The best Anki settings for medical students are settings that produce useful retention within a workload you can sustain. They are not a secret configuration that turns a large deck into effortless examination preparation. Before changing anything, decide how much time flashcards should occupy and which other learning activities must remain protected.
A student in a preclinical teaching block, a finalist on placement and a doctor revising around shifts face different constraints. Their settings may reasonably differ even when they use the same deck. Your objective is a manageable system, not numerical resemblance to somebody else's screenshot.
What FSRS changes
FSRS uses review history to model memory and schedule future reviews. Anki also documents its older SM-2-based scheduling approach; describing all current Anki use as SM-2 is therefore inaccurate. The official algorithm explanation is a better starting point than a settings video made for a different release.
The practical implication is that scheduling should respond to your performance rather than depend entirely on hand-tuned interval assumptions. It does not mean the software can decide which medical facts belong in your curriculum, judge whether a card is clinically current or certify your readiness for an examination.
Check the instructions for your installed version before following menu-by-menu advice. Interface details and available options can change. When a tutorial and the current manual disagree, investigate the discrepancy rather than applying both sets of instructions simultaneously.
Desired retention is a trade-off, not a grade
Desired retention concerns the probability of recalling scheduled material when it comes up for review. It is not a predicted examination score, a percentage of medicine mastered or a measure of clinical competence. Raising it generally means accepting more frequent reviews.
The deck-options documentation describes a default of 90% and explains the increasingly steep workload associated with higher targets. That default can be a sensible starting reference, but it is not a personalised prescription for every student or every collection.
Choose a target with the rest of your timetable in view. An apparently attractive improvement in card recall may be poor value when it removes the only available hour for clinical questions. Conversely, a low workload is not automatically efficient when essential facts repeatedly disappear before you can apply them.
Answer buttons must reflect actual recall
A scheduling algorithm can only work with the information you give it. Do not rate a forgotten answer as difficult-but-remembered simply because it feels familiar after disclosure. A card that required a substantial guess should not receive the same rating as one you recalled accurately and comfortably.
Use a simple personal marking rule before beginning a session. State or write the answer, reveal it, then compare the two. Decide in advance how much precision a card requires. If reasonable answers repeatedly fail to match the back of the card, the wording may need repair.
For complex medicine, distinguish a genuinely missing fact from an incomplete question. A card asking for "the treatment" without specifying clinical context is not rescued by more careful button pressing. Correct the learning material before trying to optimise its schedule.
Control new cards before chasing advanced settings
New-card intake is a commitment to future reviews. A large burst can seem manageable on the first day while creating work that collides with placements or mock examinations later. Increase intake gradually and observe the consequences over ordinary weeks, not just unusually quiet weekends.
Use your actual available study time as the constraint. For example, a student allocating half an hour to flashcards should not activate material based on a plan that assumes two hours. The exact number of cards will vary with their difficulty and the time needed to mark them properly.
Avoid treating a low maximum-review limit as proof that the workload has disappeared. It may simply leave due work out of view. Review the whole collection and your recent performance when deciding whether to reduce new material, suspend low-priority content or change the study plan.
Optimise carefully and protect your history
Before a substantial scheduling change, make a backup and understand whether you are changing future scheduling or immediately rescheduling existing cards. Anki's exporting guidance explains how collection exports and scheduling information relate to recoverability.
Do not alter several major variables at once. A manageable experiment might change one setting, preserve new-card intake and review the resulting workload before making another adjustment. Otherwise, you will struggle to identify why the queue became easier or harder.
Look at patterns over a meaningful period rather than reacting to one tired evening. The Anki statistics guidance is useful for interpreting retention and review data. These figures describe your interaction with the collection; they should sit beside, not replace, evidence from unfamiliar questions.
Give settings a clinical reality check
Once a week, choose a topic that looks strong in Anki and attempt new examination-style questions without help. Record whether errors reflect forgotten information or difficulty applying what you remembered. Use this as a practical self-check, not a validated assessment of examination readiness.
iatroX question banks can provide the application task, while the Socratic Tutor can help examine your reasoning afterwards. No Anki score needs to be transferred for that sequence to be useful.
When recall is improving but case performance is not, another scheduling adjustment may be the wrong intervention. You may need a clearer explanatory model, more comparison between plausible answers or practice handling uncertainty. Good settings support that work by keeping factual maintenance proportionate.
The companion guide to using Anki without burning out addresses the wider workload decisions that settings alone cannot resolve.
Frequently asked questions
Is 90% retention the best target for everyone?
No. It is a documented default and a useful starting reference. Your workload, material and examination timetable matter. Higher retention usually carries a cost in reviews, so assess that cost against the rest of your learning.
Should I copy an AnKing settings tutorial exactly?
Only after checking its date, software version and assumptions. A tutorial can explain controls, but it cannot know your existing review history, placement commitments or the relevance of every card in your collection.
Will FSRS improve my clinical reasoning?
Its role is scheduling recall practice. Better availability of relevant knowledge may help you reason, but reasoning still needs to be practised directly through unfamiliar questions, discussion and supervised clinical learning.
Test your retained knowledge on an iatroX clinical question.