I watched a second-year med student open Anki on a Monday morning to find 847 reviews waiting. She'd added 30 new cards daily for six weeks, skipped two weekends, and now faced a backlog that would take four hours to clear. She closed the app and didn't open it again for three days. By Thursday, the count was 1,200.
This is the 30-day review debt phenomenon, and it's the most common reason people abandon spaced repetition systems entirely.
TL;DR
Daily review load compounds exponentially in the first 30 days, then stabilizes. Most learners can sustain 100-150 reviews/day long-term (15-25 minutes). Prevent burnout by: (1) calculating your steady-state load before adding cards, (2) pausing new cards when reviews spike, (3) using FSRS reschedule to redistribute load, and (4) accepting that some decks need slower intake rates.
Why Review Counts Explode
When you add 20 new cards today, you're not committing to 20 reviews. You're committing to roughly 80-120 reviews over the next 30 days as those cards cycle through their early intervals (1 day, 3 days, 7 days, 14 days).
Here's the math for SM-2 (Anki's default algorithm):
- Day 1: 20 new cards → 20 reviews
- Day 2: 20 new + 20 from yesterday → 40 reviews
- Day 3: 20 new + ~16 from Day 1 (80% retention) + 20 from Day 2 → 56 reviews
- Day 4: 20 new + ~13 from Day 2 + ~16 from Day 3 → 49 reviews
- Day 8: 20 new + cards from Days 1, 4, 5, 6, 7 → 70-90 reviews
By Day 30, if you've added 20 cards every day without missing a session, your daily review load stabilizes around 100-140 reviews. That's the steady-state load — the equilibrium between new cards entering the system and mature cards graduating to longer intervals.
The problem: most people don't reach Day 30 without interruption. A single skipped day adds 20-40 reviews to the next day. A skipped weekend adds 80-120. The backlog grows faster than you can clear it, and the psychological weight becomes unbearable.
The Sustainable Daily Load
From analyzing SmartRecall user data across 12,000+ active learners, here's what daily review loads look like in practice:
15-20 minutes/day (100-150 reviews):
Sustainable for 80% of users over 6+ months. This is the sweet spot for most med students, language learners, and working professionals.
25-35 minutes/day (200-250 reviews):
Sustainable for 40% of users. Common during exam prep sprints (USMLE Step 1, MCAT, JLPT N2). Burnout risk increases after 8-10 weeks without breaks.
40+ minutes/day (300+ reviews):
Sustainable for <15% of users. Typically requires dedicated study blocks and high intrinsic motivation. Most learners in this range are full-time students or in intensive language immersion.
5-10 minutes/day (50-80 reviews):
Sustainable indefinitely. Good for maintenance decks (vocabulary, formulas) or learners with unpredictable schedules.
Your sustainable load depends on three factors:
- Review speed: 5-10 seconds/card for recognition (vocab, anatomy), 15-25 seconds for recall (cloze deletions, problem-solving)
- Cognitive load: Low-stakes reviews (Duolingo-style) vs. high-stakes (USMLE-style clinical vignettes)
- Schedule consistency: Daily 20-minute blocks are easier than 2-hour weekend cram sessions
Calculating Your Steady-State Load
Before adding new cards, calculate where your daily review count will stabilize. Use this formula:
Steady-state reviews/day ≈ (New cards/day) × (Average reviews per card in first 30 days)
For SM-2 with default settings:
- Average reviews per card in first 30 days: 5-7 reviews
- 10 new cards/day → 50-70 reviews/day steady-state
- 20 new cards/day → 100-140 reviews/day
- 30 new cards/day → 150-210 reviews/day
For FSRS (Anki's newer algorithm):
- Average reviews per card in first 30 days: 4-6 reviews (slightly more efficient)
- 10 new cards/day → 40-60 reviews/day
- 20 new cards/day → 80-120 reviews/day
- 30 new cards/day → 120-180 reviews/day
SmartRecall's load predictor runs this calculation automatically and warns you when your new-card rate will push you above your target daily time. It's saved me from overcommitting to decks I couldn't sustain.
When to Pause New Cards
Pausing new cards is not failure. It's load management.
Pause immediately when:
- Your review backlog exceeds 1.5× your normal daily count
- You've skipped 2+ days and reviews are compounding
- You're in exam week or a high-stress period
- Your retention rate drops below 75% (sign of overload)
How to pause effectively:
- Set new cards/day to 0 in deck options
- Clear your backlog over 3-5 days
- Resume at 50-75% of your previous new-card rate
- Monitor for 7 days before increasing again
I pause new cards every 6-8 weeks for a "consolidation week" where I only do reviews. This prevents creeping overload and gives me time to identify leeches (cards I keep failing) before they multiply.
The FSRS Reschedule Strategy
FSRS (Free Spaced Repetition Scheduler) is Anki's machine-learning algorithm that optimizes intervals based on your actual review history. One of its most powerful features for load management is the Reschedule function.
When you enable FSRS and run Reschedule:
- Cards with artificially short intervals (from SM-2's conservative defaults) get pushed out
- Cards you've consistently aced get longer intervals immediately
- Your daily review load drops by 15-30% in the first week
Real example from a SmartRecall user (NCLEX prep):
- Before FSRS: 180 reviews/day, 35 minutes
- After FSRS + Reschedule: 125 reviews/day, 24 minutes
- Retention rate: 87% → 89% (slightly improved)
The catch: Reschedule is a one-time redistribution. If you're still adding 30 new cards/day, you'll hit the same steady-state load within 2-3 weeks. FSRS makes the system more efficient, but it doesn't change the fundamental math of new cards → future reviews.
To enable FSRS in Anki:
- Update to Anki 23.10 or later
- Go to Deck Options → FSRS
- Click "Optimize" to train on your review history
- Click "Reschedule all cards" (do this once, not repeatedly)
Load Management Workflows
Workflow 1: The Exam Sprint (8-12 weeks)
Goal: Maximize retention for a fixed deadline (USMLE Step 1, MCAT, JLPT N1)
- Weeks 1-4: 25-30 new cards/day, build the deck
- Weeks 5-8: 15-20 new cards/day, let reviews stabilize
- Weeks 9-10: 5-10 new cards/day, focus on weak areas
- Weeks 11-12: 0 new cards, pure review + practice exams
Expected peak load: 200-250 reviews/day in Week 6-7
Workflow 2: The Maintenance Deck (indefinite)
Goal: Retain vocabulary, formulas, or clinical knowledge long-term
- 5-10 new cards/day consistently
- Pause new cards during busy weeks
- Run FSRS Reschedule every 3 months
- Accept that some days you'll skip reviews (aim for 6/7 days/week)
Expected steady-state: 50-80 reviews/day
Workflow 3: The Language Immersion (6-12 months)
Goal: Build vocabulary rapidly while living abroad or in intensive study
- Weeks 1-8: 20-25 new cards/day (high motivation phase)
- Weeks 9-16: 15 new cards/day (consolidation)
- Weeks 17+: 10 new cards/day (sustainable long-term)
- Use filtered decks to prioritize high-frequency words
Expected peak load: 150-180 reviews/day in Week 10-12
Red Flags You're Overloaded
Watch for these signs:
- Retention drops below 80%: You're reviewing too fast or adding cards too quickly
- You dread opening the app: Psychological resistance is an early warning
- You're hitting "Again" reflexively: Sign of cognitive fatigue, not genuine forgetting
- Reviews take 2× longer than usual: Mental fatigue compounds review time
- You skip 2+ days in a row: The backlog is already unsustainable
When you see these, pause new cards immediately and reduce your daily load by 30-40% for one week.
Tools for Load Prediction
Anki's built-in stats:
Go to Stats → Forecast to see projected review counts for the next 30 days. This assumes you maintain your current new-card rate and don't skip days.
SmartRecall's load predictor:
Enter your target daily time (e.g., 20 minutes) and review speed (e.g., 8 seconds/card). It calculates your maximum sustainable new-card rate and warns you when you're approaching overload. It also tracks your actual review time vs. predicted time to catch drift early.
Spreadsheet method:
Track these daily for 2 weeks:
- New cards added
- Reviews completed
- Time spent
- Retention rate
Calculate your average reviews/new-card and time/review. Use those to project your steady-state load.
The 80% Rule
Here's the principle that's kept me consistent for four years:
Aim to complete 80% of your reviews 80% of the time.
Not 100%. Not every day. 80%.
This means:
- If your steady-state is 120 reviews/day, aim for 100
- If you miss a day, don't try to clear the full backlog the next day
- If you're at 90% retention, you have room to add more cards
- If you're at 75% retention, you're overloaded
Spaced repetition is a marathon, not a sprint. The learners who succeed long-term are the ones who build sustainable habits, not the ones who burn brightest for six weeks and quit.
What About Mature Cards?
Once cards reach intervals of 30+ days, they contribute minimally to daily load. A card at a 90-day interval will only appear every three months. This is why steady-state load stabilizes — new cards cycle through their early reviews and graduate to long intervals, making room for the next batch.
The danger zone is Days 1-30 for any given card. After that, the card is "paid off" in terms of daily time commitment.
Final Recommendations
- Start conservative: 10-15 new cards/day for your first month, even if it feels slow
- Track your actual time: Use a timer for one week to see if your perceived load matches reality
- Build in pause weeks: Every 6-8 weeks, set new cards to 0 and consolidate
- Use FSRS: The efficiency gains are real and measurable
- Accept imperfection: Missing a day doesn't ruin the system; trying to catch up in one session does
Your daily review load is the single biggest predictor of whether you'll still be using spaced repetition six months from now. Manage it actively, adjust it frequently, and remember that slower progress is infinitely better than burning out and quitting.
SmartRecall's load management tools are designed around these principles — we'd rather you add 10 cards/day consistently than 30 cards/day for three weeks before abandoning the deck. If you're struggling with review overload, the load predictor and auto-pause features can help you find your sustainable rhythm.
The goal isn't to maximize cards reviewed. It's to maximize knowledge retained over years, not weeks.
