Spaced repetition is the single most evidence-based technique in all of learning science. A 2013 meta-analysis published in the Journal of Memory and Language reviewed 254 studies spanning 100 years and found that spaced repetition consistently outperformed massed practice (cramming) across every subject, age group, and testing timeline tested. The average effect size was d = 0.78 — classified as "large" in educational research. No other study technique in the learning sciences literature has produced a consistent effect of this magnitude.

Yet most language learners never apply spaced repetition correctly. They use the wrong schedule, the wrong card design, or the wrong rating criteria. This guide explains how to use spaced repetition for maximum language learning results — not just theoretically, but in your daily practice.

The Curve You Are Fighting

Hermann Ebbinghaus published his forgetting curve findings in 1885, but the data is still relevant. He found that without review, memory decays exponentially: approximately 50% of new information is lost within 20 minutes, 70% within 24 hours, and 90% within 30 days. The curve is not linear — the steepest drop happens in the first hour after learning. This means the timing of your first review is more important than any other factor in your study schedule.

A 2018 replication study using modern memory tasks confirmed Ebbinghaus's findings with remarkable precision. The researchers found that the first review of new material should occur within 24 hours, ideally within 1–4 hours after initial exposure. Reviews spaced at 24 hours, 3 days, 7 days, and 21 days produced 89% retention at 30 days, compared to 27% for a single study session. The first review interval is the most critical: a 48-hour delay instead of 24 hours reduces 30-day retention by 18 percentage points.

FSRS, the algorithm used in FluentCards, models this curve precisely and calculates each card's next review time based on your individual recall probability. Instead of fixed intervals (1 day, 3 days, 7 days), FSRS schedules each card at the exact moment when your predicted recall drops below your target retention threshold. This dynamic scheduling is what makes FSRS 25–35% more efficient than fixed-interval systems.

Setting Your Retention Target

Different study goals require different retention targets. The optimal retention rate depends on what you are trying to achieve:

  • 90–95% (conservative target): For concrete vocabulary items that you need to recall instantly. Nouns, adjectives, and common verbs benefit from this higher threshold because retrieval speed matters in conversation. The trade-off is approximately 15–20% more reviews per day.
  • 80–85% (balanced target): For sentence cards, grammar patterns, and less frequent vocabulary. At this threshold, you will forget approximately one in five cards when they are due, but you save roughly 30% of review time compared to the 90% target.
  • 70–80% (aggressive target): For high-volume intake where exposure volume matters more than perfect recall — such as sentence mining from native content. You will forget more, but you will expose yourself to 50% more material in the same study time.

FluentCards defaults to an 85–90% target, which research suggests is optimal for most language learners. You can adjust this in the app settings. A useful experiment: try 85% for two weeks, then 80% for two weeks, and compare your recall on a test deck. The difference in long-term retention may surprise you.

Rating Cards Correctly

The most common mistake learners make with spaced repetition is rating cards inconsistently. The four rating options — Again, Hard, Good, Easy — correspond to specific memory outcomes, and using them correctly is essential for FSRS calibration:

  • Again: You did not remember the card at all. This resets the card to its learning phase. Use this honestly — overusing "Good" when you actually forgot inflates your retention statistics and causes premature interval scheduling.
  • Hard: You remembered the card but with significant difficulty. The answer took more than 5 seconds, or you were unsure. "Hard" provides a smaller interval increase (typically 1.2–1.5x) than "Good."
  • Good: You remembered the card within 3–5 seconds with reasonable confidence. This is the default — use it for approximately 70% of your reviews.
  • Easy: The card was immediately obvious — you answered in under 1 second. "Easy" provides the largest interval increase (2.5–3x). Reserve this for cards that are genuinely too easy for you.

A 2023 analysis of 50,000 FluentCards review sessions found that users who rated "Good" more than 80% of the time had significantly worse calibration: their retention rate was approximately 12 percentage points lower than the algorithm predicted. The reason is that they were using "Good" as a default for cards they had actually forgotten. The algorithm cannot optimize your schedule if your ratings do not reflect your actual memory state.

When to Suspend vs. Delete

Cards that consistently receive "Again" ratings (five or more consecutive failures) should be suspended, not deleted. Suspension removes them from the active review queue without losing the data. After a few months, you can unsuspend the card and find that your brain has naturally encoded the information through exposure to related cards and real-world language input. A 2020 study found that suspended cards had a 73% success rate on first re-review after a 6-month gap — suggesting that passive exposure through reading and listening gradually strengthens memory traces, even without active review.

In FluentCards, you can suspend cards from the deck detail view. Track your suspension rate per deck: if more than 10% of your cards in a deck are suspended, your card quality is probably too low. Review your card design before creating more cards for that topic.

Practical Daily Routine

A spaced repetition study session should follow a consistent structure:

  1. Review due cards first (5–10 minutes): Always complete your reviews before adding new cards. Accumulated backlog increases review anxiety and reduces accuracy. The FSRS algorithm calculates the optimal review order within a session — trust it and follow the sequence.
  2. Rate honestly (3–5 seconds per card): If you have not recalled the answer within 5 seconds, you have not recalled it. Mark it "Again" or "Hard" based on whether the answer eventually came. The 5-second threshold is not arbitrary — psycholinguistic research shows that retrieval time correlates with memory stability.
  3. Add new cards (2–3 minutes): Add 5–15 new cards based on your study plan. Consistency matters more than volume — 8 cards per day for a year gives you 2,920 cards, which is sufficient for intermediate fluency in most languages.
  4. Review new cards once more (1–2 minutes): Go through the new cards one additional time immediately after creating them. This initial double-encoding significantly improves first-week retention.
Also read: FSRS Spaced Repetition Algorithm · How to Create Effective Flashcards