Daily Learning Module

Each session is a structured four-phase module: Warm-up → Core Focus → Integration → Reflection. This is not arbitrary sequencing — it mirrors the neurological rhythm of optimal retrieval practice as established by cognitive science research on the testing effect and spaced repetition.

Warm-up surfaces questions from the Spaced Repetition Queue (SRS) — concepts you previously answered incorrectly or rated low confidence on. The number of warm-up questions varies by your SRS backlog, typically 3–8 questions. This phase primes active retrieval and re-engages memory traces before new material is introduced.

Warm-up questions are drawn from across your full topic history, not just the current session's focus area. This cross-topic priming is intentional: it reduces the compartmentalization that leads to poor integration performance in the actual CAT paper.

Core Focus presents new questions at your current mastery level for the day's designated topic or topic cluster. This is the primary skill-building phase. Questions are drawn from the topic cluster with the highest leverage for your current proficiency level — determined by your accuracy history and the proximity of each topic to an unlock threshold.

The engine applies the ZPD band here: questions are targeted at the difficulty level just above your current operating ceiling, keeping accuracy in the 70–85% range. If you hit above 90% in Core Focus, the engine will escalate difficulty mid-session.

Integration introduces mixed-topic questions that require synthesizing concepts across domains. For example, a question that requires both Ratio reasoning and Work-Time logic, or a RC passage followed by a Parajumble that shares thematic content. This phase tests transfer — your ability to apply learned patterns in unfamiliar combinations.

Reflection generates a session summary: questions attempted, accuracy by phase, hints used, time distribution, and performance trend delta versus your 7-day baseline. Every recommendation surfaced in Reflection includes its data source — if the engine recommends revisiting a topic, it will state the exact accuracy and attempt count that triggered the recommendation.