Spaced Repetition vs Immersion: Which One Actually Works?

It's the argument that refuses to die in language learning communities. The short answer is that the question is slightly wrong: these two methods train different skills, and comparing them head-to-head misses the point.

Every few weeks somebody on a language learning forum posts the same question in a slightly different form. Flashcards or immersion? Anki or Netflix? SRS or comprehensible input? The replies are always the same mix of strong opinions, personal anecdotes, and citations that nobody actually reads.

The honest answer is less exciting than either camp wants. Spaced repetition and immersion aren't alternatives. They train fundamentally different skills, and treating them as competing methods is the reason so many learners end up with a large deck and weak listening comprehension, or hundreds of hours of shows watched and a vocabulary that plateaus.

What each method actually builds

The first confusion to clear up: these two methods are not doing the same job. Spaced repetition and immersion build different parts of your language ability, and the differences are bigger than most debates acknowledge.

Spaced repetition builds vocabulary that you can retrieve on demand. Its core mechanism is retrieval practice on a schedule that catches words just before you forget them. What you get out of it is a set of words you recognize in writing or speech and can, with effort, recall when you need them. That's valuable but narrow - it's memory storage, not language processing.

Immersion builds everything else. Processing speed. Pattern recognition for grammar you never studied explicitly. The ability to pick up meaning from context when your vocabulary is incomplete. Pragmatics - how native speakers actually use the language as opposed to how textbooks say they should. None of that lives in a flashcard deck, and none of it can be built through pure SRS.

This is why the comparison is slightly broken as a question. Asking whether SRS or immersion is better is like asking whether push-ups or running is better. It depends on what you're trying to achieve, and for most language learners the answer is both - but not in equal amounts, and not at the same stages.

Where spaced repetition clearly wins

Spaced repetition wins decisively in one area: building a working vocabulary from zero to intermediate. This is the research-backed claim that actually holds up.

At the beginning of a language, immersion barely works. If 90% of the words in a podcast are unknown to you, you're not learning from it - you're listening to noise with no way to extract meaning. Comprehensible input requires comprehension, and comprehension requires a baseline vocabulary that has to come from somewhere. Spaced repetition is how most successful learners build that baseline efficiently.

The numbers are meaningful here. A learner can move from zero to a 1,000–2,000 word working vocabulary in roughly 3–6 months of consistent SRS at 10–15 new words per day. That same vocabulary acquired purely through immersion takes substantially longer because most beginner-level content doesn't contain enough repetition of new words in comprehensible contexts to build memory efficiently.

Spaced repetition also wins for maintenance. Once you've learned a language and aren't using it daily, a 5–10 minute review session a few times a week prevents decay more efficiently than anything else. Immersion works for maintenance too, but requires more time and isn't always available.

Where immersion clearly wins

Immersion wins decisively on everything that happens in real time. Listening comprehension at native speed. Speaking fluency. The ability to follow conversations where you don't control the topic. The development of an ear for what sounds natural versus what's technically correct but wrong.

None of these skills can be built through flashcards, regardless of how good the cards are or how consistently you review them. They require actual language input being processed under real conditions - fast, noisy, ambiguous, full of cultural references that don't translate cleanly. A flashcard isolates a word from everything that makes real language hard. Immersion gives you the full mess.

This is why advanced learners who hit a plateau rarely break through it with more flashcards. At that level, the bottleneck is almost never vocabulary size - it's the gap between recognizing words on cards and processing them at real-time speed in native-speaker speech. That gap only closes with exposure to real language in real quantities.

The research: does one produce better outcomes?

Studies that compare SRS and immersion head-to-head are rarer than you'd think, because good comparisons are hard to design. The honest summary of what the research does show: spaced repetition produces larger measurable vocabulary gains in the short term, while immersion produces stronger real-world language use in the long term. Both results replicate reasonably well.

The catch is that short-term vocabulary gains measured on a test don't automatically translate into usable language. A learner who's memorized 2,000 words through SRS but never used the language for real communication will still struggle to order coffee. A learner with 1,000 hours of immersion and 500 words of solid flashcard study will often outperform the first learner in any practical test, despite having "less vocabulary" on paper.

This is part of why controlled studies can feel misleading. They measure what's measurable (word recognition on a test), which happens to be exactly what SRS is optimized to produce. They don't always measure fluency, speaking confidence, or the ability to process unfamiliar topics - which is where immersion quietly accumulates its advantages.

The failure modes of each method alone

Using only one of these methods reliably produces specific, predictable problems. If you're running into any of these, the diagnosis is usually that your method mix is unbalanced.

Pure SRS failure mode: a large deck, a high mature card count, and the humbling experience of not being able to follow a normal conversation. You recognize every word when it's on a card in front of you. You can't catch half of them when they come at you in fast native speech. This is the classic outcome of treating flashcards as a complete language learning method instead of a vocabulary tool.

Pure immersion failure mode: hundreds of hours of content consumed, a rough understanding of what's going on most of the time, but the same high-frequency words keep surprising you. Your vocabulary feels like it should be bigger than it is. Specific words that you've heard many times never quite crystallize because you've never retrieved them under any pressure to produce a definition. This is the outcome of assuming exposure alone builds usable vocabulary.

What the actual split should be

The practical question isn't which method is better. It's how to split your time between them. For intermediate learners, the research-informed answer is roughly 20–30% of study time on SRS and 70–80% on immersion - but this shifts at beginner and advanced levels.

At beginner level, SRS should go up (40–50% of time) because immersion doesn't work yet. At intermediate level, the 20–30% default applies cleanly. At advanced level, SRS drops to maintenance mode (5–10%) while almost all time goes to immersion and output practice.

The breakdown by level and the reasoning behind each range is covered in more detail in the optimal SRS-to-immersion ratio article - the short version is that no single ratio works for every stage, and locking yourself into one is a reliable way to plateau.

The one combination that multiplies both

There's a specific way to combine SRS and immersion that produces results neither method can match alone. It's not technically complicated, but most learners don't do it.

The trick is to feed your flashcards from your immersion. Instead of studying a generic top-1000 wordlist, you pull new cards from the content you're actually consuming - the show you watched last night, the podcast on your commute, the book you're reading. When a word surprises you in real content, it becomes a flashcard with the sentence you heard it in as the example.

What this does: every word in your deck already has context before it enters spaced repetition. You're not trying to memorize an abstract translation - you're consolidating a memory that already has a handle in real experience. Retention rises sharply because the card isn't the only encounter with the word. It's one of three or four encounters, all linked to the same scene in your head.

This is also what closes the feedback loop in the other direction. Words you're reviewing in the deck start appearing in your next immersion session, and recognizing them in new contexts strengthens memory further. SRS without immersion is isolated retrieval. SRS fed by immersion is retrieval on top of real exposure - and the difference shows up in how quickly vocabulary becomes usable.

Which method should you actually pick?

If you've read this far hoping for a verdict, here it is, bluntly: if you're forced to choose only one, the answer depends entirely on your level.

Complete beginner: spaced repetition. Immersion doesn't work yet. Your first 500–1,000 words have to come from somewhere structured, and flashcards are the most time-efficient way to build that base.

Lower intermediate onward: immersion, with minimal SRS. The bottleneck stops being vocabulary and starts being processing speed. Flashcards can help at the margin, but the bulk of the work is done by real language exposure.

Advanced: almost entirely immersion. SRS at this level is maintenance. The new vocabulary you need is rare enough that you'll encounter it in content rather than in top-frequency lists, and the skills that matter most - fluency, register, idiom - aren't built by flashcards at all.

But the "pick one" framing is artificial. In any realistic scenario you have access to both, and the learners who make the fastest progress use both in combination, with the mix shifting as their level changes. MindDory is built around exactly this - SRS fast enough to leave most of your study time for real input, with cards that come with context by default so the two sides of your study routine aren't disconnected.