How Often Should You Review Flashcards for Language Learning?

The honest answer is daily - but that's not the interesting part. The interesting part is what actually happens when you skip, how much new content you can safely add, and why session length matters less than most people think.

The short answer is every day. The long answer is more useful: there's a specific reason daily review matters, and understanding it changes how you handle the days when "every day" stops being realistic.

What follows is the research-informed view on review frequency - not the generic "be consistent" advice, but what actually happens to your deck at different frequencies, how much new material your system can absorb, and the failure modes that every serious SRS user eventually runs into.

Why daily, specifically

Spaced repetition scheduling algorithms - SM-2, FSRS, and their variants - calculate intervals assuming you'll actually see the card when it's due. When a card is scheduled to come back in 3 days and you see it on day 3, the algorithm gets the signal it expects and adjusts your next interval based on how well you remembered.

When you see that same card on day 6 instead, two things go wrong. First, the memory has decayed further than the algorithm expected, so you're more likely to fail it - which kicks the scheduling back, effectively losing the progress that card had built. Second, the algorithm's model of your retention gets noisier, because delayed reviews don't carry the same signal as on-time ones.

Multiply that across a 1,000-card deck where you're skipping every third day, and you get a system that's running at maybe 60% of its designed efficiency. It still works - but it's doing substantially more work for substantially less retention than if you'd just opened the app every day.

How long should each session actually be?

Session length matters less than frequency, but not in the direction most people expect. Longer sessions don't help as much as you'd think. Shorter sessions work better than intuition suggests - as long as they're daily.

The practical range for most learners: 10–20 minutes per day of focused review. Under 10 minutes and you're often not clearing a reasonable daily queue. Over 20 minutes and attention drops, grading starts getting sloppy, and you're mostly forcing through cards you should have dealt with in a different session.

If you have more time and want to study more, the research-informed move is to split it rather than extend it. Ten minutes in the morning plus ten minutes in the evening beats one twenty-minute block, because each session becomes its own retrieval event. Two separate retrievals produce stronger memory than one longer one, even controlling for total time.

There's also a diminishing-returns pattern within a single session. The first five minutes of review are high-quality. The last five minutes of a forty-minute session are noise. If you're studying for that long, the system is telling you to either reduce daily new cards or split the session.

What skipping a day actually costs

One skipped day is not a disaster. The scheduler absorbs it, you review the overdue cards the next day, and the noise washes out over time. This is the pattern your routine should be robust to, because life happens.

Two to three skipped days is where it starts costing real retention. Cards scheduled in that window degrade enough that more of them fail when you finally review them. Failures kick intervals back, which creates a pile of short-interval cards that compete with the natural daily flow - and the backlog starts eating into time you'd normally spend on new material.

A week of skipped days is effectively a partial reset. You'll come back to a review queue that looks impossible, fail more cards than usual, and spend the next two weeks digging out. This is the failure pattern that makes people quit SRS - not because the method stopped working, but because the recovery felt worse than starting over.

The protective habit is to have a fallback mode. Instead of skipping entirely on busy days, do a five-minute emergency session on just the overdue cards. You won't make progress, but you'll keep the backlog from growing, which is the only thing that matters for keeping the habit recoverable.

How many new cards per day?

New cards are the lever that decides everything else. Too many and the queue grows faster than you can maintain. Too few and progress stalls.

The practical range for language learners: 5–15 new cards per day for sustainable long-term study, scaling with your daily available time. At 10 minutes per day, 5–7 new cards is realistic. At 20 minutes, 10–15 works. Above 15 new cards daily is possible short term but very few learners sustain it beyond a few months without burning out or letting reviews slip.

The hard rule that matters more than the exact number: clear due reviews before adding new cards. Every day. If your review queue isn't at zero, you don't add new material. This single rule prevents the most common SRS failure mode, where the new-card count and the review-card count both grow, compound, and eventually overwhelm the schedule.

The math is brutal if you think it through. Ten new cards per day, at roughly 7–9 reviews per card before it matures, produces a steady-state review load of 70–90 cards per day once the deck is mature. That's manageable at 10–15 seconds per card (about 15–20 minutes). Twenty new cards per day doubles that review load, and most learners can't sustain 30–40 minutes of daily review indefinitely.

Two signals that your frequency is wrong

If your retention on mature cards (cards you've seen many times) is below 80%, your review frequency or grading honesty is off. The scheduler targets roughly 85–90% retention on mature cards; consistently lower means cards are decaying between reviews, which usually comes from skipped days or from marking cards easier than they actually were.

If your daily review count keeps growing despite stable new-card intake, your grading is too soft. Marking everything "good" when you barely remembered tells the scheduler those cards are stronger than they are, which schedules them for longer intervals, which makes them harder to remember next time - and cards that should have matured keep coming back. Honest grading, including failing cards when you actually failed, is what stabilizes the review load over time.

The practical answer for most people

If you want one routine that works for most language learners: 15 minutes per day, every day, in the morning. Add 10 new cards on days when the review queue feels light. Skip new cards entirely on days when the queue is heavy. Grade honestly. Don't chase a specific streak - chase the habit of opening the app, even when you only have five minutes.

This is the pattern that the research and practical experience converge on. More isn't better. Longer isn't better. Consistent is better. The spaced repetition best practices guide covers the habit side of this in more detail - the frequency question is solved, the consistency question is where most learners actually succeed or fail.

MindDory's defaults are tuned for this pattern: conservative new-card limits, reviews that fit into short mobile sessions, and a review-first flow that clears due cards before offering new ones. The goal is to make the sustainable rhythm the default, rather than something you have to discipline yourself into.