Reading Comprehension Mock Tests | Build Accuracy, Stamina and Timing

51. Repair windows should sit between mocks

A useful mock cycle is simulation → review → targeted repair → transfer test → next simulation. Skipping the repair window turns mocks into repeated measurement without intervention. The repair may be short if the system is already stable: a few late-session source-separation tasks, one recovery drill, one digital-navigation rehearsal. The key is that the next mock should test a changed system rather than merely a different paper.

52. Use a mock question for each simulation

Before starting, define what the mock is meant to reveal beyond the total score. Examples: “Can I preserve answer scope in the final third?” “Does the new skip-return routine keep queue size manageable?” “Can I finish paired sources without source blending?” The learner should not receive strategic hints about specific questions, but teacher and review process should know what performance hypothesis is being tested.

53. Use mock families, not isolated papers

A sequence of standard mocks can be designed to vary one meaningful dimension across comparable sections: early versus late technical text, paper versus digital, more versus fewer paired sources, stronger versus weaker summary demands. This creates a family of observations rather than unrelated scores. The learner’s system can then be tested across controlled variation.

54. Mock families reveal hidden dependencies

If scores fall only when paired sources appear late, the problem may be fatigue × source attribution rather than source attribution alone. If performance drops only digitally, interface cost is implicated. A family of simulations can expose interactions that one paper cannot. The aim is not experimental perfection; it is better causal evidence about why performance changes.

55. Use standard mocks to estimate baseline

A standard mock should approximate the expected level and format without deliberate extra stress. Several such mocks provide the best estimate of ordinary performance range. Record score, completion, reserve, queue, late accuracy and first major deviation. A baseline is a band, not a single number.

56. Use stress mocks to probe one vulnerability

If recovery is the target, include an unusually hard early question while keeping the rest of the paper realistic. If stamina is the target, place paired sources late. If digital navigation is the target, use the real interface. Change one main stressor so the result remains interpretable. A stress mock should answer a focused resilience question.

57. Use dress rehearsals to confirm, not invent

Close to the examination, dress rehearsals should validate the final routine. Major strategy redesign should be rare unless the rehearsal reveals a serious unresolved defect. The learner should practise the same pacing cues, answer medium, permitted materials and recovery routine intended for exam day. Familiarity with the operating system matters more than adding new complexity at the end.

58. Mock review should separate score variance from system variance

A score can vary because the passage mix changes while the underlying process remains stable. System variance is different: one mock uses disciplined local search; the next has five broad rereads. Track whether core behaviours remain consistent across papers. A stable system can tolerate normal score movement without triggering unnecessary strategy changes.

59. Use a mock dashboard with only high-value metrics

A useful dashboard might contain total score, first-completion time, reserve, queue peak, late accuracy, longest-question time, broad-reread count, source-attribution errors and cascade count. Not every learner needs all of them. Select metrics linked to known vulnerabilities. The dashboard should illuminate the system, not turn the mock into data administration.

60. Score by passage position as well as question family

Question-family analysis alone can miss fatigue. Track whether the same inference skill performs differently early and late. Passage-position analysis can reveal stamina boundaries, while family analysis reveals conceptual weaknesses. The intersection is often the most useful: comparison errors only in the final third, for example.

61. Compare first-pass accuracy with final accuracy

If checking improves several answers, review which risks are being caught. If checking changes correct answers into wrong ones, calibration or review discipline is weak. If final accuracy is unchanged despite large checking time, the reserve may be poorly allocated. First-pass versus final comparison measures the actual return on checking.

62. Compare flagged-item accuracy with unflagged accuracy

Flags should correlate with genuine uncertainty. If many wrong answers are unflagged, premature confidence may be a problem. If most flags are actually correct and consume large review time, confidence may be under-calibrated. The pattern helps refine both return queues and checking priorities.

63. Compare skipped-item return value

For each returned item, ask whether the later time improved the answer. If many returns produce no new evidence or reasoning, return timing may be poor. If brief returns convert partial high-value answers efficiently, the queue is working. Mock review should evaluate the return system by marks recovered per unit of reserve, qualitatively if not numerically.

64. Compare reserve use across mocks

A learner may consistently finish with reserve but spend it inefficiently. Another may use reserve well but reach it inconsistently. Track whether reserve is protected, when it begins and what it buys: blanks completed, high-value flags resolved, mechanical errors caught. Reserve quality matters as much as reserve quantity.

65. Compare the longest-question tail

Average question time can look healthy while one item occasionally consumes six or seven minutes. The longest-question tail reveals catastrophic risk. Recovery training should reduce extreme overruns even if average speed barely changes. A narrower tail often means more stable total performance.

66. Compare broad-reread count across mocks

If full-passage rereads decline while accuracy remains stable or improves, passage mapping and evidence search are likely becoming more efficient. If rereads fall but errors rise, the learner may be under-reading. Process improvements should be judged together with outcome quality.

67. Compare answer-length stability

Answer length should respond to question demand, not emotional state. If responses expand after a hard question or shrink late in the section, the mock has revealed an execution instability. Track a few representative items rather than counting every word. Stable answer contracts are a readiness signal.

68. Compare source-attribution stability

Count source-ownership mistakes and note their position. A decline across mocks suggests transfer of source mapping. If mistakes remain clustered late, stamina is still the boundary. If they appear from the beginning, conceptual or notation repair may be needed.

69. Compare scope-preservation stability

Track a small sample of quantifiers, conditions and certainty markers from source to final answer. Late broadening may reveal fatigue; random broadening may reveal a stable scope weakness. Mock review can turn subtle language control into an observable performance metric.

70. Compare recovery cascade count

After the hardest question, how many later errors are plausibly attributable to rushing, overchecking, forgotten skips or source confusion? A shrinking cascade count can be a major readiness gain even if the hard item itself remains difficult. Recovery aims to localise damage.

71. Use mocks to test independence from teacher cues

No one should remind the learner to check quantifiers, build a comparison axis or preserve the reserve during a clean mock. If a skill appears only with prompts, the simulation has identified incomplete transfer. The repair belongs back in SETC-SEC-COMP-090, where support can be faded deliberately.

72. Use mocks to test independence from familiar passages

Mock material should be sufficiently fresh that evidence locations and model answers are not known. Familiar papers can still be used as rehearsals, but readiness conclusions should rely on unseen or genuinely unfamiliar material. Recognition memory can hide weak retrieval and passage modelling.

73. Use mocks to test medium-specific fluency

A learner may understand every question yet lose time to scrolling, typing, answer transfer or page flipping. The simulation should reveal whether the intellectual system can actually run through the medium. If not, mechanical rehearsal becomes a legitimate performance repair.

74. Use mocks to test start-of-section routine

How does the learner begin? Does she orient to instructions and passage architecture efficiently, or spend the first minutes overplanning? A stable opening routine reduces early variability. The first passage should not require a special warm-up that the real examination will not provide.

75. Use mocks to test passage transitions

After one passage ends, can the learner close its unresolved state, check time and begin the next passage cleanly? Passage boundaries can serve as cognitive firebreaks. If uncertainty from one text contaminates the next, transition routines deserve targeted practice.

76. Use mocks to test the hard checkpoint

A hard checkpoint should trigger proportional action, not panic. The mock reveals whether the learner can compare planned and actual progress, assess queue size and protect high-value later tasks. If the checkpoint causes global rushing, the pacing plan is not yet operational under pressure.

77. Use mocks to test the reserve transition

When first completion occurs, does the learner know what phase begins next? The reserve should shift from first-pass answering to unresolved queue and risk-based checking. If the learner restarts at question one automatically, return and checking priorities need repair.

78. Use mocks to test final-minute behaviour

In the last minute, the learner should prioritise blanks, unfinished components and mechanical safety rather than stylistic polishing. Simulations show whether endgame rules remain available under pressure. A good final-minute routine is brief enough to survive fatigue.

79. Use mocks to test strategic acceptance

Not every ambiguous item will reach perfect certainty. A ready learner can submit a defensible qualified answer and move. If one unresolved mark repeatedly absorbs reserve, strategic acceptance has not transferred. Mocks reveal whether uncertainty tolerance exists outside isolated recovery drills.

80. Use mocks to test confidence normalisation after difficulty

After a hard item, do later ordinary questions return to ordinary time and answer length? This is a direct measure of recovery maturity. One difficult question should not create a new paper-wide confidence state. The next three questions are often diagnostically valuable.

81. Build a mock sequence around current readiness

An early sequence may use one standard mock every few weeks with many targeted sessions between. Later, as components stabilise, full simulations can become more frequent. Close to the examination, dress rehearsals and light targeted repairs may dominate. Frequency should evolve with the learner rather than follow a fixed universal calendar.

82. Early mocks should be information-rich, not frequent

At the start of a preparation cycle, one well-reviewed simulation can reveal enough for several weeks of targeted work. Running many mocks before repairs have time to transfer wastes fresh papers and repeats known weaknesses. Early mock scarcity can improve information value.

83. Middle-phase mocks should test repaired boundaries

As transfer practice progresses, mocks should ask whether specific repairs survive integration. Use a standard paper but enter review with clear questions: has late scope control improved? Does the new return queue work? Is evidence search still local after fatigue? The score matters, but boundary confirmation is the main learning value.

84. Late-phase mocks should stabilise routine

Near the examination, avoid changing the operating system after every small fluctuation. Use simulations to confirm consistency, expose only material defects and maintain stamina. The training emphasis shifts from invention to fidelity: execute the system that has already been built.

85. Tapering can improve final readiness

Heavy mock volume immediately before an examination can produce fatigue without new learning. Depending on the learner and schedule, final days may use lighter unseen passages, short confidence-calibration tasks, review of operating routines and adequate rest rather than continuous full simulations. The principle is to arrive ready to perform, not merely recently exhausted by practice.

86. Do not chase a personal-best score in every mock

A harder passage mix can produce a lower mark even when the system is improving. If every simulation is treated as a contest with the previous score, learners may change strategies irrationally or avoid difficult material. Use score as one signal inside a broader readiness profile.

87. Do not hide poor process behind a high score

A learner can score well through familiarity, generous time, lucky choices or one easy passage mix. Review high-cost correct answers, broad rereads and unstable confidence. Readiness requires a process likely to survive a less favourable paper.

88. Do not hide strong process behind one low score

A mock can expose unusually difficult content while the learner still manages time, keeps uncertainty local and preserves answer scope. That performance may be more encouraging than the raw mark appears. Diagnose the score before changing a stable system.

89. Mocks should eventually become boring

A mature simulation should feel less dramatic. The learner encounters hard questions, flags them, continues, returns, checks and finishes without large strategy swings. Boring is desirable because the operating system is stable. The text remains intellectually variable; the response architecture becomes familiar.

90. Boring does not mean easy

The learner may still miss marks, feel uncertain and meet unfamiliar subject matter. Stability means those events do not cause cascading failure. A difficult mock can be handled calmly without producing a high score. Readiness is the ability to keep the paper structured even when the content is demanding.

91. Readiness includes knowing when not to change strategy

If the first passage runs slightly long but the drift remains within the plan’s tolerance, no global intervention is needed. If one answer feels uncertain but evidence is sufficient, no extra search is needed. Mocks teach restraint as well as action. Stable systems avoid unnecessary self-disruption.

92. Readiness includes knowing when to change strategy

A severe time debt, growing queue, repeated source confusion or broken interface routine may require proportional adjustment. The learner should recognise real triggers and respond with trained actions. Readiness is not rigid adherence to a plan; it is controlled adaptation without abandoning core accuracy rules.

93. Readiness includes stable self-assessment

The learner should be able to look at a checkpoint and judge whether she is ahead, on pace or behind without emotional distortion. She should know which flagged answers genuinely need return and which are merely low-confidence. Mocks calibrate this internal monitoring against actual outcomes.

94. Readiness includes reliable mechanical execution

Answer transfer, digital submission, page navigation, flag review and word-limit handling should be routine enough that they do not compete heavily with English reasoning. Mechanical errors are preventable and belong inside mock validation.

95. Readiness includes recoverability, not perfection

No simulation can guarantee that the real paper will match expectations. The more robust target is recoverability. Can the learner handle an unfamiliar passage, a bad opening, one hard question or a temporary time debt without losing access to the rest of the paper? A system that can recover is more valuable than one that performs perfectly only when conditions remain ideal.

96. Readiness includes stamina without late-paper distortion

The final third should not show a predictable collapse in source accuracy, scope, answer completeness or route discipline. Some decline in subjective energy is normal; the operational system should remain functional. If one late mechanism still decays, targeted late-session practice remains useful.

97. Readiness includes stable performance across variation

One narrative-heavy mock and one technical paired-source mock may produce different scores, but core behaviours should remain recognisable. Passage maps, evidence search, scope, ceilings and recovery should transfer across the surface variation. Stability across variation is stronger evidence than one excellent paper.

98. Readiness includes a short exam-day routine

By the final phase, the learner’s operating routine should fit on one small page or in memory: passage-map threshold, timing checkpoints, question ceiling, skip-state note, reserve trigger and checking order. If the system requires many new instructions immediately before the test, it is not yet sufficiently compressed.

99. Readiness is confirmed through repeated normality

The strongest signal is not one spectacular mock. It is several varied simulations in which the learner handles ordinary questions normally, difficult questions locally, time debt proportionally and the final minutes rationally. Repeated normality shows that the system is no longer dependent on a favourable paper.

100. The mock is successful when it tells you what to do next

A mock has value if it changes training intelligently, confirms that a repair transferred, exposes a new integration boundary or validates readiness. A score with no diagnosis, repair or readiness conclusion is incomplete information. Simulation belongs inside a learning loop, not at the end of it.

101. Practice Lab: the first mock exposes an early hard-question cascade

Leonie enters a standard mock with good untimed comprehension. Question four is a difficult inference. She spends six minutes, finally answers, then rushes the next passage and loses four easy marks. The post-mock review separates local difficulty from cascade. The repair window uses SETC-SEC-COMP-070: one trap question, state preservation, clean reset and ordinary questions afterward. The next mock asks a new question: can the learner contain an early block? The same local inference difficulty may still occur, but later accuracy should remain normal. Improvement is measured by cascade reduction, not only by solving question four faster.

102. Practice Lab: a high score hides repeated broad rereading

Iona scores well on an easier mock but rereads most passages from the top for each question. She finishes with almost no reserve. The score looks reassuring; the process is fragile. Review logs broad-reread count and evidence-search latency. The repair window returns to SETC-SEC-COMP-050 with local-corridor drills and longer passages. The next standard mock is not expected to produce a much higher score. Its readiness question is whether evidence search remains local enough to create a safer reserve.

103. Practice Lab: a lower score shows stronger recovery

Maren’s second mock is harder and her score falls slightly. Yet the longest-question tail shrinks, no easy marks are left blank, and the final summary receives its full working window. One hard inference remains wrong, but it does not contaminate later work. The process trend is stronger than the score trend. The training plan keeps the recovery system and targets the specific inference mechanism rather than redesigning pacing after a difficult paper.

104. Practice Lab: late source blending appears only in mocks

Iona performs paired-source drills accurately in short sessions. In two full mocks, however, Source A and Source B begin to blend in the final third. The diagnosis is fatigue × source attribution, not a basic comparison weakness. Repair moves paired-source mini-sets later in realistic sessions and simplifies source notes. A later mock places paired texts late again. Success is not merely the comparison score; it is preserved ownership under fatigue without extra time.

105. Practice Lab: digital navigation adds hidden minutes

Leonie’s paper mock is stable. A digital mock runs long because she loses paragraph position after every scroll and repeatedly reopens source tabs. Conceptual accuracy is unchanged when the correct region is located. The repair window uses the actual digital interface: paragraph landmarks, minimal highlighting, flags and split-screen practice where available. The next mock tests medium-specific fluency. The goal is to reduce mechanical search cost without changing the reading method itself.

106. Practice Lab: early surplus is spent on reassurance

Maren finishes the first passage three minutes ahead and rereads every answer twice. A later paired-source task is unexpectedly dense and the reserve is gone. Post-mock review identifies reserve misuse rather than poor paired-source timing. The next simulation keeps the same pacing model but changes the rule: early surplus remains protected until first completion unless a specific high-risk issue is identified. Readiness improves when easy early passages create resilience rather than extra polishing.

107. Practice Lab: the mock begins badly but the paper recovers

Iona’s first passage is technical and costs three extra minutes. At the passage boundary she logs the debt, one flagged item and remaining reserve. She begins the second passage with the normal map instead of global rushing. Several low-cost questions restore part of the schedule. The final section finishes with a small reserve. The mock is valuable because it proves the system can survive an unfavourable opening without a full-paper confidence collapse.

108. Practice Lab: the mock is paused for teaching

Halfway through a mock, Maren asks for clarification on a confusing command. The teacher decides the learning opportunity is more valuable than preserving simulation purity and pauses to teach. That is legitimate, but the session is reclassified as guided practice from that point. The earlier timing data remain useful; the final score is not treated as a clean readiness measure. Clear labels protect interpretation of the evidence.

109. Practice Lab: a stress mock is mistaken for baseline

Leonie completes a deliberately compressed-time stress mock and scores below her standard range. She interprets the result as regression. Review corrects the frame: the mock was designed to probe decision thresholds under reduced time, not predict ordinary score. The useful measures are whether local search survived, queue size remained manageable and high-value tasks stayed protected. Stress data should answer the stress question, not replace the standard baseline.

110. Practice Lab: a dress rehearsal reveals an answer-transfer risk

A late dress rehearsal uses the actual paper-and-answer-sheet workflow. Iona completes the English reasoning well but leaves several responses on the booklet and rushes transfer. The issue is mechanical, not conceptual. The final repair is small and specific: include transfer checkpoints and row alignment in one or two further rehearsals. No major reading strategy changes are introduced near the examination.

111. Practice Lab: the learner takes another mock too soon

Maren receives a disappointing score and completes another mock the next morning without reviewing the first. The same missed concessions and late overchecking recur. The second paper adds little new information. The revised cycle inserts a repair window and fresh mini-set before the next simulation. Mock frequency falls temporarily while learning value per mock rises.

112. Practice Lab: one excellent mock creates overconfidence

Leonie achieves a personal best on a familiar-topic mock with few paired sources. She assumes readiness is complete. The next standard mock uses an unfamiliar technical topic and later paired sources; source attribution and search cost deteriorate. The lesson is not that the first score was meaningless. It measured one favourable configuration. Readiness requires stability across variation rather than one maximum.

113. Practice Lab: one poor mock triggers unnecessary strategy change

Iona scores below her recent range and wants to replace passage-first reading with questions-first, remove all annotation and change answer order. Process review shows her ordinary routine remained sound; one unusually dense passage and one source-attribution error explain most of the drop. The strategy is preserved. Targeted repair addresses the actual defects. Stable systems should not be redesigned after normal variance.

114. Practice Lab: checking adds no value

Maren spends seven reserve minutes rereading high-confidence answers and changes none. One flagged multi-part answer remains incomplete. First-pass versus final comparison shows checking produced almost no return. The next mock uses a risk hierarchy: blanks, incomplete parts, flagged low-confidence work, mechanical issues. Reserve becomes more productive even without increasing its size.

115. Practice Lab: checking actively damages answers

Iona changes two correct inferences into more elaborate unsupported versions during final review. The mock reveals a calibration problem. The repair is error-triggered checking: an answer changes only when a specific defect is identified. The next simulation tracks first-pass and final accuracy separately. Success means checking corrects known risks without creating new speculation.

116. Practice Lab: first completion happens with six flags

Leonie reaches every question with twelve minutes left, which initially looks excellent. The queue contains six unresolved items, several of which were skipped before minimum engagement. The mock exposes premature skipping rather than strong completion. Repair tightens queue entry. The next mock may reach first completion slightly later but with fewer, higher-quality flags and a more manageable reserve phase.

117. Practice Lab: first completion happens late but the queue is empty

Maren reaches the final question with only two minutes left but has no unresolved items and high first-pass accuracy. Review asks whether the narrow reserve is acceptable or whether distributed small overruns create risk. The next mock tracks answer-length inflation and repeated verification. The goal is not to force an earlier finish at the cost of accuracy; it is to create a modest resilience buffer by removing waste.

118. Practice Lab: the fatigue curve is flat but subjective tiredness is high

Iona reports feeling exhausted in the final third, yet accuracy, search behaviour and scope remain stable. The mock shows subjective fatigue without operational collapse. Training does not automatically add more stamina work. The learner may need confidence that feeling tired does not equal performing badly, plus sensible rest and tapering before important assessments.

119. Practice Lab: subjective fatigue is low but performance collapses late

Leonie feels fine but late answers drop conditions and blend sources. The performance trace reveals fatigue before she notices it. The repair targets late-session scope and source tracking, not merely self-reported energy. Objective fatigue curves can reveal boundaries that subjective feeling misses.

120. Practice Lab: a standard mock becomes boring in the best way

Maren completes a varied unseen mock. One question is hard, one passage takes slightly longer, two answers are flagged. She contains the hard item, remains on pace, reaches first completion with reserve, returns by value and finishes. The score is good but not perfect. What matters is the absence of drama in the execution. The system behaves normally across normal variation.

121. Practice Lab: late answers begin dropping conditions

Iona’s early answers preserve “only if”, “unless” and “provided that” accurately. In the final third she begins paraphrasing conditional claims as unconditional facts. The overall score suggests a mild scope weakness, but the position trace shows a fatigue boundary. The repair is not another untimed lesson on conditional language. Targeted practice places scope-sensitive items late in realistic-duration sessions, with a compact final check: actor, degree, condition. A later mock asks whether those small controls survive when working memory is already loaded.

122. Practice Lab: final-third under-answering appears

Maren begins the mock with precise complete answers. Later she gives only one side of comparisons and misses second components in multi-part items. Her writing has become shorter, but not more efficient. Review separates fatigue from command misunderstanding because she can identify the missing components immediately after the timer stops. The repair is late-session answer-contract rehearsal: mark value, number of demands, minimum complete response. The next mock tracks completeness by position rather than total answer length alone.

123. Practice Lab: final-third over-answering appears

Leonie shows the opposite fatigue response. As confidence falls, one-mark answers expand into mini-essays. The mock ends without enough reserve for the final comparison. Review reveals that the extra sentences rarely add scoring value. The repair is not “write faster”. It is restore the response contract late in the session. Short timed clusters near the end of realistic practice require the learner to stop once the demanded components are complete. The next mock asks whether answer length remains stable even when fatigue rises.

124. Practice Lab: a high-confidence wrong answer is never flagged

Iona misses a whole-text purpose question and never revisits it because she is certain. The answer was fixed from the opening and ignored a later qualification. This is not a checking failure alone; premature closure occurred before the final phase. The repair uses contrast cases where early purpose is revised by a late turn, followed by confidence logging on selected global questions. A later mock should show fewer high-confidence errors even if the number of flags does not rise dramatically.

125. Practice Lab: a low-confidence correct answer consumes reserve

Maren flags a supported inference, returns twice and leaves the original answer unchanged. The answer was correct from the first pass. Her reserve is being spent on confidence rather than evidence. Review asks what specific defect she was trying to fix. She cannot name one. The repair is evidence-sufficiency calibration: once command, support and scope are complete, low confidence alone does not justify repeated return. The next mock compares flagged-item correctness and time spent revisiting each flag.

126. Practice Lab: a mock family reveals a digital-only weakness

Two paper mocks are stable. Two digital mocks show longer evidence-search time and more forgotten return items. The comparison isolates medium rather than content. The learner practises the same passages digitally and on paper with no score pressure, identifying where paragraph position, source tabs and flag review add cost. Once the mechanical routine stabilises, another digital mock confirms whether the gap narrows. Mock families make the hidden dependency visible because the intellectual task is otherwise similar.

127. Practice Lab: standard and stress mocks tell different stories

Leonie performs consistently in standard simulations but a compressed-time stress mock produces a large queue and several under-answers. The result does not invalidate the baseline. Instead, it locates the reserve of resilience: when timing pressure becomes substantially higher than normal, answer contracts begin to fail. The repair can strengthen decision thresholds without treating the stress score as predicted exam performance. Standard mocks continue to estimate readiness; stress mocks probe failure margins.

128. Practice Lab: the next mock shows the queue shrinking

After premature-skipping repair, Iona reaches first completion with three flags instead of seven. First-completion time is slightly later, but each flag represents a genuine block and all three are resumed cheaply. The raw score rises only a little. The process improvement is larger: the reserve is no longer overloaded. The mock demonstrates why queue quality matters more than finishing the first pass as early as possible.

129. Practice Lab: a repair works only in the first half

Maren’s comparison-axis training transfers beautifully in the first passage but disappears in a late paired-source task. She begins selecting evidence before stating the shared dimension. The mock therefore does not send her back to basic comparison drills. It moves the same transfer task later in realistic sessions and simplifies the source map under fatigue. The next simulation asks whether the repaired operation remains self-starting after sustained reading.

130. Practice Lab: a mock reveals a new upstream weakness

Leonie has repaired inference scope and timing, yet a new mock shows several errors in long technical sentences. Post-mock reconstruction reveals that the main clause and condition were misparsed before any inference began. This is progress: earlier bottlenecks have become stable enough for syntax under load to become visible. The next cycle targets sentence decomposition rather than adding more inference practice. Mocks are useful partly because they reveal the next bottleneck after old ones are removed.

131. Practice Lab: the final high-value task is protected

Iona reaches the hard checkpoint behind schedule with two low-value flags and a substantial comparison still ahead. She applies the trained route: the comparison receives its protected window; the low-value flags wait. She completes the high-value task and later resolves one flag. The mock score improves because the order of opportunity changed, not because the difficult low-value items became easier. Readiness includes protecting what matters before the final minutes decide for you.

132. Practice Lab: the final high-value task is sacrificed

In another mock, Maren spends the reserve trying to perfect two one-mark items and reaches the final summary with too little time. The visible losses occur in the summary, but the first execution error is return allocation. The repair uses value-aware queue ranking and a protected minimum window for the final task. The next mock tests whether low-value uncertainty can remain unresolved temporarily without stealing high-value setup time.

133. Practice Lab: a no-return interface changes the mock

A digital platform does not allow backward navigation. Standard skip-return metrics no longer apply. The mock instead measures bounded commitment, certainty calibration and whether uncertainty from one forced choice contaminates later questions. Review focuses on decision latency and post-question reset. The simulation is still valid because it reproduces the actual degrees of freedom. Mock design must follow the assessment environment rather than import strategies that the interface does not permit.

134. Practice Lab: paper answer transfer changes readiness

Leonie’s comprehension section is strong until the last minutes, when several responses still need to be transferred to a separate sheet. Earlier practice used direct writing in the booklet. The mock uncovers a workflow mismatch. Later rehearsals budget transfer time and include a row-alignment check. No extra comprehension drill is needed. Readiness includes completing the scoring workflow, not merely reaching the correct answer in working notes.

135. Practice Lab: unfamiliar topic, stable process

Iona encounters an article on an unfamiliar technical subject. She cannot define several terms precisely, yet the passage map, reference chains and causal relations remain stable. Evidence search is local and answer scope is accurate. The score stays within her normal band. This mock provides strong evidence of transfer: domain unfamiliarity does not dismantle the operating system. Training should not manufacture a vocabulary crisis when structural comprehension remains functional.

136. Practice Lab: score plateau with process improvement

Maren’s last four mock scores are similar. She worries that practice has stopped working. Process data show fewer broad rereads, lower longest-question time, smaller queues and better late accuracy. The papers have also become more varied. The plateau is reinterpreted as stable marks under greater process quality and surface difficulty. The next goal is not simply “raise score” but convert the stronger process into fewer recurring first-failure mechanisms.

137. Practice Lab: score rises while process degrades

Leonie scores higher on an easier paper but rereads more, spends the reserve poorly and guesses two difficult items. The score trend is positive; the readiness trend is not. The review preserves the existing repair programme rather than celebrating the score as proof that weaknesses disappeared. Mock interpretation should resist both pessimism and optimism when the process evidence points elsewhere.

138. Practice Lab: the final dress rehearsal is normal

A late dress rehearsal uses realistic timing and medium. Iona encounters one hard question, flags it, continues, returns once, completes every required task and finishes with a small reserve. The score sits inside her established range. No new major weakness appears. The correct response is not to search for something dramatic to change. The rehearsal confirms the routine. Final preparation can taper and maintain rather than redesign.

139. Practice Lab: too many late mocks create fatigue

Maren completes full simulations on several consecutive days before an important assessment. Her later scores fall, concentration declines and new “weaknesses” appear that were absent earlier. The training load itself has become a confounder. The plan shifts to lighter unseen work, operating-routine review and recovery. Simulation should prepare performance, not consume the resources needed for it.

140. Practice Lab: the integrated system finally stabilises

Across several varied mocks, Leonie’s scores move within a normal band. The more important features are stable: passage maps are functional, evidence search stays local, queue size remains selective, one hard question does not cascade, final-third accuracy holds and reserve is used by risk. Difficult passages still feel difficult. The system no longer depends on them being easy. This is the purpose of mock training: not to remove uncertainty from the paper, but to make the learner’s response to uncertainty reliable.

Technical Appendix A: Mock types and what each one can legitimately tell you

Standard mock: approximates normal target conditions and estimates the learner’s current ordinary performance band. It is the best source for readiness trends when conditions are comparable. Stress mock: deliberately increases one demand such as time compression, early difficulty, late paired sources or digital navigation. It probes resilience and should not be treated as a baseline prediction. Dress rehearsal: maximises fidelity close to the examination and confirms the final routine, medium, answer workflow and pacing. Guided simulation: resembles a mock but allows teacher intervention; it is useful for learning but should not be interpreted as a clean readiness score. Partial mock: reproduces one substantial section under realistic timing and is useful when full-section fatigue is not the current question.

The same passage set can therefore serve different purposes depending on conditions. Naming the mock type before analysis prevents category errors. A low stress-mock score may coexist with strong standard readiness. A high guided-simulation score may overstate independent performance. A dress rehearsal with stable execution can be more informative than a slightly higher ordinary practice mark.

Technical Appendix B: Pre-mock fidelity checklist

• Material: sufficiently fresh for the purpose of the mock.
• Time: realistic for a standard or dress rehearsal; deliberately documented if compressed.
• Medium: paper, digital interface, answer sheet or typing environment reasonably matches the target.
• Help: no ordinary hints during a clean mock.
• Resources: only those permitted or deliberately included by the target assessment.
• Environment: interruptions reduced enough that the performance is interpretable.
• Logging: limited to preselected high-value markers so observation does not distort execution.
• Purpose: one readiness question is defined before the mock begins.
• Post-mock window: enough time is reserved later for review rather than immediately launching another paper.

Fidelity does not mean recreating every environmental detail perfectly. It means removing obvious mismatches that would make the result answer the wrong question. If a platform fails or the learner is interrupted repeatedly, note the event and interpret the score cautiously rather than pretending the simulation remained clean.

Technical Appendix C: Minimal in-mock logging sheet

A minimal log can fit in a small table or margin note. Record the start time, passage-boundary times, hard checkpoint, first major block, first-completion time, queue count at first completion, reserve remaining and final submission time. Optionally mark one or two confidence states on especially uncertain or changed answers.

The log should never require the learner to stop after each question. If a timestamp takes several seconds and breaks concentration, the measurement is too intrusive for a normal mock. Teachers observing externally can collect more detail, but self-administered simulations should keep logging sparse. The script itself supplies additional process evidence later: flags, crossed-out answers, paragraph marks, incomplete subparts and visible search traces.

Afterward, transform the sparse timestamps into a timeline: on pace → question seven block → two-minute debt → passage two normal → first completion → reserve queue → final check. This timeline becomes the spine of the post-mock review.

Technical Appendix D: Mock dashboard

A compact dashboard might include ten measures: total score, passage scores, first-completion time, reserve, queue peak, late-third accuracy, longest-question time, broad-reread count, cascade-error count and one active first-failure mechanism. Add source-attribution or scope-preservation metrics only when they are known vulnerabilities.

Interpret measures together. A faster first-completion time is good only if accuracy and queue quality remain stable. A smaller queue is good only if the learner has not stopped flagging genuine uncertainty. Fewer broad rereads are good only if passage understanding has not become shallow. A higher score is encouraging, but if it is accompanied by larger longest-question tails and poor reserve use, readiness may not have improved as much as the mark suggests.

The dashboard’s purpose is decision support. It should help answer what to repair, what to maintain and what not to change. Metrics that do not alter any training decision can usually be dropped.

Technical Appendix E: Building and reading a fatigue curve

Divide the section into early, middle and late bands, or use natural passage boundaries. Compare accuracy, answer completeness, source ownership, scope, search behaviour and question time across positions. A fatigue curve does not require mathematical modelling; a simple pattern can be enough. For example: early 90% accurate with local search, middle 85% with normal answer length, late 70% with source blending and dropped conditions.

Then test competing explanations. Move a similar difficult passage earlier. Place a familiar-topic passage late. If the error follows position, stamina is implicated. If it follows text family regardless of position, the component skill is more likely responsible. If both matter, practise their interaction.

The goal is not to eliminate feeling tired. It is to keep the operating system functional while tired. A strong fatigue curve may show stable accuracy and process even when subjective energy falls.

Technical Appendix F: Post-mock review protocol

1. Reconstruct the timeline before studying every wrong answer.
2. Identify the first major deviation: overrun, broad reread, source confusion, queue growth or confidence shock.
3. Separate local hard questions from downstream cascade.
4. Review passage maps and source states for clustered errors.
5. Use SETC-SEC-COMP-080 to diagnose one or two high-leverage first failure points.
6. Compare confidence with correctness on selected items.
7. Compare first-pass with final answers to evaluate checking.
8. Review queue entry, breadcrumbs, return order and reserve use.
9. Choose a small repair list: usually one component mechanism and one execution behaviour.
10. Design fresh transfer practice through SETC-SEC-COMP-090 before scheduling the next mock.

The review should end with testable next steps, not a page of observations. “Practise source separation late in the session” is actionable. “Need more focus” is not. “Preserve conditions in paraphrase on five fresh passages, then retest inside a timed mini-set” creates a closed loop from diagnosis to transfer.

Technical Appendix G: Standard → stress → dress-rehearsal sequence

A useful progression begins with standard mocks to establish an ordinary band. Once major component skills are stable, a stress mock can probe one known vulnerability: perhaps a hard early question, reduced reserve or late source synthesis. The repair that follows should be targeted. A later standard mock confirms that the system remains strong under normal conditions. Dress rehearsals come after this, when the main question is no longer “what skill should we teach?” but “can the final routine execute reliably under realistic conditions?”

Do not force the sequence mechanically. A learner with unstable basics may need weeks of transfer practice between standard mocks. Another close to an assessment may use one stress probe followed by a dress rehearsal. The logic is what matters: baseline, targeted resilience test, repair, fidelity confirmation.

Technical Appendix H: Four-week mock cycle

Week 1: one standard comprehension mock. Review timeline, first-failure points, fatigue curve and reserve. Select one major component repair and one execution repair. Week 2: targeted transfer work through short passages and mini-sets; no need for another full mock unless the first result was unusable. Week 3: delayed transfer tests plus one stress mini-simulation aimed at the known vulnerability. Week 4: another standard mock under comparable conditions. Ask whether the repair survived integration and whether the system’s variance narrowed.

This four-week structure deliberately leaves more days for learning than for measuring. It is a template, not a universal schedule. Closer to an examination, the cycle can compress; earlier in preparation, it can expand.

Technical Appendix I: Six-week mock programme

Week 1: Baseline standard mock. Establish the ordinary score/process band and identify first failure points. Week 2: Repair and transfer. No full mock is required unless the first result was unusable; use targeted work and unseen mini-sets. Week 3: Standard mock focused on repair integration. Ask whether the repaired mechanism survives realistic pacing. Week 4: Stress probe. Increase one known vulnerability—perhaps late paired sources or an early hard item—then repair. Week 5: Standard mock again. Confirm that the stress repair did not destabilise ordinary performance. Week 6: Dress-rehearsal-style simulation. Use the real medium and final operating routine with minimal experimentation.

The sequence deliberately alternates measurement and learning. A learner with larger component gaps may stretch the programme and place several weeks between full mocks. A learner near an examination may compress it. The principle remains: every simulation should either establish a baseline, probe a boundary, confirm a repair or validate final readiness.

Technical Appendix J: Twelve-week simulation arc

Weeks 1–2 establish baseline and diagnose major comprehension mechanisms. Weeks 3–4 emphasise SETC-SEC-COMP-090 transfer work, with one standard mock near the end of the phase. Weeks 5–6 add timing, route and recovery stress through partial and full simulations. Weeks 7–8 compare standard mocks across varied text families and medium where relevant. Weeks 9–10 use a small number of targeted stress mocks plus targeted repair. Weeks 11–12 emphasise dress rehearsal, tapering and stable execution rather than major new strategy changes.

Older repaired mechanisms remain in light maintenance throughout. The active repair list should shrink rather than grow. If every mock produces a completely new operating system, the review process is overreacting to variance or the foundational diagnosis remains weak.

Technical Appendix K: Teacher mock protocol

1. Define the mock type and one main readiness question.
2. Confirm material freshness and realistic conditions.
3. Collect only minimal in-mock data unless observing externally.
4. Do not intervene during ordinary difficulty in a clean simulation.
5. Afterward, reconstruct the timeline before opening the full error list.
6. Review the first major deviation and any cascade.
7. Use SETC-SEC-COMP-080 for one or two high-leverage first failure points.
8. Compare late versus early performance for fatigue effects.
9. Review reserve, queue and checking return.
10. Choose a small repair plan.
11. Use SETC-SEC-COMP-090 to transfer the repair on fresh material.
12. Schedule the next mock only when it can answer a new question.

The teacher should resist the temptation to explain every lost mark during one review. Mocks create abundant data; useful teaching requires prioritisation. The learner should leave knowing what changed the score most, what will be repaired next and which parts of the system are already stable enough to leave alone.

Technical Appendix L: Self-study mock protocol

Before the mock, write one sentence: “Today I am testing whether…” Examples include finishing with a usable reserve, preserving source identity late, or containing one hard question. Run the simulation cleanly. Immediately afterward, before deep marking, note passage times, first block, queue, reserve and subjective fatigue. Then mark and choose no more than three diagnostic items.

For each item, reconstruct the first failure point. Do not merely copy the correct answer. Design one short fresh repair exercise and one delayed transfer test. Record whether the mock’s original readiness question was answered. The next simulation should not happen solely because a calendar says “mock day”; it should happen when enough repair has occurred that a new full-section test can add information.

Technical Appendix M: Parent support around mocks

Parents can help keep mocks informative by avoiding score-only interrogation. Useful questions are: “Where did the paper first become difficult?” “Did that difficulty spread?” “What stayed stable?” “What one thing are you repairing before the next mock?” “Was this a standard mock or a stress mock?” These questions encourage mechanism-based review.

A disappointing score should not automatically trigger another paper the next day. A strong score should not automatically end targeted practice. Encourage the learner to use the result as evidence about readiness rather than as a verdict about ability. Close to important assessments, protect rest and tapering when additional full simulations would add fatigue more than information.

Technical Appendix N: Full worked mock walkthrough — early disruption, stable finish

Imagine a seventy-minute comprehension section with three passage groups: narrative, argumentative paired sources and a final summary task. The learner’s tested plan includes a small reserve and two major checkpoints. The narrative begins normally. Passage mapping is efficient and the first questions are straightforward. Then a two-mark motive inference becomes ambiguous. Two readings remain viable after a serious local attempt. The learner spends slightly longer than planned and notices repeated state.

SETC-SEC-COMP-070 activates. The learner writes a short breadcrumb—“P5; embarrassment strongest; pride possible; need second clue”—flags the item and moves. The next two questions are handled normally. At the first passage boundary, the section is ninety seconds behind with one unresolved item. The learner does not rush the paired sources. She begins with the usual source maps.

Source A and Source B are more demanding than expected. The first source-specific questions proceed well. A comparison question is high-value and has high restart cost, so when it runs slightly over but the shared axis is stable, the learner allows a modest extension rather than skipping mechanically. At the hard checkpoint, cumulative drift is two minutes, but the reserve remains partly intact and the final summary’s protected window is still safe.

The learner makes proportional adjustments: short direct answers remain concise, annotation is reduced to paragraph jobs and no completed high-confidence answer is checked immediately. Later, one narrative-related question in the section activates the same emotional pattern as the skipped motive inference. That new context does not prove the earlier answer, but it strengthens the embarrassment reading. At the next natural boundary the learner returns, completes the inference quickly and removes it from the queue.

The final summary begins with slightly less reserve than planned but enough time for proper selection. Causes are separated from consequences before writing. First completion occurs with four minutes left and no blanks. The remaining time is used for one low-confidence comparison and a mechanical scan for incomplete subparts.

Post-mock review shows that the score is within the learner’s normal range. The more important evidence is systemic: an early block did not cascade, a high-restart-cost comparison was given elastic time appropriately, the final high-value task remained protected and the return queue never exceeded one item. The mock validates recovery and pacing integration even though the opening inference itself was genuinely difficult.

Technical Appendix O: Full worked mock walkthrough — fatigue boundary discovered

Now consider a learner whose early passages are consistently strong. In a standard mock she builds accurate maps, answers direct and inferential questions efficiently and reaches the halfway point on schedule. The final third contains paired sources and a graph. The learner feels only mildly tired, but source notes become less precise. Source B’s conditional recommendation is accidentally attributed to Source A. A graph question also drops the unit attached to the vertical axis.

The final score is respectable, so a score-only review might miss the problem. Position analysis shows both errors occur late, while similar source and graph tasks completed early in previous practice are accurate. Untimed retesting immediately after a break is also accurate. The leading diagnosis is integration under fatigue rather than missing conceptual knowledge.

The repair window therefore places short paired-source and graph tasks after forty to fifty minutes of mixed reading. The learner uses a simpler late-session source map and a mandatory variable-unit-group check for graphs. Full mocks are not repeated immediately; targeted late-position work first confirms the repair.

In the next standard mock, late source attribution remains correct and graph units are preserved. The total score rises only modestly, but final-third accuracy moves closer to the early-section baseline. The fatigue curve has flattened. This is strong evidence of readiness improvement because the system has become more position-stable.

Technical Appendix P: Final readiness dashboard and taper decision

Before reducing simulation volume, review several dimensions together. Component stability: major first-failure mechanisms have transferred on fresh texts. Timing stability: checkpoint drift remains within a manageable range across varied standard mocks. Recovery: one hard question no longer creates a large cascade. Stamina: final-third accuracy and scope remain functional. Queue quality: flags are selective and restart cost is low. Reserve: first completion usually leaves enough time for unresolved work and high-risk checking. Medium fluency: paper or digital mechanics no longer create major hidden cost. Self-assessment: confidence is reasonably calibrated and the learner knows when a strategy change is actually needed.

If most dimensions are stable, more full mocks may produce diminishing returns. Shift toward light maintenance, a small amount of fresh unseen work, operating-routine review and sufficient recovery. If one dimension remains unstable, target it specifically rather than returning to indiscriminate paper volume. Tapering is not stopping preparation; it is protecting the performance system from unnecessary fatigue once the evidence says integration is reliable.

141. How SETC-SEC-COMP-100 connects to the lane

This page owns comprehension-specific full-section simulation and readiness validation. It integrates the passage model from 020, pacing from 030, order from 040, evidence search from 050, ambiguity control from 060, recovery from 070, diagnosis from 080, and transfer practice from 090.

The broad cross-paper practice owner remains English Exam Practice Design. Return to SETC-SEC-COMP-000 for the complete Secondary English comprehension performance system.

142. Final principle: the mock is where the system proves it can survive reality

Drills can make one mechanism strong. Transfer practice can make it portable. A mock asks whether all those mechanisms can still cooperate when the learner is timed, uncertain, tired, switching passages and making decisions without prompts.

Maren protects structure and complete answers. Iona protects evidence, source and scope. Leonie protects time, route, recovery and the final minutes. When those roles remain stable across varied simulations, the mock becomes less a test of luck and more a validation of an operating system.

DO NOT USE A MOCK ONLY TO ASK, “WHAT SCORE DID I GET?” USE IT TO ASK, “DID THE SYSTEM STILL WORK WHEN THE PAPER STOPPED BEING PREDICTABLE?”