Reviews & Spaced Repetition
Understand how FSRS scheduling works, review your questions, and let AI evaluate your answers.
What is spaced repetition?
Spaced repetition is a study technique based on a simple insight: you remember things better when you review them at increasing intervals over time. Instead of cramming everything at once, you revisit material right before you would naturally forget it — strengthening the memory each time and pushing the next review further into the future.
Research consistently shows that spaced repetition is one of the most effective ways to move knowledge into long-term memory. Temar automates the scheduling so you never have to guess when to study — just show up, review what is due, and the algorithm handles the rest.
The FSRS algorithm
Temar uses FSRS (Free Spaced Repetition Scheduler), a modern algorithm that models how your memory works. For each item you review, FSRS tracks two key properties:
- Stability — how long the memory will last before you are likely to forget it. Higher stability means longer intervals between reviews.
- Difficulty — how inherently challenging the item is for you. Harder items get shorter intervals and more frequent reviews.
After each review, FSRS updates these values based on your rating and calculates the optimal date for your next review. The algorithm adapts to your personal learning pace — items you find easy get spaced out quickly, while items you struggle with are reviewed more often until they stick.
You do not need to understand the math behind FSRS to use Temar effectively. Just answer questions honestly, rate your recall accurately, and the algorithm takes care of the rest.
How a review session works
When items are due for review, they appear on your Reviews page. Here is what a typical session looks like:
- 1
Visit the Reviews page
Navigate to Reviews in the dashboard sidebar. You will see a count of items currently due and a list of review cards waiting for you.
- 2
Read the question
Each review presents a question generated from one of your tracked Chunks. The question type (open-ended, MCQ, or LeetCode-style) determines how you will respond.
- 3
Write your answer
For open-ended questions, write your answer using the rich-text editor. For MCQ, select the correct option. For LeetCode-style, write your solution code. Take your time — the goal is to demonstrate understanding, not speed.
- 4
AI evaluates your response
For open-ended questions, the AI compares your answer against the grading rubric. It identifies which key points you covered, highlights strengths, and points out gaps or misconceptions. MCQ questions are evaluated instantly on the client side.
- 5
Rate your recall
After seeing the evaluation, you rate how well you remembered the material: Again, Hard, Good, or Easy. Each rating shows you when the next review will be scheduled, so you can make an informed choice.
Understanding the ratings
Your self-rating after each question is what drives the FSRS scheduling. Each rating tells the algorithm something different about your memory state:
| Rating | When to use it | Effect on scheduling |
|---|---|---|
| Again | You forgot completely or got the answer entirely wrong | Resets the interval — you will review this item again very soon (often within the same day) |
| Hard | You struggled significantly but eventually recalled some of the answer | Shortens the interval — you will review again in a few days |
| Good | You remembered the material with some effort — your answer was mostly correct | Normal progression — the interval increases at the standard rate, typically about a week |
| Easy | You recalled the answer instantly and completely, with no hesitation | Extends the interval significantly — you will not see this item for a much longer period |
Good is the most common rating
Most reviews should be rated Good. Reserve Easy for items you answered instantly and completely — if you had to think about it at all, Good is the right choice. Overusing Easy can push intervals too far out, leading to forgotten material later.
AI answer analysis
For open-ended questions, the AI does more than just tell you whether your answer was right or wrong. It provides a detailed analysis that includes:
- Key points covered — which rubric criteria your answer addressed successfully
- Strengths — what you explained well, including any insights or connections that went beyond the basics
- Gaps and weaknesses — specific points you missed or areas where your explanation was incomplete or inaccurate
- Suggestions — targeted feedback on what to focus on next time you review this material
This feedback loop is what makes Temar different from simple flashcard apps. Every review teaches you something, even when you get the answer right — the analysis might reveal subtle gaps you did not realize you had.
Past attempts are hidden for due items
When an item is due within the next 24 hours, your previous attempts and the AI's past analysis are hidden. This prevents you from simply re-reading your last answer instead of genuinely recalling the material. Past attempts become visible again once you complete the current review.
Review history
Every review you complete is recorded in your review history. You can access this from each Chunk's detail view to see:
- All past attempts — your previous answers and the AI's evaluations, showing how your understanding has evolved over time
- Performance trends — how your ratings have changed across reviews, helping you identify material that is getting easier or harder
- Scheduling details — the current stability and difficulty values, next review date, and how many times the item has been reviewed
For a broader view of your learning progress across all topics, visit the Analytics page where you can explore forgetting curves, review heatmaps, and topic-level performance breakdowns.
Tips for effective reviews
- Review consistently — short daily sessions are far more effective than long infrequent ones. Even 10 minutes a day keeps your schedule on track.
- Be honest with ratings — the algorithm only works well if your ratings accurately reflect your recall. Do not inflate your ratings to avoid seeing items again — you will just forget them later.
- Read the AI feedback carefully — the analysis often reveals gaps you did not notice. Use it as a learning opportunity, not just a grade.
- Do not skip due items — items are due because the algorithm predicts you are about to forget them. Skipping reviews lets memories decay, making the next review harder.