Anki Review: The Brute Force Algorithm for Memory Sovereignty and Biological Encoding

You crushed it on Sunday. The chapter, the new language, the framework for the meeting — it all clicked, and you closed the laptop feeling sharp. By Thursday it’s fog. You reach for the fact in the meeting and it’s gone, leaving only the humiliating certainty that you knew this. So you re-read. You re-watch. You pay for the course again. And somewhere a product manager smiles, because your forgetting is the business model.

The short version: Anki is a free, open-source flashcard app that uses spaced repetition — the SuperMemo-2 (SM-2) algorithm, with an optional newer one called FSRS — to show you each fact at the precise moment you’re about to forget it, locking it into long-term memory with about 15 minutes of review a day. Unlike Duolingo, Quizlet, or Coursera, it has no streaks, no badges, and no servers holding your data hostage; your decks live as files you own. The trade-off is honesty over comfort: Anki is plain, demands daily consistency, and only works with well-built “atomic” cards. If you want guaranteed retention instead of the illusion of progress, that trade is worth making.

Why does your brain forget — and why do apps profit from it?

You’ve felt it: learn something intensely, then watch most of it evaporate within days. Here’s the reframe that changes the whole picture — you’re not bad at remembering; you’re using a tool built to make you forget. Your brain isn’t broken. It’s doing exactly what it evolved to do: prune neural pathways it doesn’t see being used, to save energy. Forgetting is the default, not the failure — which means the fix isn’t trying harder, it’s changing when the material comes back in front of you.

Here’s the part that should make you angry. The mainstream learning industry isn’t fighting your forgetting — it’s monetising it. Duolingo’s owl and its streaks, Quizlet’s badges, Coursera’s certificates: these gamify a loop where you feel productive, forget anyway, and come back to learn (and pay) again. The dopamine is real. The retention often isn’t.

Anki breaks the loop precisely because it refuses to entertain you. It doesn’t distract you with rewards — it confronts you with what you’re losing and shows you each card at the last moment before it would slip away. No streaks. No hooks. Just the blunt biology of memory, used in your favour instead of against you.

What makes Anki different? The SuperMemo-2 algorithm

Most flashcard apps hide their scheduling logic in a black box. Anki exposes it completely, and that transparency is the whole point. The SuperMemo-2 (SM-2) algorithm decides when to resurface a card based on four inputs:

• Ease factor — how hard you find the card (1.3 = hard, 2.5 = normal, 3.0+ = trivial). Easier cards return less often.
• Interval — days until the next review, spaced roughly exponentially: 1 day, then 4, then 10, then 30, then months.
• Your response (1–4) — you grade your own recall after each card. A “4” pushes the interval out; a “1” resets it.
• Retention target — the percentage you want to remember, default around 90%, and tunable.

Because Anki lets you adjust these variables, you own the algorithm rather than renting access to one tuned for someone else’s ad revenue. You’re optimising for your own long-term memory, full stop.

Which Anki card types actually lock information in?

Anki doesn’t force one format. How you encode a fact changes how well it sticks, and the strongest cards make your brain produce the answer, not just recognise it.

• Basic (question and answer): simple front-and-back. “What is an electron?” → “A negatively charged subatomic particle.”
• Cloze deletion: hide the critical words — “An electron is a {{c1::negatively charged}} subatomic particle.” Forces active retrieval of the missing piece.
• Image occlusion: hide parts of a diagram and name them. The brain evolved to remember places and objects, so visual encoding tends to outlast plain text.
• Type-in-answer: you have to spell it out, which makes the brain work harder than passive recognition does.

Encode one fact across text, image, and audio and it holds far better than any single mode. The cards that feel slightly effortful are the ones that work — easy cards are the ones that quietly fail you.

How to build an Anki system that sticks

A good algorithm can’t rescue bad cards or broken habits. Three phases turn Anki from a clever app into a retention machine.

1. Atomic card design — the input. One card, one fact. “All the causes of anaemia” is not atomic; “Vitamin B12 deficiency causes megaloblastic anaemia” is. When each card holds a single retrievable fact, reviews become 15-second bursts of clean focus instead of vague staring contests.
2. The daily ritual — the maintenance. Anki only works if you review every day. Skip a week and you accrue “review debt” — hundreds of stacked cards that feel crushing enough to make you quit, and the dread of that backlog is what kills most people’s habit before it ever compounds. The simplest fix is a fixed slot: a short Anki session before the day’s noise, capped tight, no exceptions. Attach it to something you already do without fail — coffee, the commute, the first sit-down at your desk — so the trigger is automatic and you never have to decide to start.
3. Synthesis with a second brain — the output. Isolated facts aren’t understanding. Connect Anki to a notes system like Obsidian: Anki for atomic recall, Obsidian for relationships, context, and argument. Together they form a complete stack — fast retrieval plus real comprehension.

Why you’ll hit “ease hell” — and how to escape it

After a few weeks, some cards get stuck. You keep failing them, the ease factor drops, and the algorithm starts showing them every single day. This is “ease hell,” and it’s where most people give up, deciding Anki “doesn’t work for them.”

It does work. The card is the problem. A card you fail repeatedly is almost always a card trying to hold too much — the fix is to split, not to grind. Delete it and rewrite it as two smaller, atomic cards. Instead of “What are the symptoms of diabetes?”, make “What is hyperglycaemia?” and, separately, “What is polydipsia?” Your brain wasn’t equipped to hold the bundle; broken into pieces, each one finally lands.

Does Anki beat subscription flashcard apps? The honest comparison

The gap isn’t about polish — the subscription apps are far prettier. It’s about who owns the system and what it’s optimised for.

• Cost — Anki: free on desktop and Android, a one-time roughly $25 on iOS. Quizlet: free with ads, around $60–100/year for premium. Duolingo: free with ads, roughly $7–12/month.
• Algorithm control — Anki: full customisation (SM-2 and FSRS). Quizlet and Duolingo: proprietary black boxes.
• Data ownership — Anki: 100% local files you control. Quizlet and Duolingo: locked on their servers.
• Card types — Anki: unlimited custom types. The others: limited to flashcards, multiple choice, fill-in.
• Gamification — Anki: none, by design. The others: heavy streaks, badges, leaderboards.

The honest verdict: for fun, low-stakes dabbling, the gamified apps genuinely win on motivation — but for serious, durable retention you control, Anki has no real competitor. Its ugliness is a feature; nothing in it is engineered to keep you scrolling instead of learning. If you need a streak to show up, start on Duolingo and graduate to Anki when you’re ready to actually remember things.

Does Anki really work? What the evidence supports

It’s worth separating the documented mechanism from the hype, because plenty of breathless claims circulate about people memorising entire degrees overnight. The credible part is the engine underneath: spaced repetition is one of the most robustly supported findings in the science of learning. Reviewing material at expanding intervals produces markedly stronger long-term retention than massed study (cramming) or re-reading — that’s the well-replicated “spacing effect,” and SM-2 and FSRS are simply algorithms for scheduling it.

What the evidence does not promise is a specific outcome for you. Anki is a tool, and a demanding one — its results depend on card quality, daily consistency, and whether you pair raw recall with genuine understanding. Treat dramatic “I learned X in half the time” anecdotes as motivation, not data. The realistic claim, and the one worth banking on, is this: used consistently with atomic cards, spaced repetition reliably moves facts into durable memory far more efficiently than re-reading ever will.

Frequently asked questions

How long does it take to see results with Anki?
Most people notice clearer recall within 2–3 weeks of daily reviews. Durable, high-accuracy retention typically takes 3–6 months of consistent use, after which the effect compounds — well-encoded cards resurface less and less often because they’ve genuinely stuck. Consistency matters more than session length.

What’s the difference between Anki and FSRS?
FSRS (Free Spaced Repetition Scheduler) is a newer scheduling algorithm that uses machine learning to predict your memory more precisely than the older SM-2, often delivering similar retention with somewhat less review time. It’s built into modern Anki. If you’re starting fresh, turning it on is reasonable; if you already have a working SM-2 setup, the difference is modest.

Should I use pre-made decks or build my own?
Pre-made decks are useful scaffolding — AnkiWeb hosts a huge library of shared decks — but they’re often poorly designed: cards too complex, sources missing, encoding that doesn’t match how you think. For maximum retention, build your own from sources you trust, and use shared decks only as a starting frame you then customise.

How do I stop Anki from feeling like a chore?
Set a hard cap — 15 to 20 minutes a day, no exceptions. If you’re spending an hour, your cards are too complex; break them down. Review at the same time each day so it becomes automatic, and automaticity removes most of the friction. The chore feeling almost always signals overbuilt cards, not a lack of willpower.

Is Anki better than active-recall study groups?
They do different jobs, so use both. Anki encodes individual facts into long-term memory; study groups build understanding of how those facts relate and apply. Anki alone risks a pile of disconnected trivia. Pair it with discussion and real projects and you get the full picture.

You came here because something you’d genuinely learned vanished when you needed it, and you were tired of paying twice for the same knowledge. Anki won’t make that feeling disappear with a clever interface — it has no interest in your feelings. What it offers is plainer and rarer: a system where the things you choose to learn actually stay, on rails you own, in files no company can lock or delete. It asks for 15 honest minutes a day and good cards in return. Give it those, and a year from now you won’t be the person who frantically re-reads before every high-stakes moment. You’ll be the one who simply knows — and who never has to rent their own memory back again.