How Many Flashcards a Day Can You Actually Do?

Twenty new cards a day isn't twenty cards a day

This is the arithmetic almost everyone gets wrong, and it's why decks collapse.

A card you learn today doesn't go away. It comes back — repeatedly — and every new card you add raises the review load for weeks afterwards. So the daily cost of new cards is the new cards plus everything the previous days send back.

600 cards over 60 days:

Ten new cards a day sounds trivial. Eighty cards a day is a real commitment — about eleven minutes at eight seconds a card, and considerably more if your material is slower than that.

Why week one lies to you

The review load builds over weeks, then plateaus.

On day two there's almost nothing to review, so the daily total is close to the new-card count and the whole thing feels comfortable. By week three the reviews have stacked up and the same schedule is several times heavier — at exactly the point where quitting feels like a reasonable decision rather than a scheduling failure.

Check the steady-state figure, not the first-week one. That's the plan you're actually signing up for, and knowing it on day two costs nothing.

The review multiple is an input on purpose

How many reviews a new card generates per day depends on the scheduling algorithm, your accuracy, and how the deck is tuned. Published figures vary widely enough that asserting a single one would be misleading.

A commonly cited working range for a well-behaved deck is roughly 5 to 10 reviews per new card per day at steady state. But if your app reports your own ratio, use that — it's a measurement of your actual deck rather than a general figure, and it will be considerably more accurate.

This is the same reason the spacing planner won't tell you an optimal interval: better to compute honestly from a number you supply than to present a contested figure as settled.

If the load is unmanageable, look at the cards first

The instinct is to cut the deck. Often the real problem is accuracy.

Cards you get wrong come back sooner and more often, so a deck you're failing generates far more reviews than one you're getting right. Which means the load is frequently a quality problem wearing a quantity costume.

The usual culprits:

• Cards that are too long. If the answer is a paragraph, it isn't one card.
• Cards testing two things at once. These fail half the time by construction, and each failure schedules another review.
• Material you never actually understood. Cards are for retaining things you've learned. They're very poor at teaching you something in the first place, and a deck built on top of a gap generates failures indefinitely.

Fixing those reduces the load more than cutting the deck does — and it's the version that leaves you knowing the material.

Time yourself, then check it fits

Seconds per card varies enormously. Simple vocabulary recognition runs a few seconds; a card needing a derivation or a multi-part answer can take thirty seconds or more.

Since that figure is what converts a card count into a time commitment — and time is the constraint that actually decides whether this happens — it's worth one real session to measure rather than guess.

Then run it against the minutes you genuinely have. If the answer is over 100% of your available time, the plan doesn't fit, and the honest responses are a longer runway, a smaller deck, or better cards. The one that doesn't work is starting anyway and hoping.