How to Track Study Progress Across 800 Flashcards Without Losing Your Mind
If you've ever generated 800 flashcards from a textbook PDF and then realized you have no idea which ones you've actually learned, this article is for you. I've been on both sides of that mess — once with a hand-built Anki deck for USMLE Step 1, and again recently when I turned a 52-page chapter into 812 cards in 12 minutes. Generating cards is the easy part now. Tracking what's happening across hundreds of them is where most people quietly give up.
This is the system I use, the three numbers I actually look at every morning, and the failure modes I've learned to spot before they eat a week of study time.
Why 800 cards is the breaking point
Under 200 cards, you can run an entire deck on vibes. You roughly remember which ones you bombed yesterday, your daily review queue is 30–40 cards, and a single 20-minute session clears it. No spreadsheet, no dashboard, no problem.
Between 200 and 500 cards, the wheels start wobbling. The daily queue jumps to 80–120, you start skipping cards "just for today," and a small backlog forms. You can usually recover with a heroic Sunday session.
Past 500 cards, and definitely by 800, the math turns on you. With healthy SM-2 intervals and a realistic 85% recall rate, an 800-card mature deck will surface roughly 60–110 reviews per day indefinitely. Miss two days and you're staring at 200+ cards in the queue. Miss a week and the queue passes 500 and you mentally write off the deck. This is the silent killer of self-study.
The fix is not "more discipline." The fix is structured progress tracking so that you can see the deck breathing in real time and intervene early.
The three numbers that actually matter
Forget total cards. Forget cards-studied-today. Those are vanity metrics. Here are the three that predict whether you'll actually master the deck.
1. Mature card percentage
Definition: % of cards in the deck whose next-review interval is ≥ 21 days.
This is the closest thing to a "what % of this deck is in long-term memory" number. A card with a 21-day interval has been recalled correctly at least 4–5 times across spaced sessions. That's a real memory trace, not a 24-hour cram echo.
| Mature % | What it means |
|---|---|
| < 20% | You're still in acquisition phase. Normal for week 1–2. |
| 20–50% | Healthy mid-stage. Most of your time is on new + young cards. |
| 50–80% | You're consolidating. Daily load is dropping. |
| > 80% | The deck is mostly in maintenance mode. |
For an 800-card deck, expect to hit 50% mature around day 35–45 of consistent study, and 80% somewhere between day 70 and day 100, depending on subject difficulty.
2. Daily review backlog
Definition: number of cards whose due_date is before today but haven't been reviewed yet.
Today's due cards are not a backlog — they're just today's work. Backlog is yesterday's, last weekend's, last week's. This number should hover at zero. The moment it crosses 50 for two days in a row, you have a decision to make: catch up aggressively, or cut scope (split the deck, suspend a sub-topic, lower the daily new-card limit).
Healthy range: 0–30. Warning zone: 30–100. Triage zone: 100+.
3. Lapse rate
Definition: % of mature card reviews where you pressed "Again" (failed recall) over the last 7 days.
This is the truth-teller. A high lapse rate on mature cards means cards graduated to long intervals before you actually knew them — usually because you were guessing "Good" too aggressively in the first week.
Healthy: 5–12%. Warning: 12–20%. Something's broken: > 20%.
If lapse rate is > 20%, the answer is almost never "study harder." It's: rewrite the failing cards. They're either ambiguous, multi-answer, or testing the wrong atom.
How SmartRecall's dashboard surfaces these
When you open a deck of 800+ cards in SmartRecall, the deck overview puts these three numbers directly under the deck title — not buried six clicks deep. The mature % shows as a horizontal stacked bar (new / learning / young / mature), the backlog shows as a red badge if non-zero, and the 7-day lapse rate sits next to a sparkline so you can see whether it's trending up or down.
Under the hood this is the same SM-2 scheduler described here in detail, with one tweak: cards that lapse twice in a row get auto-flagged for review (not deleted, just surfaced) so you can rewrite them before they keep eating your morning queue.
The point isn't that the dashboard is fancy. The point is that the three numbers that decide whether you'll finish the deck are the three numbers you see first. If you have to dig for them, you won't look at them, and the deck will quietly rot.
Building a 21-day review rhythm for 800 cards
Here's the schedule I run for a fresh 800-card deck. The numbers assume SM-2-style scheduling and a daily new-card limit of 40.
| Day | New cards | Reviews due | Total session | Cumulative mature % |
|---|---|---|---|---|
| 1 | 40 | 0 | ~25 min | 0% |
| 2 | 40 | 38 | ~40 min | 0% |
| 3 | 40 | 70 | ~50 min | 1% |
| 5 | 40 | 95 | ~55 min | 4% |
| 7 | 40 | 110 | ~60 min | 8% |
| 10 | 40 | 120 | ~65 min | 14% |
| 14 | 40 | 130 | ~65 min | 22% |
| 17 | 20 | 135 | ~60 min | 28% |
| 21 | 20 | 130 | ~55 min | 35% |
Two things to notice. First, the daily session length plateaus around 60 minutes — it does not keep growing forever, which is the fear that scares people off big decks. Second, around day 17 I drop the new-card rate from 40 to 20. That's deliberate: by then your mature pool is large enough that adding 40 new cards a day pushes the daily queue past 60 minutes and the backlog starts to grow.
If you're using this for a real exam, mark the deck completion date (day 21–25 for new card intake) at least 30 days before exam day so the long-interval reviews have time to actually mature.
When to split a 800-card deck
An 800-card deck of one coherent subject (one chapter, one course unit) is fine. An 800-card deck spanning four unrelated chapters is a trap, because:
- You can't suspend a topic without manual tagging
- A bad week on chapter 3 contaminates your sense of how chapter 1 is doing
- Mature % becomes meaningless because it averages across difficulties
Split when any of these is true: deck spans more than ~3 distinct topics, you want to delay one topic by weeks, lapse rate on one tag is >2× the deck average, or total session time exceeds 75 minutes for 5+ days in a row.
The split itself is mechanical — in SmartRecall, multi-select by tag and "move to new deck" preserves card history and intervals, so you don't reset progress.
Common failure modes
Backlog denial. You let the backlog hit 200 and tell yourself you'll catch up "this weekend." You won't. The fix is a hard rule: backlog > 100 for two days = no new cards until backlog is back under 30.
Mature self-deception. You press "Good" on cards you half-knew because the answer was obvious in retrospect. Six weeks later those cards are at 60-day intervals and you fail every one. The fix is to be brutally honest in the first 14 days; "Good" should mean "I produced the answer before flipping," nothing softer.
No tagging on import. You generate 800 cards from a PDF and never tag them by subsection. Three weeks later you can't tell whether the 22% lapse rate is from one bad subsection or evenly spread. Always carry the section structure into tags at import time. SmartRecall does this automatically when generating from PDFs with detected H2/H3 headings.
Comparing rituals instead of outcomes. "I studied 90 minutes today" is not a metric. Mature % delta over 7 days is. If you've put in 7 hours this week and mature % moved 1 percentage point, the deck or your card writing is the problem, not your effort.
If you're comparing the underlying scheduler choices that drive these numbers, this breakdown of SM-2 vs FSRS vs Leitner vs Anki goes into why SmartRecall picked SM-2 for the 800-card use case specifically.
FAQ
How long does it take to truly master 800 flashcards?
For dense factual content (medical, legal, language vocab), expect 8–12 weeks from first import to >80% mature with a 7-day lapse rate under 10%. For lighter conceptual content, 5–7 weeks. Anyone promising 2 weeks is selling cram, not memory.
Should I review every card every day?
No, and trying to is the fastest way to kill an 800-card deck. The whole point of spaced repetition is that the algorithm decides which cards are due today. On a healthy day 30 of an 800-card deck, you'll see roughly 80–110 of them, not 800.
What if I miss several days of review?
Miss 1–2 days: just clear the backlog over the next two sessions, do not double the new-card rate to "catch up." Miss 3–7 days: clear backlog over a week with zero new cards added. Miss > 7 days: it is almost always better to suspend the deck for a deliberate restart than to grind through a 600-card backlog while resenting it.
Can I import 800 cards from Anki to track them in SmartRecall?
Yes. SmartRecall imports .apkg files and preserves card content, tags, and the SM-2 scheduling state (interval, ease factor, repetition count). Your mature cards stay mature. The dashboard numbers above start working from the first day after import.
Is 800 cards too many for one deck?
Depends on coherence, not count. 800 cards on one well-scoped topic is fine and arguably ideal. 800 cards spanning a whole textbook is a project — split it into 3–5 sub-decks aligned with how you'll actually use the knowledge.
If you're sitting on a 800-card deck right now and dreading opening it, the move isn't more willpower. It's looking at the three numbers above and intervening on whichever one is broken. SmartRecall is built to surface them by default — start a free account and your existing decks (or fresh PDF imports) will show mature %, backlog, and lapse rate from day one. If you're deciding whether to upgrade for higher daily generation limits, the pricing page lays out the per-tier card and PDF caps.
