marathon study sessions—feel effective but produce weak, fleeting memories. Meanwhile, the techniques that actually work feel harder and less satisfying in the moment.
The research is remarkably clear: spacing your learning over time produces 15–40% better retention than cramming, and testing yourself beats rereading by a wide margin (effect sizes of d = 0.5–0.8).[1][2] When you combine these two approaches—spaced practice with active self-testing—you get a system that’s nearly optimal for building durable knowledge. This is exactly what flashcard apps like Anki implement, and it’s why medical students, language learners, and competitive memory athletes swear by them.
What makes this particularly frustrating is that the brain systematically misjudges which study methods work. This isn’t a minor bias—it’s a robust finding replicated across decades of research. Zechmeister and Shaughnessy documented the paradox as early as 1980: despite spacing dramatically improving recall, participants’ confidence ratings failed to reflect this benefit.[3] The pattern extends beyond verbal memory to motor learning, where trainees consistently prefer blocked practice even as their retention suffers.[4] In one striking demonstration, 90% of participants learned more from spaced practice, yet 72% believed that cramming had been more effective—a judgment they made after experiencing both conditions.[5] The illusion persists even in complex inductive learning: participants who studied paintings in spaced versus massed conditions performed better with spacing, yet most still rated massing as more effective for learning new artistic styles.[6] Your intuitions about learning are often wrong—and this report explains why.
The Spacing Effect: Why Cramming Fails You
When Hermann Ebbinghaus first mapped the “forgetting curve” in 1885, he discovered something that every student learns the hard way: memories decay rapidly after initial learning, with most forgetting happening in the first hours and days. What’s less intuitive is his other discovery—that spacing out your review sessions dramatically slows this decay.
The effect is robust across nearly every domain researchers have tested. Nurse anesthesia students randomized to spaced learning showed significantly higher retention across three semesters compared to traditional instruction, with large effect sizes (η ≈ 0.58–0.60).[1] Dental students using mobile flashcards with spaced repetition performed similarly to controls immediately after lectures, but showed substantially better retention at one and three months.[7] Even surgical residents learning suturing skills performed better when they practiced in spaced sessions rather than one continuous block.[8]
Why does spacing work so well? The key insight is that forgetting is not the enemy—it’s the teacher. When you space out your practice, you’re forced to reconstruct the memory trace rather than simply keeping it warm in working memory. This effortful reconstruction strengthens the memory in ways that easy, massed repetition cannot. Neuroimaging studies confirm this: spaced learning reduces a phenomenon called “neural repetition suppression” and increases pattern similarity across brain regions involved in memory, both of which predict better long-term retention.[9][10]
The Testing Effect: Quizzes Are Better Than Rereading
The second pillar of effective learning is counterintuitive: testing yourself produces better memory than additional studying, even when you don’t get feedback and even when the test counts for nothing. Psychologists call this the “testing effect” or “retrieval practice effect,” and it’s one of the most replicated findings in cognitive science.
The mechanism appears to involve the act of retrieval itself. When you try to pull information from memory—rather than passively reviewing it—you engage a different set of neural processes. fMRI studies show that retrieval practice strengthens representations in the hippocampus and increases connectivity between memory regions in ways that simple restudy does not.[2] Behaviourally, both younger and older adults show robust testing effects over retention intervals of a week or more.[11]
There’s also a metacognitive benefit: testing gives you accurate information about what you actually know versus what you merely recognize. Students who engage in regular self-testing become better calibrated—they’re less likely to overestimate their knowledge (the Dunning-Kruger problem) and better able to allocate their study time to weak areas.[12]
The practical implication is clear: replace a substantial portion of your “review time” with testing. Instead of rereading notes, close the book and try to recall the key points. Instead of highlighting passages, quiz yourself on them later. The effort feels less pleasant, but it produces dramatically better results.
The Sweet Spot: Combining Spacing and Testing
When researchers pit different study conditions against each other—massed versus spaced practice, testing versus restudy—one combination consistently wins: spaced testing. Experiments that manipulate both factors find a clear hierarchy: spaced testing outperforms spaced restudy, which outperforms massed testing, which outperforms massed restudy.[13]
This is precisely what modern spaced repetition software implements. Every review session is a test—you see a prompt and try to recall the answer before checking. The software then schedules your next review based on how well you performed, using expanding intervals that grow longer as items become more firmly established in memory. It’s not a fancy trick; it’s applying the two most robust findings in memory research simultaneously.
How much effort does this require? Estimates suggest that a typical flashcard costs about 5 minutes of total review time over 20 years, with the bulk of that time front-loaded in the first few months.[14] This leads to a useful rule of thumb: if memorizing something will save you more than 5 minutes of lookup time over your lifetime, it’s worth adding to a spaced repetition system.
The 7-Opportunity Rule: How Long Mastery Really Takes
One of the most striking findings from recent large-scale learning research is how consistent the path to mastery turns out to be. Analysing data from nearly 7,000 students across 27 educational technology datasets, researchers found that students need approximately 7 practice opportunities to move from initial learning (~65% accuracy) to reasonable mastery (~80% accuracy).[15]
What’s remarkable is the uniformity in learning rate. Despite enormous variation in where students start—some begin at 55% accuracy, others at 75%—their rate of improvement per opportunity is nearly identical: about 2.5% per practice trial. The students who need 13 opportunities to reach mastery aren’t learning more slowly; they’re starting from further behind. This suggests that the variation we see in academic performance is largely about accumulated prior learning, not fundamental differences in learning capacity.
The practical takeaway is that mastery requires more practice than most people expect. Lectures and readings get you to about 65% accuracy—a failing grade. The remaining 20 percentage points to mastery require dedicated practice, ideally spaced and tested.
Why Your Brain Lies to You About Studying
Perhaps the cruelest aspect of all this research is that the study methods that feel most effective are often the least effective, and vice versa. Rereading feels productive because the material becomes increasingly familiar—but familiarity is not the same as retrievable knowledge. Cramming feels efficient because you can demonstrate short-term gains—but those gains evaporate rapidly. Spaced practice and testing feel frustrating precisely because they’re harder—but that difficulty is the signal that real learning is occurring.
Researchers call this the “illusion of fluency.” When information comes to mind easily during study, you predict you’ll remember it later. But easy retrieval during study often reflects recency or repetition, not durable encoding. The students who reported that cramming worked better were simply wrong about their own learning.[5]
Understanding this can help you tolerate the discomfort of effective learning. When you struggle to recall something during a practice test, that struggle is productive—it’s strengthening the memory trace. When you feel confident after rereading your notes, that confidence is probably misplaced.
Practical Implementation
Translating these findings into practice doesn’t require fancy software, though tools like Anki can help. The key insight from experienced practitioners: Anki makes memory a choice—it transforms remembering from a haphazard event into a deliberate decision.[16]
When to Add a Card
Apply the 10-minute rule: if memorizing a fact will save you more than 10 minutes of lookup time over your lifetime, add it.[16][14] But also trust your intuition—if something feels striking or important even when you can’t justify why, add it anyway. The most valuable knowledge is often what you didn’t know would matter.
Critically, tie cards to active projects. Ankifying material for a creative project you’re emotionally invested in produces far better retention than stockpiling knowledge “just in case.” Abstract goals like “learn African geography” generate cold, lifeless cards that fail to stick. Questions born from genuine curiosity connect to your existing knowledge network.
How to Craft Cards
Make cards atomic. Each card should test exactly one idea. If you’re routinely getting a card wrong, break it into smaller pieces. A question like “How do I create a Unix soft link?” becomes two cards: “What command creates a soft link?” (ln -s) and “In what order do the arguments go?” (filename linkname).[16]
Avoid sets and lists. Asking “What countries are in the EU?” is a memorization nightmare because you’ll retrieve members in different orders each time, creating interference. Convert essential lists into ordered enumerations or overlapping cloze deletions.[17]
Prevent orphan questions. Single cards disconnected from your other knowledge decay quickly. Aim for at least 5 questions per topic—fewer suggests the material isn’t worth memorizing at all. If you can’t generate 5 good questions from a paper, add none.[16]
Build your own deck. Creating cards is itself an act of elaborative encoding—you’re forced to think through alternate phrasings, best answers, and connections to existing knowledge. Using shared decks forfeits this benefit. Merge everything into one big deck; the collision of unrelated topics can stimulate creative connections.[16]
Qualify uncertain claims. When Ankifying research findings, phrase questions to reflect the source: “What does Jones 2011 claim about Nobel physicist ages?” rather than treating it as established fact. You’re memorizing that a paper made a claim, not that the claim is true.[16]
Avoiding Common Traps
Don’t chase optimization rabbit holes. Anki offers roughly a 20-fold improvement over ordinary flashcards. People who abandon it because they weren’t using it “perfectly” are giving up a 2,000% improvement over worries about missing 1% gains.[16] Start simple, stay consistent.
References
[1] “Enhancing learning outcomes in nursing education: The role of spaced learning.” Nurse Education Today 2024. https://pmc.ncbi.nlm.nih.gov/articles/PMC10958887/
[2] Agarwal PK, et al. “Retrieval practice facilitates learning by strengthening processing in both the anterior and posterior hippocampus.” Cerebral Cortex 2021. https://pmc.ncbi.nlm.nih.gov/articles/PMC7821628/
[3] Zechmeister EB, Shaughnessy JJ. “When you know that you know and when you think that you know but you don’t.” Memory & Cognition 1980;8(6):549-559. https://link.springer.com/article/10.3758/BF03329756
[4] Simon DA, Bjork RA. “Metacognition in motor learning.” Journal of Experimental Psychology: Learning, Memory, and Cognition 2001;27(4):907-912. https://pubmed.ncbi.nlm.nih.gov/11486923/
[5] Kornell N, Bjork RA. “Optimising learning using flashcards: Spacing is more effective than massing.” Applied Cognitive Psychology 2008;22(9):1209-1220. https://onlinelibrary.wiley.com/doi/abs/10.1002/acp.1537
[6] Kornell N, Bjork RA. “Learning concepts and categories: Is spacing the ‘enemy of induction’?” Psychological Science 2008;19(6):585-592. https://journals.sagepub.com/doi/abs/10.1111/j.1467-9280.2008.02127.x
[7] “Impact of spaced repetition flashcards on dental student retention.” Journal of Dental Education 2024. https://onlinelibrary.wiley.com/doi/10.1002/jdd.13561
[8] “Spaced vs. massed practice in surgical skill acquisition.” Annals of Medicine 2024. https://www.tandfonline.com/doi/full/10.1080/07853890.2024.2363940
[9] “Neural repetition suppression and spaced learning.” NeuroImage 2012. https://pmc.ncbi.nlm.nih.gov/articles/PMC3297428/
[10] “Pattern similarity during spaced retrieval.” Journal of Cognitive Neuroscience 2019. https://pmc.ncbi.nlm.nih.gov/articles/PMC6607761/
[11] “Age-comparative effects of retrieval practice.” Memory & Cognition 2020. https://pmc.ncbi.nlm.nih.gov/articles/PMC6990689/
[12] “Practice testing improves metacognitive calibration.” Educational Psychology Review 2023. https://pmc.ncbi.nlm.nih.gov/articles/PMC10045345/
[13] “Desirable difficulties in learning.” Perspectives on Psychological Science 2018. https://pmc.ncbi.nlm.nih.gov/articles/PMC6289840/
[14] Branwen G. “Spaced Repetition.” Gwern.net 2009–2024. https://gwern.net/spaced-repetition
[15] Koedinger KR, et al. “An Astonishing Regularity in Student Learning Rate.” PsyArXiv 2024. https://osf.io/preprints/psyarxiv/pxsfh_v1
[16] Nielsen M. “Augmenting Long-term Memory.” 2018. https://augmentingcognition.com/ltm.html
[17] Wozniak P. “Twenty rules of formulating knowledge.” SuperMemo 1999 (updated). https://www.supermemo.com/en/blog/twenty-rules-of-formulating-knowledge
