How Mathematics Examination Works | How to Use Maths Past Papers, Mock Exams and Practice Tests

Past-paper practice in mathematics works when an examination paper is treated as a measurement instrument, not merely as a pile of questions to finish. A maths past paper can reveal method-selection gaps, weak prerequisite knowledge, slow algebra, calculator errors, poor checking, unfamiliar wording and pacing problems—but only if the student records where the first failure occurred and uses that evidence to design the next practice.

Understanding how to use maths past papers, mock exams and exam practice effectively prevents a common revision trap: repeatedly sitting full papers, collecting scores and hoping familiarity becomes mastery. Full papers are valuable for integration and timing, but targeted sets, delayed retests, mixed practice and error repair are often better tools for fixing a specific weakness.

This world-facing guide extends How Mathematics Examination Works, Maths Mark Schemes, Method Marks, Working and Partial Credit, and Maths Exam Time Management, Pacing and Difficult Questions. It does not duplicate the broader Examination Preparations hub; it owns the mathematics-specific paper-to-diagnosis-to-repair loop.

The 50-second answer

CHOOSE PURPOSE → ATTEMPT UNDER MATCHED CONDITIONS → MARK WITH OFFICIAL EVIDENCE → LOCATE FIRST FAILURE → CLASSIFY → REPAIR THE COMPONENT → RETEST WITH A FRESH QUESTION → RETURN TO FULL PAPER.

1. A score is an outcome, not a diagnosis

Two students can score sixty percent for completely different reasons. One may lack algebra knowledge; another may know every method but run out of time. Revision should respond to the mechanism producing the score, not the score alone.

2. Decide what the paper is measuring today

A first attempt can measure current integration. A timed attempt can test pacing. A topic-filtered paper can test retrieval. A reattempt can test repair. Do not use every paper for every purpose simultaneously.

3. Match conditions when readiness is the question

If you want evidence about examination performance, use the permitted calculator, formula resources, timing and interruption conditions as closely as practical. If the goal is learning, pause, annotate and use hints deliberately. Label which mode you are in.

4. Mark the first failure, not only the final wrong line

A wrong answer may begin with misreading the target, choosing the wrong formula, dropping a negative sign or rounding too early. The repair belongs at the earliest failing transition. Later errors may simply be consequences.

5. Separate selection from execution

If the student cannot decide whether a question needs similarity or trigonometry, more trigonometric arithmetic is not the first repair. If the method is chosen correctly but algebra fails, mixed-method recognition may already be strong. Train the failing component.

6. Immediate correction is not proof of learning

After seeing a solution, the route feels familiar. Test the same relationship later with a fresh question and no model visible. Successful delayed transfer is stronger evidence that the repair has become available independently.

7. Full papers and targeted sets have different jobs

A full paper tests integration, endurance and pacing. A six-question contrast set can repair reverse percentages far more efficiently than another entire paper. Move between scales according to the diagnosis.

8. Repeating the same paper can inflate familiarity

Remembering an answer or layout can make a reattempt look fluent. Reattempts are useful, but pair them with unseen transfer questions so recognition of the paper is not mistaken for general skill.

9. Error logs should lead to actions

“Careless” is too broad. Record “copied negative coefficient incorrectly,” “used original total after remainder changed,” or “stayed on stalled geometry for seven minutes.” Specific descriptions suggest specific drills.

10. The loop ends when performance transfers

A corrected past-paper question is not finished until the learner can solve a structurally related unseen question and later recognise the same relationship inside mixed work. Repair must survive both delay and disguise.

Part II. One hundred sixty past-paper decisions

11. Baseline paper

For baseline paper, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

12. Diagnostic paper

For diagnostic paper, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

13. Timed paper

For timed paper, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

14. Untimed paper

For untimed paper, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

15. Open-book attempt

For open-book attempt, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

16. Closed-book attempt

For closed-book attempt, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

17. Calculator conditions

For calculator conditions, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

18. Formula-sheet conditions

For formula-sheet conditions, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

19. Paper selection

For paper selection, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

20. Syllabus matching

For syllabus matching, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

21. Specification matching

For specification matching, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

22. Old-format papers

For old-format papers, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

23. Current-format papers

For current-format papers, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

24. Specimen papers

For specimen papers, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

25. Official past papers

For official past papers, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

26. Teacher-made mocks

For teacher-made mocks, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

27. Commercial mocks

For commercial mocks, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

28. Difficulty matching

For difficulty matching, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

29. Topic coverage

For topic coverage, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

30. Cumulative coverage

For cumulative coverage, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

31. Question freshness

For question freshness, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

32. Unseen questions

For unseen questions, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

33. Repeat papers

For repeat papers, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

34. Memory contamination

For memory contamination, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

35. Answer familiarity

For answer familiarity, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

36. Mark-scheme familiarity

For mark-scheme familiarity, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

37. Score recording

For score recording, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

38. Raw score

For raw score, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

39. Percentage score

For percentage score, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

40. Topic score

For topic score, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

41. Method score

For method score, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

42. Timing data

For timing data, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

43. Completion rate

For completion rate, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

44. Blank rate

For blank rate, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

45. Confidence rating

For confidence rating, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

46. Confidence calibration

For confidence calibration, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

47. First failure

For first failure, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

48. Misread target

For misread target, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

49. Missing knowledge

For missing knowledge, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

50. Method-selection error

For method-selection error, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

51. Representation error

For representation error, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

52. Formula-selection error

For formula-selection error, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

53. Algebra error

For algebra error, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

54. Arithmetic error

For arithmetic error, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

55. Calculator error

For calculator error, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

56. Sign error

For sign error, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

57. Unit error

For unit error, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

58. Rounding error

For rounding error, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

59. Domain error

For domain error, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

60. Interpretation error

For interpretation error, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

61. Proof gap

For proof gap, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

62. Checking failure

For checking failure, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

63. Pacing failure

For pacing failure, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

64. Restart failure

For restart failure, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

65. Error propagation

For error propagation, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

66. Dependent parts

For dependent parts, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

67. Independent parts

For independent parts, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

68. Careless label

For careless label, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

69. Error taxonomy

For error taxonomy, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

70. Error frequency

For error frequency, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

71. Error severity

For error severity, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

72. Error leverage

For error leverage, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

73. Repair priority

For repair priority, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

74. Prerequisite repair

For prerequisite repair, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

75. Worked example

For worked example, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

76. Self explanation

For self explanation, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

77. Contrast set

For contrast set, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

78. Blocked practice

For blocked practice, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

79. Mixed practice

For mixed practice, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

80. Interleaving

For interleaving, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

81. Retrieval practice

For retrieval practice, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

82. Spaced retest

For spaced retest, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

83. Delayed retest

For delayed retest, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

84. Fresh transfer

For fresh transfer, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

85. Near transfer

For near transfer, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

86. Far transfer

For far transfer, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

87. Same-topic transfer

For same-topic transfer, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

88. Mixed-topic transfer

For mixed-topic transfer, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

89. Timed transfer

For timed transfer, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

90. Untimed transfer

For untimed transfer, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

91. Full-paper return

For full-paper return, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

92. Partial-paper sets

For partial-paper sets, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

93. Section practice

For section practice, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

94. Question clusters

For question clusters, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

95. Difficulty ladders

For difficulty ladders, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

96. Easy-to-hard progression

For easy-to-hard progression, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

97. Hard-question practice

For hard-question practice, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

98. Unfamiliar-question practice

For unfamiliar-question practice, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

99. Proof practice

For proof practice, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

100. Diagram practice

For diagram practice, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

101. Calculator practice

For calculator practice, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

102. Non-calculator practice

For non-calculator practice, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

103. Number fluency

For number fluency, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

104. Algebra fluency

For algebra fluency, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

105. Method-selection drills

For method-selection drills, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

106. Checking drills

For checking drills, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

107. Pacing drills

For pacing drills, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

108. Formula drills

For formula drills, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

109. Mark-scheme study

For mark-scheme study, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

110. Examiner-report study

For examiner-report study, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

111. Model-answer study

For model-answer study, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

112. Alternative methods

For alternative methods, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

113. Correction notebook

For correction notebook, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

114. Error cards

For error cards, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

115. Revision queue

For revision queue, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

116. Weakness queue

For weakness queue, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

117. Mastery criteria

For mastery criteria, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

118. Exit criteria

For exit criteria, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

119. Retest threshold

For retest threshold, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

120. Paper spacing

For paper spacing, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

121. Paper conservation

For paper conservation, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

122. Resource scarcity

For resource scarcity, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

123. Paper reuse

For paper reuse, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

124. Question remixing

For question remixing, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

125. Number changes

For number changes, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

126. Condition changes

For condition changes, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

127. Context changes

For context changes, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

128. Reverse questions

For reverse questions, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

129. Explain-why questions

For explain-why questions, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

130. Teach-back

For teach-back, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

131. Peer explanation

For peer explanation, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

132. Teacher conference

For teacher conference, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

133. Hint levels

For hint levels, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

134. Scaffolding

For scaffolding, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

135. Fading support

For fading support, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

136. Independent attempt

For independent attempt, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

137. Second attempt

For second attempt, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

138. Third attempt

For third attempt, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

139. Solution exposure

For solution exposure, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

140. Solution copying

For solution copying, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

141. Solution reconstruction

For solution reconstruction, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

142. Closed-model reconstruction

For closed-model reconstruction, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

143. Partial cues

For partial cues, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

144. Worked-example completion

For worked-example completion, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

145. Self marking

For self marking, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

146. Teacher marking

For teacher marking, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

147. Peer marking

For peer marking, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

148. Official scheme

For official scheme, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

149. Marking uncertainty

For marking uncertainty, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

150. Partial credit

For partial credit, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

151. Follow-through

For follow-through, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

152. Score inflation

For score inflation, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

153. Score volatility

For score volatility, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

154. Mock anxiety

For mock anxiety, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

155. Fatigue

For fatigue, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

156. Break conditions

For break conditions, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

157. Session length

For session length, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

158. Paper order

For paper order, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

159. Morning/evening effects

For morning/evening effects, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

160. Environment

For environment, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

161. Interruptions

For interruptions, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

162. Device setup

For device setup, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

163. Stationery setup

For stationery setup, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

164. Time stamps

For time stamps, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

165. Question time

For question time, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

166. Selection time

For selection time, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

167. Execution time

For execution time, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

168. Checking time

For checking time, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

169. Return time

For return time, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

170. End scan

For end scan, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

171. Unfinished work

For unfinished work, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

172. Post-paper review

For post-paper review, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

173. Next-day review

For next-day review, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

174. Weekly review

For weekly review, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

175. Trend tracking

For trend tracking, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

176. Moving average

For moving average, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

177. False progress

For false progress, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

178. Plateaus

For plateaus, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

179. Regression

For regression, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

180. Retention

For retention, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

181. Transfer

For transfer, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

182. Readiness

For readiness, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

183. Prediction limits

For prediction limits, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

184. Exam-day simulation

For exam-day simulation, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

185. Final-week papers

For final-week papers, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

186. Tapering

For tapering, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

187. Overpractice

For overpractice, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

188. Burnout

For burnout, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

189. Sleep and performance

For sleep and performance, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

190. Paper debrief

For paper debrief, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

191. Next-action assignment

For next-action assignment, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

192. Repair loop

For repair loop, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

193. Verification loop

For verification loop, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

194. Release from repair

For release from repair, decide what evidence this practice activity is meant to produce. A past paper can measure integrated performance, expose a weakness or rehearse conditions, but the interpretation changes with the setup. Record the purpose before the attempt so the result is not overread afterwards.

Diagnosis drill. If the response fails, locate the first transition that stopped matching the question. Classify it narrowly: knowledge, selection, representation, execution, tool use, checking, interpretation or pacing. Avoid the label “careless” unless it is replaced by an observable action.

Repair drill. Choose the smallest practice format capable of training the failing component. A formula-selection error may need contrast pairs; slow arithmetic may need fluency practice; a proof gap may need reconstruction; a pacing failure may need timed mixed sets. Another full paper is not automatically the best repair.

Retest drill. After correction, wait long enough for immediate memory to fade and use a fresh question with the same underlying relationship. Then place that relationship back into mixed practice. Success should survive both delay and removal of the topic label.

Evidence rule. Treat a rising score as one signal, not a complete explanation. Check completion, error types, timing, question freshness and transfer. Improvement is strongest when several independent indicators move together.

Part III. Forty original past-paper diagnosis laboratories

Laboratory 1. reverse percentage

Observed response. Wrong:72×1.2 after20% discount.

Diagnosis and repair. Diagnosis: reference-base/model error. Repair with forward/reverse percentage contrast pairs.

Retest. Retest on unseen changed percentage and context. The retest should be sufficiently fresh that remembering the corrected answer cannot carry the learner.

Evidence note. Record whether the new question was solved independently, how long selection took and whether the same error reappeared under disguise. Do not close the repair merely because the original question now feels easy.

Return rule. Once the targeted component survives fresh practice, return it to a mixed set and then a full paper. This restores the retrieval and pacing demands that targeted repair temporarily removed.

Laboratory 2. ratio change

Observed response. Student repartitions old total after blue counters added.

Diagnosis and repair. Diagnosis: invariant error.

Retest. Practise changed-ratio pairs where different quantity stays fixed. The retest should be sufficiently fresh that remembering the corrected answer cannot carry the learner.

Evidence note. Record whether the new question was solved independently, how long selection took and whether the same error reappeared under disguise. Do not close the repair merely because the original question now feels easy.

Return rule. Once the targeted component survives fresh practice, return it to a mixed set and then a full paper. This restores the retrieval and pacing demands that targeted repair temporarily removed.

Laboratory 3. average speed

Observed response. Student averages30 and60 to45.

Diagnosis and repair. Diagnosis: method-selection/definition error.

Retest. Repair total-distance/total-time examples with unequal durations/distances. The retest should be sufficiently fresh that remembering the corrected answer cannot carry the learner.

Evidence note. Record whether the new question was solved independently, how long selection took and whether the same error reappeared under disguise. Do not close the repair merely because the original question now feels easy.

Return rule. Once the targeted component survives fresh practice, return it to a mixed set and then a full paper. This restores the retrieval and pacing demands that targeted repair temporarily removed.

Laboratory 4. algebra sign

Observed response. Correct method, copied −3 as +3.

Diagnosis and repair. Diagnosis: transcription/sign error.

Retest. Use bracketed substitution and line-to-line sign checks. The retest should be sufficiently fresh that remembering the corrected answer cannot carry the learner.

Evidence note. Record whether the new question was solved independently, how long selection took and whether the same error reappeared under disguise. Do not close the repair merely because the original question now feels easy.

Return rule. Once the targeted component survives fresh practice, return it to a mixed set and then a full paper. This restores the retrieval and pacing demands that targeted repair temporarily removed.

Laboratory 5. quadratic

Observed response. Finds one root only.

Diagnosis and repair. Diagnosis: completeness error.

Retest. Retest equations with two, repeated and no real roots. The retest should be sufficiently fresh that remembering the corrected answer cannot carry the learner.

Evidence note. Record whether the new question was solved independently, how long selection took and whether the same error reappeared under disguise. Do not close the repair merely because the original question now feels easy.

Return rule. Once the targeted component survives fresh practice, return it to a mixed set and then a full paper. This restores the retrieval and pacing demands that targeted repair temporarily removed.

Laboratory 6. inequality

Observed response. Fails to reverse sign dividing by negative.

Diagnosis and repair. Diagnosis: rule/meaning gap.

Retest. Use number-line demonstrations then mixed inequalities. The retest should be sufficiently fresh that remembering the corrected answer cannot carry the learner.

Evidence note. Record whether the new question was solved independently, how long selection took and whether the same error reappeared under disguise. Do not close the repair merely because the original question now feels easy.

Return rule. Once the targeted component survives fresh practice, return it to a mixed set and then a full paper. This restores the retrieval and pacing demands that targeted repair temporarily removed.

Laboratory 7. Pythagoras

Observed response. Uses on non-right triangle.

Diagnosis and repair. Diagnosis: theorem-condition error.

Retest. Contrast right/non-right diagrams before calculation. The retest should be sufficiently fresh that remembering the corrected answer cannot carry the learner.

Evidence note. Record whether the new question was solved independently, how long selection took and whether the same error reappeared under disguise. Do not close the repair merely because the original question now feels easy.

Return rule. Once the targeted component survives fresh practice, return it to a mixed set and then a full paper. This restores the retrieval and pacing demands that targeted repair temporarily removed.

Laboratory 8. trigonometry

Observed response. Wrong opposite/adjacent labels.

Diagnosis and repair. Diagnosis: representation error.

Retest. Relabel sides relative to changing target angles. The retest should be sufficiently fresh that remembering the corrected answer cannot carry the learner.

Evidence note. Record whether the new question was solved independently, how long selection took and whether the same error reappeared under disguise. Do not close the repair merely because the original question now feels easy.

Return rule. Once the targeted component survives fresh practice, return it to a mixed set and then a full paper. This restores the retrieval and pacing demands that targeted repair temporarily removed.

Laboratory 9. similarity

Observed response. Uses non-corresponding sides.

Diagnosis and repair. Diagnosis: correspondence error.

Retest. Mark vertex correspondence before ratio practice. The retest should be sufficiently fresh that remembering the corrected answer cannot carry the learner.

Evidence note. Record whether the new question was solved independently, how long selection took and whether the same error reappeared under disguise. Do not close the repair merely because the original question now feels easy.

Return rule. Once the targeted component survives fresh practice, return it to a mixed set and then a full paper. This restores the retrieval and pacing demands that targeted repair temporarily removed.

Laboratory 10. probability

Observed response. Uses replacement denominator when no replacement.

Diagnosis and repair. Diagnosis: state-update error.

Retest. Tree pairs differing only in replacement condition. The retest should be sufficiently fresh that remembering the corrected answer cannot carry the learner.

Evidence note. Record whether the new question was solved independently, how long selection took and whether the same error reappeared under disguise. Do not close the repair merely because the original question now feels easy.

Return rule. Once the targeted component survives fresh practice, return it to a mixed set and then a full paper. This restores the retrieval and pacing demands that targeted repair temporarily removed.

Laboratory 11. statistics

Observed response. Averages two group means.

Diagnosis and repair. Diagnosis: weighting error.

Retest. Reconstruct totals from mean×count. The retest should be sufficiently fresh that remembering the corrected answer cannot carry the learner.

Evidence note. Record whether the new question was solved independently, how long selection took and whether the same error reappeared under disguise. Do not close the repair merely because the original question now feels easy.

Return rule. Once the targeted component survives fresh practice, return it to a mixed set and then a full paper. This restores the retrieval and pacing demands that targeted repair temporarily removed.

Laboratory 12. histogram

Observed response. Reads height as frequency with unequal widths.

Diagnosis and repair. Diagnosis: representation-definition error.

Retest. Frequency-density contrast sets. The retest should be sufficiently fresh that remembering the corrected answer cannot carry the learner.

Evidence note. Record whether the new question was solved independently, how long selection took and whether the same error reappeared under disguise. Do not close the repair merely because the original question now feels easy.

Return rule. Once the targeted component survives fresh practice, return it to a mixed set and then a full paper. This restores the retrieval and pacing demands that targeted repair temporarily removed.

Laboratory 13. function

Observed response. Treats f(x) as f×x.

Diagnosis and repair. Diagnosis: notation meaning gap.

Retest. Mapping and substitution drills. The retest should be sufficiently fresh that remembering the corrected answer cannot carry the learner.

Evidence note. Record whether the new question was solved independently, how long selection took and whether the same error reappeared under disguise. Do not close the repair merely because the original question now feels easy.

Return rule. Once the targeted component survives fresh practice, return it to a mixed set and then a full paper. This restores the retrieval and pacing demands that targeted repair temporarily removed.

Laboratory 14. sequence

Observed response. Uses n rather than n−1 transitions.

Diagnosis and repair. Diagnosis: indexing error.

Retest. Draw term-position transitions. The retest should be sufficiently fresh that remembering the corrected answer cannot carry the learner.

Evidence note. Record whether the new question was solved independently, how long selection took and whether the same error reappeared under disguise. Do not close the repair merely because the original question now feels easy.

Return rule. Once the targeted component survives fresh practice, return it to a mixed set and then a full paper. This restores the retrieval and pacing demands that targeted repair temporarily removed.

Laboratory 15. calculus

Observed response. Calls every stationary point maximum.

Diagnosis and repair. Diagnosis: classification error.

Retest. Sign/second-derivative contrast sets. The retest should be sufficiently fresh that remembering the corrected answer cannot carry the learner.

Evidence note. Record whether the new question was solved independently, how long selection took and whether the same error reappeared under disguise. Do not close the repair merely because the original question now feels easy.

Return rule. Once the targeted component survives fresh practice, return it to a mixed set and then a full paper. This restores the retrieval and pacing demands that targeted repair temporarily removed.

Laboratory 16. integral

Observed response. Forgets +C.

Diagnosis and repair. Diagnosis: indefinite-integral form error.

Retest. Differentiate families differing by constants. The retest should be sufficiently fresh that remembering the corrected answer cannot carry the learner.

Evidence note. Record whether the new question was solved independently, how long selection took and whether the same error reappeared under disguise. Do not close the repair merely because the original question now feels easy.

Return rule. Once the targeted component survives fresh practice, return it to a mixed set and then a full paper. This restores the retrieval and pacing demands that targeted repair temporarily removed.

Laboratory 17. bounds

Observed response. Treats rounded measurement exact.

Diagnosis and repair. Diagnosis: precision model error.

Retest. Reverse-rounding interval drills. The retest should be sufficiently fresh that remembering the corrected answer cannot carry the learner.

Evidence note. Record whether the new question was solved independently, how long selection took and whether the same error reappeared under disguise. Do not close the repair merely because the original question now feels easy.

Return rule. Once the targeted component survives fresh practice, return it to a mixed set and then a full paper. This restores the retrieval and pacing demands that targeted repair temporarily removed.

Laboratory 18. capacity

Observed response. Rounds11.375 to11.

Diagnosis and repair. Diagnosis: contextual integer error.

Retest. Minimum/maximum feasibility pairs. The retest should be sufficiently fresh that remembering the corrected answer cannot carry the learner.

Evidence note. Record whether the new question was solved independently, how long selection took and whether the same error reappeared under disguise. Do not close the repair merely because the original question now feels easy.

Return rule. Once the targeted component survives fresh practice, return it to a mixed set and then a full paper. This restores the retrieval and pacing demands that targeted repair temporarily removed.

Laboratory 19. proof

Observed response. Lists examples for universal claim.

Diagnosis and repair. Diagnosis: evidence-type error.

Retest. General-variable proof reconstruction. The retest should be sufficiently fresh that remembering the corrected answer cannot carry the learner.

Evidence note. Record whether the new question was solved independently, how long selection took and whether the same error reappeared under disguise. Do not close the repair merely because the original question now feels easy.

Return rule. Once the targeted component survives fresh practice, return it to a mixed set and then a full paper. This restores the retrieval and pacing demands that targeted repair temporarily removed.

Laboratory 20. modelling

Observed response. Linear model chosen for repeated percentage growth.

Diagnosis and repair. Diagnosis: mechanism-selection error.

Retest. Compare fixed-addition versus fixed-proportion data. The retest should be sufficiently fresh that remembering the corrected answer cannot carry the learner.

Evidence note. Record whether the new question was solved independently, how long selection took and whether the same error reappeared under disguise. Do not close the repair merely because the original question now feels easy.

Return rule. Once the targeted component survives fresh practice, return it to a mixed set and then a full paper. This restores the retrieval and pacing demands that targeted repair temporarily removed.

Laboratory 21. calculator

Observed response. Correct formula, wrong degree/radian mode.

Diagnosis and repair. Diagnosis: tool-state error.

Retest. Mode checks before trig sets. The retest should be sufficiently fresh that remembering the corrected answer cannot carry the learner.

Evidence note. Record whether the new question was solved independently, how long selection took and whether the same error reappeared under disguise. Do not close the repair merely because the original question now feels easy.

Return rule. Once the targeted component survives fresh practice, return it to a mixed set and then a full paper. This restores the retrieval and pacing demands that targeted repair temporarily removed.

Laboratory 22. formula

Observed response. Uses diameter as radius.

Diagnosis and repair. Diagnosis: variable-map error.

Retest. Label formula symbols before substitution. The retest should be sufficiently fresh that remembering the corrected answer cannot carry the learner.

Evidence note. Record whether the new question was solved independently, how long selection took and whether the same error reappeared under disguise. Do not close the repair merely because the original question now feels easy.

Return rule. Once the targeted component survives fresh practice, return it to a mixed set and then a full paper. This restores the retrieval and pacing demands that targeted repair temporarily removed.

Laboratory 23. graph

Observed response. Reads one square as one unit despite scale5.

Diagnosis and repair. Diagnosis: scale extraction error.

Retest. Rescaled graph drills. The retest should be sufficiently fresh that remembering the corrected answer cannot carry the learner.

Evidence note. Record whether the new question was solved independently, how long selection took and whether the same error reappeared under disguise. Do not close the repair merely because the original question now feels easy.

Return rule. Once the targeted component survives fresh practice, return it to a mixed set and then a full paper. This restores the retrieval and pacing demands that targeted repair temporarily removed.

Laboratory 24. word problem

Observed response. Stops at remainder though asks per box.

Diagnosis and repair. Diagnosis: target/completeness error.

Retest. Write target phrase before solving. The retest should be sufficiently fresh that remembering the corrected answer cannot carry the learner.

Evidence note. Record whether the new question was solved independently, how long selection took and whether the same error reappeared under disguise. Do not close the repair merely because the original question now feels easy.

Return rule. Once the targeted component survives fresh practice, return it to a mixed set and then a full paper. This restores the retrieval and pacing demands that targeted repair temporarily removed.

Laboratory 25. pacing

Observed response. Spends12 minutes repeating same stalled route.

Diagnosis and repair. Diagnosis: move-on failure.

Retest. Timed sets with progress-based exit triggers. The retest should be sufficiently fresh that remembering the corrected answer cannot carry the learner.

Evidence note. Record whether the new question was solved independently, how long selection took and whether the same error reappeared under disguise. Do not close the repair merely because the original question now feels easy.

Return rule. Once the targeted component survives fresh practice, return it to a mixed set and then a full paper. This restores the retrieval and pacing demands that targeted repair temporarily removed.

Laboratory 26. checking

Observed response. Recalculates same wrong equation.

Diagnosis and repair. Diagnosis: non-independent checking.

Retest. Choose substitution/units/reconstruction checks. The retest should be sufficiently fresh that remembering the corrected answer cannot carry the learner.

Evidence note. Record whether the new question was solved independently, how long selection took and whether the same error reappeared under disguise. Do not close the repair merely because the original question now feels easy.

Return rule. Once the targeted component survives fresh practice, return it to a mixed set and then a full paper. This restores the retrieval and pacing demands that targeted repair temporarily removed.

Laboratory 27. rounding

Observed response. Rounds every intermediate line.

Diagnosis and repair. Diagnosis: precision workflow error.

Retest. Compare early-rounding versus preserved-precision routes. The retest should be sufficiently fresh that remembering the corrected answer cannot carry the learner.

Evidence note. Record whether the new question was solved independently, how long selection took and whether the same error reappeared under disguise. Do not close the repair merely because the original question now feels easy.

Return rule. Once the targeted component survives fresh practice, return it to a mixed set and then a full paper. This restores the retrieval and pacing demands that targeted repair temporarily removed.

Laboratory 28. exactness

Observed response. Gives decimal for exact π answer.

Diagnosis and repair. Diagnosis: answer-form error.

Retest. Exact/approximate instruction pairs. The retest should be sufficiently fresh that remembering the corrected answer cannot carry the learner.

Evidence note. Record whether the new question was solved independently, how long selection took and whether the same error reappeared under disguise. Do not close the repair merely because the original question now feels easy.

Return rule. Once the targeted component survives fresh practice, return it to a mixed set and then a full paper. This restores the retrieval and pacing demands that targeted repair temporarily removed.

Laboratory 29. mark scheme

Observed response. Correct result but requested proof omitted.

Diagnosis and repair. Diagnosis: command/evidence error.

Retest. Command-word sorting and proof reconstruction. The retest should be sufficiently fresh that remembering the corrected answer cannot carry the learner.

Evidence note. Record whether the new question was solved independently, how long selection took and whether the same error reappeared under disguise. Do not close the repair merely because the original question now feels easy.

Return rule. Once the targeted component survives fresh practice, return it to a mixed set and then a full paper. This restores the retrieval and pacing demands that targeted repair temporarily removed.

Laboratory 30. dependency

Observed response. Wrong part a poisons b and c.

Diagnosis and repair. Diagnosis: high-leverage unchecked intermediate.

Retest. Insert checkpoint before carry-forward. The retest should be sufficiently fresh that remembering the corrected answer cannot carry the learner.

Evidence note. Record whether the new question was solved independently, how long selection took and whether the same error reappeared under disguise. Do not close the repair merely because the original question now feels easy.

Return rule. Once the targeted component survives fresh practice, return it to a mixed set and then a full paper. This restores the retrieval and pacing demands that targeted repair temporarily removed.

Laboratory 31. confidence

Observed response. Changes correct answer because it feels odd.

Diagnosis and repair. Diagnosis: confidence calibration.

Retest. Require objective failed check before changing. The retest should be sufficiently fresh that remembering the corrected answer cannot carry the learner.

Evidence note. Record whether the new question was solved independently, how long selection took and whether the same error reappeared under disguise. Do not close the repair merely because the original question now feels easy.

Return rule. Once the targeted component survives fresh practice, return it to a mixed set and then a full paper. This restores the retrieval and pacing demands that targeted repair temporarily removed.

Laboratory 32. blank

Observed response. Leaves hard problem empty despite knowing formula.

Diagnosis and repair. Diagnosis: access under pressure.

Retest. Practise minimum viable evidence and restart notes. The retest should be sufficiently fresh that remembering the corrected answer cannot carry the learner.

Evidence note. Record whether the new question was solved independently, how long selection took and whether the same error reappeared under disguise. Do not close the repair merely because the original question now feels easy.

Return rule. Once the targeted component survives fresh practice, return it to a mixed set and then a full paper. This restores the retrieval and pacing demands that targeted repair temporarily removed.

Laboratory 33. repeat paper

Observed response. Score rises sharply on same paper.

Diagnosis and repair. Diagnosis: possible familiarity contamination.

Retest. Retest with unseen isomorphic questions. The retest should be sufficiently fresh that remembering the corrected answer cannot carry the learner.

Evidence note. Record whether the new question was solved independently, how long selection took and whether the same error reappeared under disguise. Do not close the repair merely because the original question now feels easy.

Return rule. Once the targeted component survives fresh practice, return it to a mixed set and then a full paper. This restores the retrieval and pacing demands that targeted repair temporarily removed.

Laboratory 34. solution exposure

Observed response. Can reproduce model immediately.

Diagnosis and repair. Diagnosis: recognition may masquerade as recall.

Retest. Close model, delay, reconstruct, transfer. The retest should be sufficiently fresh that remembering the corrected answer cannot carry the learner.

Evidence note. Record whether the new question was solved independently, how long selection took and whether the same error reappeared under disguise. Do not close the repair merely because the original question now feels easy.

Return rule. Once the targeted component survives fresh practice, return it to a mixed set and then a full paper. This restores the retrieval and pacing demands that targeted repair temporarily removed.

Laboratory 35. full papers

Observed response. Many papers completed, same percentage error repeats.

Diagnosis and repair. Diagnosis: repair loop missing.

Retest. Pause full papers for targeted base-reference work. The retest should be sufficiently fresh that remembering the corrected answer cannot carry the learner.

Evidence note. Record whether the new question was solved independently, how long selection took and whether the same error reappeared under disguise. Do not close the repair merely because the original question now feels easy.

Return rule. Once the targeted component survives fresh practice, return it to a mixed set and then a full paper. This restores the retrieval and pacing demands that targeted repair temporarily removed.

Laboratory 36. topic sets

Observed response. Excellent chapter scores, weak mixed paper.

Diagnosis and repair. Diagnosis: method-selection/access gap.

Retest. Interleave near-neighbour topics. The retest should be sufficiently fresh that remembering the corrected answer cannot carry the learner.

Evidence note. Record whether the new question was solved independently, how long selection took and whether the same error reappeared under disguise. Do not close the repair merely because the original question now feels easy.

Return rule. Once the targeted component survives fresh practice, return it to a mixed set and then a full paper. This restores the retrieval and pacing demands that targeted repair temporarily removed.

Laboratory 37. timed score

Observed response. Untimed strong, timed weak.

Diagnosis and repair. Diagnosis: pacing/fluency bottleneck.

Retest. Measure selection versus execution time. The retest should be sufficiently fresh that remembering the corrected answer cannot carry the learner.

Evidence note. Record whether the new question was solved independently, how long selection took and whether the same error reappeared under disguise. Do not close the repair merely because the original question now feels easy.

Return rule. Once the targeted component survives fresh practice, return it to a mixed set and then a full paper. This restores the retrieval and pacing demands that targeted repair temporarily removed.

Laboratory 38. untimed score

Observed response. Untimed weak too.

Diagnosis and repair. Diagnosis: knowledge/access precedes pacing.

Retest. Repair concept before speed. The retest should be sufficiently fresh that remembering the corrected answer cannot carry the learner.

Evidence note. Record whether the new question was solved independently, how long selection took and whether the same error reappeared under disguise. Do not close the repair merely because the original question now feels easy.

Return rule. Once the targeted component survives fresh practice, return it to a mixed set and then a full paper. This restores the retrieval and pacing demands that targeted repair temporarily removed.

Laboratory 39. examiner report

Observed response. Common cohort error matches student error.

Diagnosis and repair. Use report as clue, not substitute for personal diagnosis.

Retest. Retest student’s actual failing mechanism. The retest should be sufficiently fresh that remembering the corrected answer cannot carry the learner.

Evidence note. Record whether the new question was solved independently, how long selection took and whether the same error reappeared under disguise. Do not close the repair merely because the original question now feels easy.

Return rule. Once the targeted component survives fresh practice, return it to a mixed set and then a full paper. This restores the retrieval and pacing demands that targeted repair temporarily removed.

Laboratory 40. readiness

Observed response. One high mock score after familiar revision.

Diagnosis and repair. Evidence insufficient alone.

Retest. Look for repeated unseen performance under matched conditions. The retest should be sufficiently fresh that remembering the corrected answer cannot carry the learner.

Evidence note. Record whether the new question was solved independently, how long selection took and whether the same error reappeared under disguise. Do not close the repair merely because the original question now feels easy.

Return rule. Once the targeted component survives fresh practice, return it to a mixed set and then a full paper. This restores the retrieval and pacing demands that targeted repair temporarily removed.

Part IV. A 20-day past-paper repair cycle

Day 1. Paper → diagnosis → repair → transfer

Attempt a manageable section or mixed set under conditions chosen for today’s purpose. If measuring readiness, match examination conditions. If diagnosing knowledge, allow enough time to reveal the student’s best reasoning rather than turning every session into a speed test.

Mark with the official scheme where available and locate the first failure in every lost-mark response. Select at most a few high-leverage repairs for the day. Build short contrast or fluency sets around those mechanisms rather than trying to repair the entire paper at once.

Close the model solutions and retest later with fresh questions. Finish the cycle by placing at least one repaired relationship into mixed practice. Record independent success, selection time and whether support was needed.

Day 2. Paper → diagnosis → repair → transfer

Attempt a manageable section or mixed set under conditions chosen for today’s purpose. If measuring readiness, match examination conditions. If diagnosing knowledge, allow enough time to reveal the student’s best reasoning rather than turning every session into a speed test.

Mark with the official scheme where available and locate the first failure in every lost-mark response. Select at most a few high-leverage repairs for the day. Build short contrast or fluency sets around those mechanisms rather than trying to repair the entire paper at once.

Close the model solutions and retest later with fresh questions. Finish the cycle by placing at least one repaired relationship into mixed practice. Record independent success, selection time and whether support was needed.

Day 3. Paper → diagnosis → repair → transfer

Attempt a manageable section or mixed set under conditions chosen for today’s purpose. If measuring readiness, match examination conditions. If diagnosing knowledge, allow enough time to reveal the student’s best reasoning rather than turning every session into a speed test.

Mark with the official scheme where available and locate the first failure in every lost-mark response. Select at most a few high-leverage repairs for the day. Build short contrast or fluency sets around those mechanisms rather than trying to repair the entire paper at once.

Close the model solutions and retest later with fresh questions. Finish the cycle by placing at least one repaired relationship into mixed practice. Record independent success, selection time and whether support was needed.

Day 4. Paper → diagnosis → repair → transfer

Attempt a manageable section or mixed set under conditions chosen for today’s purpose. If measuring readiness, match examination conditions. If diagnosing knowledge, allow enough time to reveal the student’s best reasoning rather than turning every session into a speed test.

Mark with the official scheme where available and locate the first failure in every lost-mark response. Select at most a few high-leverage repairs for the day. Build short contrast or fluency sets around those mechanisms rather than trying to repair the entire paper at once.

Close the model solutions and retest later with fresh questions. Finish the cycle by placing at least one repaired relationship into mixed practice. Record independent success, selection time and whether support was needed.

Day 5. Paper → diagnosis → repair → transfer

Attempt a manageable section or mixed set under conditions chosen for today’s purpose. If measuring readiness, match examination conditions. If diagnosing knowledge, allow enough time to reveal the student’s best reasoning rather than turning every session into a speed test.

Mark with the official scheme where available and locate the first failure in every lost-mark response. Select at most a few high-leverage repairs for the day. Build short contrast or fluency sets around those mechanisms rather than trying to repair the entire paper at once.

Close the model solutions and retest later with fresh questions. Finish the cycle by placing at least one repaired relationship into mixed practice. Record independent success, selection time and whether support was needed.

Day 6. Paper → diagnosis → repair → transfer

Attempt a manageable section or mixed set under conditions chosen for today’s purpose. If measuring readiness, match examination conditions. If diagnosing knowledge, allow enough time to reveal the student’s best reasoning rather than turning every session into a speed test.

Mark with the official scheme where available and locate the first failure in every lost-mark response. Select at most a few high-leverage repairs for the day. Build short contrast or fluency sets around those mechanisms rather than trying to repair the entire paper at once.

Close the model solutions and retest later with fresh questions. Finish the cycle by placing at least one repaired relationship into mixed practice. Record independent success, selection time and whether support was needed.

Day 7. Paper → diagnosis → repair → transfer

Attempt a manageable section or mixed set under conditions chosen for today’s purpose. If measuring readiness, match examination conditions. If diagnosing knowledge, allow enough time to reveal the student’s best reasoning rather than turning every session into a speed test.

Mark with the official scheme where available and locate the first failure in every lost-mark response. Select at most a few high-leverage repairs for the day. Build short contrast or fluency sets around those mechanisms rather than trying to repair the entire paper at once.

Close the model solutions and retest later with fresh questions. Finish the cycle by placing at least one repaired relationship into mixed practice. Record independent success, selection time and whether support was needed.

Day 8. Paper → diagnosis → repair → transfer

Attempt a manageable section or mixed set under conditions chosen for today’s purpose. If measuring readiness, match examination conditions. If diagnosing knowledge, allow enough time to reveal the student’s best reasoning rather than turning every session into a speed test.

Mark with the official scheme where available and locate the first failure in every lost-mark response. Select at most a few high-leverage repairs for the day. Build short contrast or fluency sets around those mechanisms rather than trying to repair the entire paper at once.

Close the model solutions and retest later with fresh questions. Finish the cycle by placing at least one repaired relationship into mixed practice. Record independent success, selection time and whether support was needed.

Day 9. Paper → diagnosis → repair → transfer

Attempt a manageable section or mixed set under conditions chosen for today’s purpose. If measuring readiness, match examination conditions. If diagnosing knowledge, allow enough time to reveal the student’s best reasoning rather than turning every session into a speed test.

Mark with the official scheme where available and locate the first failure in every lost-mark response. Select at most a few high-leverage repairs for the day. Build short contrast or fluency sets around those mechanisms rather than trying to repair the entire paper at once.

Close the model solutions and retest later with fresh questions. Finish the cycle by placing at least one repaired relationship into mixed practice. Record independent success, selection time and whether support was needed.

Day 10. Paper → diagnosis → repair → transfer

Attempt a manageable section or mixed set under conditions chosen for today’s purpose. If measuring readiness, match examination conditions. If diagnosing knowledge, allow enough time to reveal the student’s best reasoning rather than turning every session into a speed test.

Mark with the official scheme where available and locate the first failure in every lost-mark response. Select at most a few high-leverage repairs for the day. Build short contrast or fluency sets around those mechanisms rather than trying to repair the entire paper at once.

Close the model solutions and retest later with fresh questions. Finish the cycle by placing at least one repaired relationship into mixed practice. Record independent success, selection time and whether support was needed.

Day 11. Paper → diagnosis → repair → transfer

Attempt a manageable section or mixed set under conditions chosen for today’s purpose. If measuring readiness, match examination conditions. If diagnosing knowledge, allow enough time to reveal the student’s best reasoning rather than turning every session into a speed test.

Mark with the official scheme where available and locate the first failure in every lost-mark response. Select at most a few high-leverage repairs for the day. Build short contrast or fluency sets around those mechanisms rather than trying to repair the entire paper at once.

Close the model solutions and retest later with fresh questions. Finish the cycle by placing at least one repaired relationship into mixed practice. Record independent success, selection time and whether support was needed.

Day 12. Paper → diagnosis → repair → transfer

Attempt a manageable section or mixed set under conditions chosen for today’s purpose. If measuring readiness, match examination conditions. If diagnosing knowledge, allow enough time to reveal the student’s best reasoning rather than turning every session into a speed test.

Mark with the official scheme where available and locate the first failure in every lost-mark response. Select at most a few high-leverage repairs for the day. Build short contrast or fluency sets around those mechanisms rather than trying to repair the entire paper at once.

Close the model solutions and retest later with fresh questions. Finish the cycle by placing at least one repaired relationship into mixed practice. Record independent success, selection time and whether support was needed.

Day 13. Paper → diagnosis → repair → transfer

Attempt a manageable section or mixed set under conditions chosen for today’s purpose. If measuring readiness, match examination conditions. If diagnosing knowledge, allow enough time to reveal the student’s best reasoning rather than turning every session into a speed test.

Mark with the official scheme where available and locate the first failure in every lost-mark response. Select at most a few high-leverage repairs for the day. Build short contrast or fluency sets around those mechanisms rather than trying to repair the entire paper at once.

Close the model solutions and retest later with fresh questions. Finish the cycle by placing at least one repaired relationship into mixed practice. Record independent success, selection time and whether support was needed.

Day 14. Paper → diagnosis → repair → transfer

Attempt a manageable section or mixed set under conditions chosen for today’s purpose. If measuring readiness, match examination conditions. If diagnosing knowledge, allow enough time to reveal the student’s best reasoning rather than turning every session into a speed test.

Mark with the official scheme where available and locate the first failure in every lost-mark response. Select at most a few high-leverage repairs for the day. Build short contrast or fluency sets around those mechanisms rather than trying to repair the entire paper at once.

Close the model solutions and retest later with fresh questions. Finish the cycle by placing at least one repaired relationship into mixed practice. Record independent success, selection time and whether support was needed.

Day 15. Paper → diagnosis → repair → transfer

Attempt a manageable section or mixed set under conditions chosen for today’s purpose. If measuring readiness, match examination conditions. If diagnosing knowledge, allow enough time to reveal the student’s best reasoning rather than turning every session into a speed test.

Mark with the official scheme where available and locate the first failure in every lost-mark response. Select at most a few high-leverage repairs for the day. Build short contrast or fluency sets around those mechanisms rather than trying to repair the entire paper at once.

Close the model solutions and retest later with fresh questions. Finish the cycle by placing at least one repaired relationship into mixed practice. Record independent success, selection time and whether support was needed.

Day 16. Paper → diagnosis → repair → transfer

Attempt a manageable section or mixed set under conditions chosen for today’s purpose. If measuring readiness, match examination conditions. If diagnosing knowledge, allow enough time to reveal the student’s best reasoning rather than turning every session into a speed test.

Mark with the official scheme where available and locate the first failure in every lost-mark response. Select at most a few high-leverage repairs for the day. Build short contrast or fluency sets around those mechanisms rather than trying to repair the entire paper at once.

Close the model solutions and retest later with fresh questions. Finish the cycle by placing at least one repaired relationship into mixed practice. Record independent success, selection time and whether support was needed.

Day 17. Paper → diagnosis → repair → transfer

Attempt a manageable section or mixed set under conditions chosen for today’s purpose. If measuring readiness, match examination conditions. If diagnosing knowledge, allow enough time to reveal the student’s best reasoning rather than turning every session into a speed test.

Mark with the official scheme where available and locate the first failure in every lost-mark response. Select at most a few high-leverage repairs for the day. Build short contrast or fluency sets around those mechanisms rather than trying to repair the entire paper at once.

Close the model solutions and retest later with fresh questions. Finish the cycle by placing at least one repaired relationship into mixed practice. Record independent success, selection time and whether support was needed.

Day 18. Paper → diagnosis → repair → transfer

Attempt a manageable section or mixed set under conditions chosen for today’s purpose. If measuring readiness, match examination conditions. If diagnosing knowledge, allow enough time to reveal the student’s best reasoning rather than turning every session into a speed test.

Mark with the official scheme where available and locate the first failure in every lost-mark response. Select at most a few high-leverage repairs for the day. Build short contrast or fluency sets around those mechanisms rather than trying to repair the entire paper at once.

Close the model solutions and retest later with fresh questions. Finish the cycle by placing at least one repaired relationship into mixed practice. Record independent success, selection time and whether support was needed.

Day 19. Paper → diagnosis → repair → transfer

Attempt a manageable section or mixed set under conditions chosen for today’s purpose. If measuring readiness, match examination conditions. If diagnosing knowledge, allow enough time to reveal the student’s best reasoning rather than turning every session into a speed test.

Mark with the official scheme where available and locate the first failure in every lost-mark response. Select at most a few high-leverage repairs for the day. Build short contrast or fluency sets around those mechanisms rather than trying to repair the entire paper at once.

Close the model solutions and retest later with fresh questions. Finish the cycle by placing at least one repaired relationship into mixed practice. Record independent success, selection time and whether support was needed.

Day 20. Paper → diagnosis → repair → transfer

Attempt a manageable section or mixed set under conditions chosen for today’s purpose. If measuring readiness, match examination conditions. If diagnosing knowledge, allow enough time to reveal the student’s best reasoning rather than turning every session into a speed test.

Mark with the official scheme where available and locate the first failure in every lost-mark response. Select at most a few high-leverage repairs for the day. Build short contrast or fluency sets around those mechanisms rather than trying to repair the entire paper at once.

Close the model solutions and retest later with fresh questions. Finish the cycle by placing at least one repaired relationship into mixed practice. Record independent success, selection time and whether support was needed.

Part V. Frequently asked questions

When should I start maths past papers?

Choose the practice format according to the information you need. Full timed papers are excellent for integration and pacing, but inefficient for repairing a narrow misconception. Targeted sets are excellent for repair, but cannot by themselves demonstrate whole-paper readiness.

For every correction, require a fresh retest after solution memory has faded. Then mix the repaired method with neighbouring methods so the learner must select it rather than merely execute it. That is the bridge from correction to examination readiness.

Should I finish the syllabus before using past papers?

Choose the practice format according to the information you need. Full timed papers are excellent for integration and pacing, but inefficient for repairing a narrow misconception. Targeted sets are excellent for repair, but cannot by themselves demonstrate whole-paper readiness.

For every correction, require a fresh retest after solution memory has faded. Then mix the repaired method with neighbouring methods so the learner must select it rather than merely execute it. That is the bridge from correction to examination readiness.

How many past papers should I do?

Choose the practice format according to the information you need. Full timed papers are excellent for integration and pacing, but inefficient for repairing a narrow misconception. Targeted sets are excellent for repair, but cannot by themselves demonstrate whole-paper readiness.

For every correction, require a fresh retest after solution memory has faded. Then mix the repaired method with neighbouring methods so the learner must select it rather than merely execute it. That is the bridge from correction to examination readiness.

Should every past paper be timed?

Choose the practice format according to the information you need. Full timed papers are excellent for integration and pacing, but inefficient for repairing a narrow misconception. Targeted sets are excellent for repair, but cannot by themselves demonstrate whole-paper readiness.

For every correction, require a fresh retest after solution memory has faded. Then mix the repaired method with neighbouring methods so the learner must select it rather than merely execute it. That is the bridge from correction to examination readiness.

When should I use untimed papers?

Choose the practice format according to the information you need. Full timed papers are excellent for integration and pacing, but inefficient for repairing a narrow misconception. Targeted sets are excellent for repair, but cannot by themselves demonstrate whole-paper readiness.

For every correction, require a fresh retest after solution memory has faded. Then mix the repaired method with neighbouring methods so the learner must select it rather than merely execute it. That is the bridge from correction to examination readiness.

Should I repeat the same past paper?

Choose the practice format according to the information you need. Full timed papers are excellent for integration and pacing, but inefficient for repairing a narrow misconception. Targeted sets are excellent for repair, but cannot by themselves demonstrate whole-paper readiness.

For every correction, require a fresh retest after solution memory has faded. Then mix the repaired method with neighbouring methods so the learner must select it rather than merely execute it. That is the bridge from correction to examination readiness.

How long should I wait before a reattempt?

Choose the practice format according to the information you need. Full timed papers are excellent for integration and pacing, but inefficient for repairing a narrow misconception. Targeted sets are excellent for repair, but cannot by themselves demonstrate whole-paper readiness.

For every correction, require a fresh retest after solution memory has faded. Then mix the repaired method with neighbouring methods so the learner must select it rather than merely execute it. That is the bridge from correction to examination readiness.

How do I mark a past paper properly?

Choose the practice format according to the information you need. Full timed papers are excellent for integration and pacing, but inefficient for repairing a narrow misconception. Targeted sets are excellent for repair, but cannot by themselves demonstrate whole-paper readiness.

For every correction, require a fresh retest after solution memory has faded. Then mix the repaired method with neighbouring methods so the learner must select it rather than merely execute it. That is the bridge from correction to examination readiness.

Should I use the official mark scheme immediately?

Choose the practice format according to the information you need. Full timed papers are excellent for integration and pacing, but inefficient for repairing a narrow misconception. Targeted sets are excellent for repair, but cannot by themselves demonstrate whole-paper readiness.

For every correction, require a fresh retest after solution memory has faded. Then mix the repaired method with neighbouring methods so the learner must select it rather than merely execute it. That is the bridge from correction to examination readiness.

What should go in an error log?

Choose the practice format according to the information you need. Full timed papers are excellent for integration and pacing, but inefficient for repairing a narrow misconception. Targeted sets are excellent for repair, but cannot by themselves demonstrate whole-paper readiness.

For every correction, require a fresh retest after solution memory has faded. Then mix the repaired method with neighbouring methods so the learner must select it rather than merely execute it. That is the bridge from correction to examination readiness.

Is ‘careless mistake’ a useful category?

Choose the practice format according to the information you need. Full timed papers are excellent for integration and pacing, but inefficient for repairing a narrow misconception. Targeted sets are excellent for repair, but cannot by themselves demonstrate whole-paper readiness.

For every correction, require a fresh retest after solution memory has faded. Then mix the repaired method with neighbouring methods so the learner must select it rather than merely execute it. That is the bridge from correction to examination readiness.

What should I do after a low mock score?

Choose the practice format according to the information you need. Full timed papers are excellent for integration and pacing, but inefficient for repairing a narrow misconception. Targeted sets are excellent for repair, but cannot by themselves demonstrate whole-paper readiness.

For every correction, require a fresh retest after solution memory has faded. Then mix the repaired method with neighbouring methods so the learner must select it rather than merely execute it. That is the bridge from correction to examination readiness.

What should I do after a high mock score?

Choose the practice format according to the information you need. Full timed papers are excellent for integration and pacing, but inefficient for repairing a narrow misconception. Targeted sets are excellent for repair, but cannot by themselves demonstrate whole-paper readiness.

For every correction, require a fresh retest after solution memory has faded. Then mix the repaired method with neighbouring methods so the learner must select it rather than merely execute it. That is the bridge from correction to examination readiness.

How do I know whether improvement is real?

Choose the practice format according to the information you need. Full timed papers are excellent for integration and pacing, but inefficient for repairing a narrow misconception. Targeted sets are excellent for repair, but cannot by themselves demonstrate whole-paper readiness.

For every correction, require a fresh retest after solution memory has faded. Then mix the repaired method with neighbouring methods so the learner must select it rather than merely execute it. That is the bridge from correction to examination readiness.

Should I practise only weak topics?

Choose the practice format according to the information you need. Full timed papers are excellent for integration and pacing, but inefficient for repairing a narrow misconception. Targeted sets are excellent for repair, but cannot by themselves demonstrate whole-paper readiness.

For every correction, require a fresh retest after solution memory has faded. Then mix the repaired method with neighbouring methods so the learner must select it rather than merely execute it. That is the bridge from correction to examination readiness.

How do I combine topic practice with full papers?

Choose the practice format according to the information you need. Full timed papers are excellent for integration and pacing, but inefficient for repairing a narrow misconception. Targeted sets are excellent for repair, but cannot by themselves demonstrate whole-paper readiness.

For every correction, require a fresh retest after solution memory has faded. Then mix the repaired method with neighbouring methods so the learner must select it rather than merely execute it. That is the bridge from correction to examination readiness.

How do I use examiner reports?

Choose the practice format according to the information you need. Full timed papers are excellent for integration and pacing, but inefficient for repairing a narrow misconception. Targeted sets are excellent for repair, but cannot by themselves demonstrate whole-paper readiness.

For every correction, require a fresh retest after solution memory has faded. Then mix the repaired method with neighbouring methods so the learner must select it rather than merely execute it. That is the bridge from correction to examination readiness.

How do I prevent memorising model answers?

Choose the practice format according to the information you need. Full timed papers are excellent for integration and pacing, but inefficient for repairing a narrow misconception. Targeted sets are excellent for repair, but cannot by themselves demonstrate whole-paper readiness.

For every correction, require a fresh retest after solution memory has faded. Then mix the repaired method with neighbouring methods so the learner must select it rather than merely execute it. That is the bridge from correction to examination readiness.

How should I practise in the final week?

Choose the practice format according to the information you need. Full timed papers are excellent for integration and pacing, but inefficient for repairing a narrow misconception. Targeted sets are excellent for repair, but cannot by themselves demonstrate whole-paper readiness.

For every correction, require a fresh retest after solution memory has faded. Then mix the repaired method with neighbouring methods so the learner must select it rather than merely execute it. That is the bridge from correction to examination readiness.

When is a weakness repaired?

Choose the practice format according to the information you need. Full timed papers are excellent for integration and pacing, but inefficient for repairing a narrow misconception. Targeted sets are excellent for repair, but cannot by themselves demonstrate whole-paper readiness.

For every correction, require a fresh retest after solution memory has faded. Then mix the repaired method with neighbouring methods so the learner must select it rather than merely execute it. That is the bridge from correction to examination readiness.

A creative-writing lens: revision is not rereading

A writer improves a draft by identifying what failed—a missing motive, flat scene, unclear reference or weak ending—and revising that mechanism. Simply rereading the same draft does not guarantee improvement. Mathematics past papers work similarly: the useful unit is not exposure to pages but a diagnosed weakness followed by a tested revision.

Use the eduKate ecosystem as a route

For broader revision planning, use Examination Preparations | The Complete Guide to Preparing for Any Exam. For interpreting mock scores, use How Practice Test Scores Work. For maths-specific checking, use How to Check Maths Answers and Find Careless Mistakes, and for scheme literacy use Maths Mark Schemes, Method Marks, Working and Partial Credit.

Scope note and final answer

Past-paper availability, formats, calculator rules and marking conventions vary by qualification and year. Use current official materials and respect copyright and access conditions. The examples in this article are original teaching material, not reproduced examination questions. Adrian, Jo, Ben, Aisha, Ryan, Mira, Clara and Ethan are fictional teaching characters.

The central habit is: do not count papers; count repaired mechanisms that survive fresh transfer. A past paper becomes powerful when it tells you what to practise next and when the next unseen question confirms that the repair worked.