The empirical literature on memory consolidation has, for over a century, converged on a single robust finding: distributed practice — that is, studying material across multiple spaced sessions rather than in a single concentrated period — produces substantially superior long-term retention compared to massed practice. Hermann Ebbinghaus first documented this effect in 1885; subsequent research across cognitive psychology, educational psychology, and neuroscience has replicated and extended the finding repeatedly.
Anki is the most widely-used software implementation of these principles among learners. It applies algorithms (originally SM-2, more recently FSRS) to schedule review of individual facts at intervals computed to maximize retention per minute of study time.
This guide addresses the operational questions that a new user must resolve to use the system effectively: how to construct cards, how to manage daily volume, how to recover from periods of disuse, and how to avoid the design errors that cause the system to fail in practice.
What spaced repetition actually does
The core insight, dating to Hermann Ebbinghaus’s 1885 forgetting curve experiments, is that memory decays predictably. A piece of information you learn today will be ~50% forgotten by tomorrow if untouched, and nearly 0% remembered after a week.
But every successful recall resets the curve and slows down the next decay. Re-test on day 1, and the curve flattens. Re-test on day 3, flatter. Day 7, flatter still. Eventually a single annual review keeps something fresh forever.
Spaced repetition algorithms (SM-2 in Anki, FSRS in newer versions) calculate the optimal next review interval based on how easily you recalled it last time:
- Forgot it → review again in a few minutes
- Recalled with effort → review in 1-3 days
- Recalled easily → review in a week, then a month, then 3 months, etc.
The result: with about 5-15 minutes per day of review, you can keep tens of thousands of facts in your long-term memory.
What Anki is good for
Excellent fit:
- Vocabulary (foreign language, medical terminology, legal terms)
- Formulas (math, physics, chemistry)
- Anatomy / structure / locations
- Definitions and key concepts
- Historical dates and people
- Translation pairs
Bad fit:
- Procedural skills (writing essays, debugging code, playing piano) — practice these instead
- Open-ended understanding (“why did the Roman empire fall?”) — read about it, write summaries
- Material that changes meaning in context — flashcards over-decontextualize
Mixed fit:
- Equations you need to derive, not just recognize — Anki helps with the recognition piece, not the derivation
Daily volume: how many cards?
This is where most people fail. They start ambitious, add 100 new cards/day, the reviews pile up, and within 3 weeks they’re staring at 800 due reviews and quit.
Sustainable volume:
| Goal | New cards/day | Total time/day |
|---|---|---|
| Casual learning | 5-10 | 5-10 min |
| Serious student | 15-25 | 15-25 min |
| Intense study (med school, law) | 25-40 | 30-60 min |
| Anything above 40 | Burn-out within weeks | Don’t do it |
Each new card creates ~5-10 future reviews (over weeks/months) before it stabilizes at long intervals. So 20 new cards today = roughly 100-200 future reviews scheduled across the next year.
Critical rule: never let your daily review queue exceed what you can do in 25 minutes. If reviews exceed that, lower your new-card rate, not just push harder.
Card design: the make-or-break
Bad cards cause everything else to fail. Three core principles:
1. One fact per card. If a card has multiple things to remember, you’ll get it wrong on whichever you forget — and never learn the others reliably.
Bad:
Front: “What is photosynthesis?”
Back: “The process by which plants convert sunlight, water, and CO2 into glucose and oxygen, occurring in chloroplasts in two stages: light reactions in the thylakoid and Calvin cycle in the stroma.”
Better, split into 6+ cards:
- Front: “Photosynthesis takes , , and ___ as inputs” → “sunlight, water, CO2”
- Front: “Photosynthesis produces ___ and ” → “glucose, oxygen”
- Front: “Photosynthesis occurs in which organelle?” → “chloroplasts”
- Front: “Stage 1 of photosynthesis is called ” → “light reactions”
- …
2. Minimum information principle. The smaller the unit, the easier to recall, the more reliable the long-term memory.
3. Use cloze deletion liberally. Anki has a {{c1::word}} format that hides one word at a time. Excellent for sentences, definitions, lists.
Example:
”{{c1::Hippocampus}} is the brain region most associated with forming new {{c2::declarative}} memories.”
This generates two cards from one note: one hiding “hippocampus”, one hiding “declarative.”
Image cards
For anatomy, geography, or anything spatial, image occlusion is the killer feature. You upload an image, drag rectangles over the labels, and Anki hides one label at a time — testing you on each.
Example uses:
- Brain anatomy with regions hidden one at a time
- World map with countries hidden
- Organic chemistry molecule diagrams
- Software UI screenshots for tool memorization
The Anki addon “Image Occlusion Enhanced” adds this. Download separately (anki-addons.com), takes 2 minutes to install.
When to make cards vs. learn first
A common mistake: making cards on material you don’t yet understand. The card becomes a chore because the answer doesn’t make sense to you, only sounds memorized.
Rule: Read and understand the material first (lectures, textbooks, explanations). Then make cards on the parts you specifically need to memorize.
Cards aren’t for learning — they’re for not forgetting what you’ve already learned. If a topic isn’t clear to you, no flashcard will fix that.
The daily routine
Morning (recommended):
1. Open Anki
2. Do reviews first (the older cards that are due today)
3. Then add new cards if you have time
4. Total: 15-30 minutes
Why morning: willpower is highest, and the slight cognitive load of doing flashcards “warms up” your brain for harder study work later. Combine with the morning study routine.
Evening (alternative):
- Reviewing within 1-2 hours of sleep significantly boosts retention via overnight consolidation.
- Drawback: late-evening reviews can disrupt sleep if too cognitively demanding.
Don’t do this:
- Batched on weekends only — the daily rhythm is what makes it work
- All in one 2-hour Sunday session — defeats the spacing
- Skipping days — accumulates a pile that becomes psychologically painful
What to do when reviews pile up
If you missed a week or two and now have 500+ due cards, the recovery protocol is:
- Stop adding new cards immediately. Until your due queue is back to ~50/day.
- Do reviews in 25-minute Pomodoros. Don’t try to clear it all in one sitting.
- Be liberal with “Hard” responses. If a card is barely recallable, mark Hard rather than Again. Else you’ll bury yourself in re-reviews.
- Use Anki’s “Review ahead” if you’re planning to be away — preview what’s due in the next few days while you have time.
If you have 1000+ cards backlog, consider using Reset and Reschedule on the deck, which redistributes the dates and prevents an avalanche on a single day.
The most common mistakes
Mistake 1: Cards that are too long. If the back of a card has 3+ sentences, split it into 3+ cards. Always.
Mistake 2: Pre-made decks vs. self-made. Pre-made decks (downloaded from AnkiWeb) almost never stick. The act of making the card is part of the learning. Make your own, even if pre-made exists. Use pre-made only as raw material to selectively rewrite.
Mistake 3: Adding too many at once. Better to add 10 cards/day for 30 days (300 cards mastered) than 50 cards in a weekend (300 cards added, mostly forgotten by month’s end).
Mistake 4: No context. A card that asks “what is X?” with no context (subject area, chapter) is hard to remember. Add a tag or short context cue: “Bio101: What is X?”
Mistake 5: Reviewing in environments full of distractions. Anki on your phone in a busy café = poor encoding. Anki at a quiet desk with lofi music = strong encoding. The same card studied in different conditions has very different retention.
Compound effect over years
The thing that makes Anki almost magical: it compounds. After 1 year of consistent 15-minute daily reviews, you have 5,000-10,000 facts in long-term memory. After 2 years: 15,000-25,000. After 5 years of medical/law student-level use: 50,000+.
Most people couldn’t recall 50,000 facts even with infinite study time, because the forgetting curve outpaces their re-study schedule. With Anki it’s normal.
Pairing with other study methods
Anki is one tool among several. Use it alongside:
- Pomodoro — do Anki at the start of a Pomodoro to warm up, then deep work.
- Note-taking — take Cornell notes from a lecture, then turn the cue column items into Anki cards.
- Active recall practice — Anki is active recall, but for raw facts. Other forms (essay writing, problem solving) cover application.
- Sleep — the consolidation that happens during sleep is where Anki reviews really stick. Don’t undermine it with all-nighters.
Final practical advice
If you’ve never used Anki and want to start: install Anki Desktop (free, ankiweb.net) plus the AnkiMobile app ($25 one-time on iOS — there’s a free Android equivalent called AnkiDroid). Sync your deck across devices.
Set new cards/day to 10. Review queue limit to 100. Do reviews every morning before anything else for 7 days. After a week, you’ll know whether the system fits your study style.
For most students, the first 2-3 weeks feel slow — the cards are easy, reviews are short. That’s the trap of underestimating the system. By month 3, you’ll be reviewing material from week 1 and remembering it with no extra effort. That compound is the entire point.
Combine with calm study environment and study music, and you have the highest-leverage memory system available to a student today.
References
The science backing spaced repetition:
- Karpicke, J. D., & Roediger, H. L. (2008). The Critical Importance of Retrieval for Learning. Science, 319(5865), 966-968. The foundational paper showing testing dramatically beats re-reading.
- Cepeda, N. J., Vul, E., Rohrer, D., Wixted, J. T., & Pashler, H. (2008). Spacing effects in learning: A temporal ridgeline of optimal retention. Psychological Science, 19(11), 1095-1102.
- Bjork, R. A., & Bjork, E. L. (2011). Making things hard on yourself, but in a good way. Psychology and the real world, 56-64. The “desirable difficulties” framework that explains why retrieval practice works.
