Anki Settings Guide 2026: Optimize for FSRS and Avoid Ease Hell

The Optimization Problem: Why Default Anki Settings Burn You Out

Anki is a remarkably powerful tool, but it ships with a dirty secret: the default settings are designed for a generic user who doesn't exist. Most students — especially medical students and language learners who rack up thousands of cards — never touch the settings panel. They install Anki, download a shared deck, and start reviewing. A few months later, they're drowning in 300+ daily reviews, watching their ease percentages plummet, and wondering why a tool that promised efficient spaced repetition feels like a second job.

This phenomenon has a name in the Anki community: ease hell. It happens when the default SM-2 algorithm's ease factor drops below 130%, making every review cycle punishingly short. The result is review burnout, and it's the primary reason many students abandon Anki despite the tool's proven effectiveness — a 2022 peer-reviewed study of first-year medical students at the University of Central Florida found that roughly 30% of students do not use Anki at all, and that proportion remained steady throughout the course.

The good news is that the fix doesn't require a complete overhaul. Switching to Anki's newer FSRS algorithm and adjusting just four or five key settings — learning steps, graduating interval, new interval percentage, desired retention, and max interval — can cut your daily review load by an estimated 20–30% while maintaining or even improving your retention rate. This isn't a theoretical claim; it's the community-reported benchmark that has driven thousands of users to make the switch since FSRS became a built-in option in Anki 23.10.

Understanding Anki's Two Algorithms: SM-2 vs. FSRS

Before you change anything, it's worth understanding what's running under the hood. Since Anki 23.10, the application ships with two available scheduling algorithms: the legacy SM-2 (based on SuperMemo 2) and the newer FSRS (Free Spaced Repetition Scheduler). Most users who installed Anki before late 2023 are still on SM-2 by default, and that's where the trouble starts.

How SM-2 Works (and Where It Breaks)

SM-2 uses a single ease factor per card, starting at 250%. Each time you rate a card, the algorithm adjusts that ease factor by a fixed amount: Again drops it by 20 percentage points, Hard drops it by 15, and Easy increases it by 15. The ease factor can never fall below 130%. The problem is that a single difficult card — one where you hit Again a few times in a row — can tank its ease factor, and that card will then show up far more frequently than it should, forever. This is the core mechanism of ease hell.

SM-2 also treats all cards as roughly the same difficulty. It doesn't learn from your specific memory patterns. A card you've seen ten times and a card you've seen twice get scheduled using the same rigid interval multiplier, which is wildly inefficient.

How FSRS Changes the Game

FSRS replaces the single ease factor with a three-component memory model: Retrievability (R), Stability (S), and Difficulty (D). Instead of applying a fixed penalty for a wrong answer, FSRS uses machine learning to analyze your review history and build a personalized model of how your memory works. It then schedules each card based on that model.

The practical result is that FSRS can schedule cards much more efficiently than SM-2. According to the official Anki FAQ, "with FSRS, users have to do fewer reviews than with Anki's default algorithm to achieve the same retention level." Preliminary tests indicate FSRS is roughly on par with SuperMemo's SM-17, which is widely considered the most advanced spaced repetition algorithm available. FSRS also handles delayed reviews — cards you missed for a day or two — much better than SM-2, which tends to over-schedule them.

For a deeper technical comparison of the two algorithms, including benchmark data and community migration experiences, read The Algorithm Divide: Why FSRS Is Making SM-2 Obsolete.

The 4 Settings That Matter Most (Before You Touch FSRS)

Whether you switch to FSRS or stay on SM-2, four settings have an outsized impact on your daily review load and your risk of falling into ease hell. These settings are often left at their defaults, which is a mistake.

1. Learning Steps

Learning steps control how many times you see a new card before it graduates to the review queue. The default is often just one step (1 minute), which is far too short. Both the LeanAnki and Zach Highley sources recommend a three-step sequence: 15 minutes, 1 day, and 3 days. This gives you three chances to see a new card at increasing intervals before it becomes a permanent review card.

Why this matters: With a single short learning step, cards that you barely know can graduate into the review queue, where they'll be scheduled at much longer intervals. You'll forget them, hit Again, and start the ease-hell cycle. Three learning steps ensure that only cards you've successfully recalled at least three times — including after a 3-day gap — make it into long-term memory.

2. Graduating Interval

The graduating interval is the first interval a card receives after it passes all its learning steps. The default is often 1 day, which is too short. Both LeanAnki and Zach Highley recommend 6–7 days. With a 7-day graduating interval and 250% starting ease, your card's intervals will grow like this: 7 days → 17.5 days → 43.75 days → ~109 days. After just five successful recalls, a card can be scheduled at roughly a 3-month interval.

A shorter graduating interval means cards return sooner, which increases your daily review count without a proportional retention benefit. The extra reviews are wasted effort.

3. New Interval %

This is the single most overlooked setting for preventing ease hell. When you lapse on a card (hit Again on a review card), the New Interval % determines what happens to its previous interval. The default is 0%, which resets the card's interval to zero — as if you'd never seen it before. That means a card you've successfully recalled ten times over six months is treated like a brand-new card after one lapse.

LeanAnki recommends setting this to 50–60%. At 60%, a card with a 100-day interval that you lapse on will be rescheduled to a 60-day interval — still a meaningful review, but not a complete reset. This prevents well-learned cards from wasting your time by climbing back up from zero.

4. Maximum Interval

The max interval caps how far into the future a card can be scheduled. The default is often 100 days, which is far too short for long-term retention. Zach Highley recommends at least 180 days for general use and 240 days for the first two years of medical school. LeanAnki suggests 180 or 365 days depending on your exam timeline.

A max interval that's too short forces cards to return more frequently than necessary, inflating your daily review count. If you're studying for a multi-year exam like the MCAT or USMLE, set it to at least 365 days. Cards that have been stable for a year don't need to be seen every three months.

FSRS Deep Dive: Desired Retention and One-Click Optimization

If you're ready to switch to FSRS — and you should be — the process is straightforward. Go to Anki's Deck Settings, find the Scheduler section, and change the algorithm from SM-2 to FSRS. That's the first step. The second step is setting your desired retention.

Setting Desired Retention

Desired retention is the single most important FSRS setting. It's a slider that lets you tell the algorithm what percentage of cards you want to recall correctly on the first try. The default is 0.90 (90%), which is a good starting point for most users.

Here's the trade-off: a higher desired retention (say, 0.95) means you'll see cards more frequently, which increases your daily review count. A lower desired retention (0.80) means fewer reviews but more forgotten cards. The 0.90 sweet spot balances review volume with recall accuracy for most high-stakes exam contexts.

One-Click Optimization with FSRS Helper

The FSRS Helper add-on (code 759844606) is the most important add-on for anyone using FSRS. It provides three critical features:

• One-click parameter optimization: The add-on analyzes your review history and finds the optimal FSRS parameters for your specific memory patterns. You don't need to understand the math — just click "Optimize" and let it run.
• Postpone and advance: These features let you smooth out daily review spikes. If you have 500 reviews due on Monday and 100 on Tuesday, you can postpone some Monday cards to Tuesday to balance the load.
• Load balancing: This automates the postpone/advance logic, distributing your reviews evenly across the week.

To install it, open Anki, go to Tools → Add-ons → Get Add-ons, and enter the code 759844606. After installation, you'll find the FSRS Helper options in the Tools menu.

Add-on code: 759844606
Install path: Tools → Add-ons → Get Add-ons → paste code → OK

Recommended Settings Table (With Rationale)

The table below consolidates the recommended values for both SM-2 holdouts and FSRS users. Each row includes a brief rationale so you understand why the value is recommended, not just what to set.

How to Interpret Answer Buttons: Avoiding Ease Hell

The way you use the Again, Hard, Good, and Easy buttons is the single biggest behavioral factor in whether you end up in ease hell. Most users develop bad habits — hitting Hard when they should hit Good, or hitting Again when they almost knew the answer — that slowly degrade the algorithm's effectiveness.

SM-2 Button Mechanics

In SM-2, each button has a specific effect on the card's ease factor:

The critical insight is that Hard is almost never the right choice. It decreases your ease factor by almost as much as Again, but it only gives you a slightly longer interval. If you're hesitating on a card, you should either hit Again (if you genuinely didn't know it) or Good (if you knew it but took a moment). Hard is the fastest path to ease hell because it punishes your ease factor without giving you the learning benefit of a full repeat.

FSRS Button Mechanics

FSRS handles buttons differently. Instead of modifying a single ease factor, each button affects the card's memory state — Retrievability, Stability, and Difficulty — in a more nuanced way. The practical advice is similar, though: use Again only when you truly forgot, use Good as your default for correct recalls, and avoid Hard unless you have a specific reason. Easy is rarely needed because FSRS already schedules cards efficiently.

The biggest difference is that FSRS recalculates difficulty dynamically. A card that you struggle with today won't be permanently penalized the way it would be in SM-2. This is one of the primary reasons FSRS users report much lower rates of ease hell.

Interval Modifier Tuning (SM-2 Users Only)

If you're not ready to switch to FSRS — or if you're using an older version of Anki that doesn't support it — the interval modifier is your primary tool for tuning retention. This setting acts as a global multiplier on all review intervals. A value of 100% means intervals are used as calculated. Increasing it to 110% makes all intervals 10% longer (fewer reviews, lower retention). Decreasing it to 90% makes intervals shorter (more reviews, higher retention).

Zach Highley recommends starting at 100% and adjusting by ±5% based on your True Retention data. The goal is to hit 88–90% retention. If your True Retention is above 92%, you're reviewing too frequently — increase the interval modifier by 5%. If it's below 85%, you're not reviewing enough — decrease it by 5%.

For a complete step-by-step walkthrough of interval modifier tuning and other SM-2 settings, see the How to Configure Anki Spaced Repetition Settings guide.

Helpful Add-Ons for Optimization

Beyond the FSRS Helper, several add-ons can help you monitor and optimize your review process. The table below lists the most useful ones for the optimization workflow described in this guide.

To install any of these, open Anki, go to Tools → Add-ons → Get Add-ons, and enter the code. Install add-ons one at a time and restart Anki after each installation to avoid conflicts.

Troubleshooting Common Problems

Even with optimal settings, you may encounter issues during the transition. Here are the most common problems and their solutions.

"My retention dropped after switching to FSRS"

This is normal for the first 2–3 weeks. FSRS needs enough review data to build an accurate model of your memory patterns. During this calibration period, retention may fluctuate. Keep reviewing consistently and run the FSRS Helper optimization again after you've accumulated 1,000+ reviews. If retention is still low after three weeks, try increasing your desired retention from 0.90 to 0.92.

"I'm still seeing too many reviews"

Check two things. First, verify that your learning steps are set to 15m 1d 3d — if they're shorter, cards are graduating too quickly and returning as reviews sooner than necessary. Second, check your desired retention. If it's set above 0.92, you're asking the algorithm to prioritize recall frequency over efficiency. Dropping it to 0.88 will reduce your daily load noticeably.

"My ease is stuck at 130%" (SM-2 users)

This is classic ease hell. The card's ease factor has hit the floor and can't go lower, so every review cycle is punishingly short. Recovery strategies include: (1) use the "Reset ease" option in the card browser to reset the card's ease to 250%, (2) change the New Interval % to 50–60% to prevent future ease drops from lapsed cards, and (3) consider switching to FSRS, which doesn't use the same ease mechanism and will recalculate the card's difficulty dynamically.

Frequently Asked Questions

Should I switch to FSRS if my current SM-2 setup is working fine?

Yes, for long-term efficiency. Even if your current retention is acceptable, FSRS will likely maintain that retention with fewer reviews. The official Anki FAQ confirms that FSRS reduces reviews for the same retention level. The only reason to delay switching is if you're in the middle of an intensive exam period and don't want to risk the 2–3 week calibration dip.

What desired retention should I use?

Start at 0.90 (90%). This is the default and works well for most users. If you're studying for a high-stakes exam where forgetting is costly (like the MCAT or USMLE), you may prefer 0.92–0.95. If you're studying for a less critical subject and want to minimize review time, try 0.85–0.88. Adjust by 0.02 increments and give each setting 2–3 weeks to stabilize.

Will FSRS work with my existing decks?

Yes. When you switch to FSRS, it recalculates intervals for all your existing cards based on your review history. You don't need to recreate decks or re-enter cards. The transition is seamless — just change the algorithm in deck settings and run the FSRS Helper optimization.

How long does it take to see results after switching to FSRS?

Most users report noticeable improvements within 2–3 weeks of consistent reviews. The algorithm needs enough data to calibrate — roughly 1,000 reviews for a reliable optimization. After that, you should see a reduction in daily review count and more stable retention.

Can I go back to SM-2 after switching to FSRS?

Yes, but you'll lose the FSRS-optimized intervals. When you switch back to SM-2, Anki will recalculate intervals using the SM-2 algorithm, which will likely result in different — and potentially longer or shorter — intervals than what FSRS had scheduled. If you decide to switch back, expect a period of adjustment similar to the initial transition.

Do I need to change settings for each deck individually?

Settings are deck-specific in Anki. If you have multiple decks, you'll need to apply the recommended settings to each one. However, you can set default settings for new decks in Anki's preferences, which saves time if you frequently create new decks. For existing decks, you'll need to update each deck's settings manually.