skills/build-review-plan/references/planning-heuristics.md

# Review planning heuristics

Apply these heuristics conservatively and label any assumption.

## Fit the constraint first

- Sum planned minutes and keep them within stated availability.
- Show the arithmetic equation for each available window and the overall total. Count buffers and transitions explicitly.
- Cross-check any prose claim about minutes per priority against the actual blocks; do not state an allocation total that the table does not support.
- Prefer several focused retrieval passes over one long rereading block when the deadline permits.
- Preserve a small buffer instead of allocating every available minute.

## Order work by evidence

1. Start with a brief diagnostic retrieval task.
2. Give more time to user-identified weak or important material.
3. Correct a mismatch against the learner's source before adding another source-backed task. Use integration only when one supplied sentence explicitly states the entire relationship.
4. Re-check weak material after a gap when the schedule permits.
5. Finish with another evidence-prioritized, single-topic source check. When only topic labels are available, do not create a mixed or comparison task.

## Make sessions executable

- Give each session one primary goal and an observable completion condition.
- Use the learner's actual source titles and topic names.
- Break sessions longer than 60 minutes into focused blocks with a short transition.
- Use a minimum viable fallback that protects the highest-priority retrieval task.
- When only topic labels are available, use content-neutral actions. Never invent subtopics, examples, mechanisms, components, likely confusions, or questions that were not supplied by the learner.

## Avoid false precision

- Do not infer topic weights, exam likelihood, memory strength, or grade impact unless the learner supplied that evidence.
- Do not prescribe exact spacing as scientifically optimal for an individual.
- Describe the plan as a practical starting point that can be adjusted with new quiz evidence.
- Do not present outside domain knowledge as planning detail. A topic label establishes priority and time allocation, not its factual contents.