Why You Shouldn't Put New Words Straight Into Spaced Repetition

The short answer

You should not put brand-new words straight into spaced repetition. SRS algorithms like FSRS are designed to maintain an existing memory trace by scheduling reviews right before you forget. They are not built to create that trace from zero. When a completely unseen word enters your review queue, you either pass by re-reading a sensory echo still in working memory, or you fail repeatedly in one session and permanently skew the card's difficulty rating. Acquisition and maintenance are different jobs. Treat them that way.

SRS is a maintenance tool, not an onboarding tool

Spaced repetition works brilliantly once you have already learned something. The spacing effect and modern schedulers like FSRS exploit a simple fact: memory decays on a predictable curve, and a well-timed review can reset that curve with minimal effort. That is why SRS feels like magic for vocabulary you already know.

But every SRS card carries an implicit assumption: a baseline memory trace already exists. The algorithm needs at least one honest retrieval attempt to estimate difficulty, stability, and when you will forget. Drop in a word you saw for the first time five seconds ago, and the scheduler has no real data to work with. It is trying to predict the forgetting curve for a memory that was never properly encoded in the first place.

This gap has a name in learning-science circles: the First-Turn Bottleneck. For a deeper look at the cognitive mechanisms and one structured fix, see Fixing the First-Turn Bottleneck.

Problem 1: You "pass" using working memory, not long-term recall

Show someone a new word, hide it, and test them one or two seconds later. Many will get it right every time. That looks like learning. It is not.

The brain is not retrieving from long-term storage. It is re-reading the sensory echo still floating in phonological or visual working memory. Classic short-term memory research shows unattended verbal items decay within roughly 10 to 18 seconds when mental rehearsal is blocked.¹ Flashcard studies find the same pattern: essentially immediate ("0-second") tests contribute little to long-term retention compared with tests after a short delay.²

Press "Good" on that first review and the SRS scheduler thinks you know the word. In reality, you only knew the afterimage. The first interval may be days or weeks. By then, the echo is long gone and you fail hard, often repeatedly.

Problem 2: You corrupt the card's difficulty rating

Modern SRS algorithms personalize scheduling from your review history. FSRS tracks Difficulty, Stability, and Retrievability for every card (see our guide to how spaced repetition works). Those values are only as good as the input you give them.

Feed the algorithm misleading first-turn data and you get misleading intervals:

• False "Good" on turn one: The card gets a long first interval even though you never encoded it. You fail later, the algorithm overcorrects, and you enter a cycle of easy-then-hard scheduling.
• Multiple "Again" presses in one session: You fail the same new word three times in ten minutes. The scheduler registers extreme difficulty from repeated misses that reflect missing encoding, not true item hardness. The card can get trapped in short intervals ("ease hell" in older systems).

Either way, day-one telemetry poisons the card. You are not giving FSRS a forgetting curve. You are giving it noise from a word that was never ready for spaced repetition.

Problem 3: Frustration loops waste your study session

Beyond bad data, immediate SRS onboarding feels terrible. You add twenty new words, open your deck, and the same unfamiliar items keep reappearing because you keep failing them. Twenty minutes later you have reviewed five cards and feel like you are getting nowhere.

That is not a discipline problem. It is a sequencing problem. You asked a maintenance system to do acquisition work. The session becomes a grind of re-exposure without the structured encoding pass that makes first contact stick.

What to do instead: acquisition before scheduling

The fix is not to abandon spaced repetition. It is to separate acquisition (first contact and encoding) from maintenance (timed review).

A practical acquisition pass for new vocabulary should include:

1. Meaningful first exposure. See the word in context, hear the pronunciation, and connect it to a translation or image. Passive exposure alone is weak; production strengthens encoding.³
2. Active recall, not recognition. Type or say the word from a cue. Avoid multiple-choice first tests: wrong options can encode false associations that hurt later recall.⁴
3. A short delay before the first blind test. Complete one unrelated task (5 to 10 seconds) between exposure and recall so you are not testing the sensory echo. Robert Bjork's desirable difficulty principle says retrieval should require genuine effort.²
4. Graduation into SRS. Only after a successful blind recall should the word enter your spaced repetition queue, with its attempt history passed to the scheduler.

If you use Anki or similar tools manually, you can approximate this with a "learning" or "intraday" step before cards graduate to the main interval queue. The key is: do not treat the first seconds after seeing a word the same as a day-3 or day-30 review. For a full walkthrough of priming, a short interleaved distractor, and blind recall before FSRS takes over, see Fixing the First-Turn Bottleneck.

LinGoat fixes this sequencing problem with a dedicated pre-SRS laddering system. New words go through priming (typed exposure while the word is visible), a 5 to 10 second interleaved distractor that clears working memory, and a blind active recall test before they ever enter the main FSRS schedule. Every attempt in that buffer is logged and passed to FSRS on graduation, so day-one intervals reflect real difficulty instead of a lucky first guess. Read the full design in Fixing the First-Turn Bottleneck, or see how LinGoat works and open the app to try it.

References

1. Peterson, L. R., & Peterson, M. J. (1959). Short-term retention of individual verbal items. Journal of Experimental Psychology, 58(3), 193-198. https://doi.org/10.1037/h0049234
2. Karpicke, J. D., & Roediger, H. L. (2007). Expanding retrieval practice promotes short-term retention, but equally spaced retrieval enhances long-term retention. Journal of Experimental Psychology: Learning, Memory, and Cognition, 33(4), 704-719. https://doi.org/10.1037/0278-7393.33.4.704
3. McDaniel, M. A., Howard, D. C., & Einstein, G. O. (2009). The read-recite-review study strategy: Effective and portable. Psychological Science, 20(4), 516-522. https://doi.org/10.1111/j.1467-9280.2009.02325.x
4. Roediger, H. L., & Marsh, E. J. (2005). The positive and negative consequences of multiple-choice testing. Journal of Experimental Psychology: Learning, Memory, and Cognition, 31(5), 1155-1159. https://doi.org/10.1037/0278-7393.31.5.1155