Spaced Repetition System
How the SRS queue is built, what triggers it, and how the Confidence Rating feeds into the next session's Warm-up.
The Spaced Repetition Queue (SRS) is the engine's primary defense against memory decay. Forgetting is not random — it follows a predictable exponential curve. The SRS interrupts this curve by re-presenting flagged material at intervals calibrated to maximize long-term retention with minimum repetition overhead.
Trigger Conditions
A question enters the SRS queue under two conditions: you answered it incorrectly, or you answered it correctly but rated your confidence at 1–2 stars on the post-question rating. The second condition is the more important one. A correct answer with low confidence is a false positive — you happened to select the right option without a reliable underlying model. If the same question recurred under slightly different phrasing, you would likely fail.
The Confidence Rating (1–5 stars) is optional but strongly recommended. It is the mechanism through which you communicate genuine uncertainty to the engine. Use it accurately — gaming it upward harms only your own preparation.
How SRS Questions Are Delivered
The SRS does not create a separate revision module. It injects flagged questions directly into the Warm-up phase of future Daily Modules at algorithmically determined intervals. The interval for each question starts short (1–2 days) and expands as you demonstrate consistent correct recall. A question exits the SRS queue when you answer it correctly with a confidence rating of 4–5 stars across three consecutive encounters.
If you consistently skip the Warm-up phase or rate all questions at maximum confidence regardless of performance, the SRS queue will grow without clearing, and your Daily Modules will increasingly be dominated by review material. The SRS is most effective when you engage the Warm-up phase deliberately and rate confidence honestly.
The Concept Library Integration
When a question enters the SRS queue due to an incorrect answer tagged as a Conceptual Gap, the engine simultaneously surfaces a link to the relevant Concept Library article in the feedback view. The Concept Library contains topic articles and worked examples mapped to the same topic taxonomy as the question content. Reviewing the concept before the question re-appears in the Warm-up queue significantly improves re-encounter accuracy.