How Mock Exams Fail | Why Simulation Without Fidelity, Diagnosis and Repair Creates False Readiness

A mock examination is supposed to answer a serious question.

If the real examination were tomorrow, what would probably happen?

That question sounds simple.

It is not.

A mock examination is not the real event. It is a model of the real event. Like every model, it keeps some features, removes others, and makes assumptions about what matters.

If the wrong features are preserved, the mock can become misleading theatre. The room looks serious. The timer runs. The student sits quietly. Pages turn. A score appears. Everybody feels that something authentic has happened.

But authenticity of appearance is not the same as validity of evidence.

A mock can overestimate readiness because the learner received hints, paused the clock, used the wrong paper format, practised only familiar sources, revised the exact topics immediately beforehand, marked generously, or repeated the same paper until memory carried the score.

A mock can underestimate readiness because it was deliberately harder than the real assessment, taken in an exhausted state, marked harshly for motivation, or overloaded with unusual questions that would not be representative of the target examination.

A mock can also be perfectly representative and still fail educationally because nobody does anything useful with the result.

The student receives 67%.

The number goes into a spreadsheet.

Another mock is scheduled.

The same weaknesses appear again.

That is not a simulation system.

That is repeated measurement without controlled repair.

This article examines the failure edge.

Alicia, Tricia and Kai Kai return as the resident learners of the eduKateSG examination-performance series. Alicia looks for structure, dependency and where a full-paper failure began. Tricia asks whether the evidence is representative enough to support the readiness claim being made. Kai Kai keeps asking the question that punctures exam theatre: “What exactly did this mock prove?”

This page has a deliberately narrow canonical role. How Mock Exams Work | Simulate Before It Counts owns the positive architecture. How Exam Preparation Fails owns the wider readiness problem. How Examination Performance Fails owns the broad live-performance system. How Past Papers Fail owns past-paper misuse. This article owns one edge: how simulation produces false readiness, false unreadiness, or no useful learning because fidelity, diagnosis, repair, retesting and stopping are poorly designed.

It also connects directly to How Feedback Fails, How Error Analysis Fails, How Checking Fails, How Confidence Fails, and How Practice Fails.

The central rule: a mock is a measurement instrument before it is a training activity

A mock examination can train. It can also measure. Those functions overlap, but they are not identical.

If the purpose is measurement, conditions matter because the learner is making an inference from performance to readiness.

If the purpose is training, conditions can be deliberately altered to improve a particular capability.

Problems begin when the mode is not declared.

A learner completes a paper with notes open, pauses for dinner, asks two questions, resumes later and scores eighty-two per cent.

That can still be useful training.

It is not strong evidence of independent timed readiness.

Another learner sits an unusually difficult paper under strict conditions and scores sixty-five per cent.

That can still be useful stretch training.

It may not be a fair estimate of expected performance on the target examination.

Before every mock, ask:

• What are we trying to measure?
• Which conditions must therefore be realistic?
• Which deviations from the real event are intentional?
• How will the result be interpreted?
• What decision can this mock change?

Without these questions, simulation drifts from measurement into ritual.

The minimum useful mock-exam loop

A full simulation should enter a larger control loop:

1. Define the readiness question. Timing? Endurance? Full-paper integration? Interface? Recovery? Score stability?
2. Set the simulation conditions. Time, resources, paper, environment, support, interface and sequence.
3. Perform independently enough for the question.
4. Capture evidence. Score, timing, omissions, confidence, major errors, support used, state and sequence where relevant.
5. Diagnose. Find the first weak links rather than merely counting wrong answers.
6. Prioritise. Choose the highest-value repair targets.
7. Repair. Use targeted training rather than immediately doing another full paper.
8. Retest locally. Confirm that the mechanism changed.
9. Reintegrate. Return to another representative simulation.
10. Update readiness. Raise, lower or leave confidence unchanged according to evidence.
11. Stop or taper when enough evidence exists.

The mock is one node in the loop.

It is not the whole loop.

Failure Mode 1: the mock becomes a score event

The learner sits a full paper, receives a mark, and the mark becomes the entire meaning of the simulation.

Seventy-two per cent.

Good or bad?

That question is too compressed.

The same score can come from very different systems. One student loses marks because three topics are missing. Another knows the content but runs out of time. Another makes one catastrophic interpretation error. Another has strong performance except for careless answer transfer. Another has volatile confidence and changes correct answers during checking.

The score is an outcome. The training system needs mechanisms.

After the paper, decompose enough of the mark loss to answer: what should change before the next simulation?

If the answer is only “get a higher score,” the mock has not yet done its diagnostic job.

Failure Mode 2: the conditions are easier than the real examination

The student pauses the clock, checks a phone, asks a quick question, gets a snack, looks at a formula and continues.

None of these actions may feel serious individually.

Together they change the task.

A comfortable simulation can overestimate readiness because it removes the constraints that make knowledge fragile: sustained attention, retrieval without rescue, pacing, uncertainty and fatigue.

That does not mean all practice must be strict.

Learning mode can include support.

But readiness claims require condition honesty.

If the learner is allowed support, label the performance as supported. When the question becomes “Can this survive the actual examination?”, remove the support that will not exist there.

Failure Mode 3: the conditions are harder than the real examination

Some adults believe a mock should be brutal.

Harder questions. Less time. No breaks. Extra pressure. Harsh marking.

The logic is that surviving something worse will make the real event easy.

This can train robustness in selected doses.

It is weak readiness measurement.

If the student fails a deliberately distorted simulation, what has been learned about the target examination?

Perhaps very little.

Use stretch tasks when you want stretch. Use representative tasks when you want readiness evidence.

Do not call both outcomes by the same name.

Failure Mode 4: the mock starts before prerequisite knowledge is stable

A full paper is an expensive way to discover that half the syllabus has not been learned.

If large foundations are missing, the mock repeatedly reports the obvious.

Wrong.

Blank.

Wrong.

Blank.

There is little integration to diagnose because the components do not yet exist.

Use targeted teaching and partial-paper work first.

Introduce full simulation when enough underlying capability exists for interaction effects to become informative.

A mock is most valuable when the question is not “Have you learned anything?” but “Can what you learned survive the whole event?”

Failure Mode 5: the first realistic mock happens too late

The learner completes revision and finally sits a serious simulation five days before the examination.

The paper reveals a pacing collapse, weak answer transfer, poor checking and fatigue in the final section.

Excellent diagnosis.

Almost no runway.

Simulation should arrive early enough that important system failures can still alter training.

The exact timing depends on the course and examination, but the principle is universal: a diagnostic event has more value when there is time to act on the diagnosis.

Failure Mode 6: every practice session becomes a mock

Full papers feel serious.

They produce a score. They look like the examination. They create visible effort.

So students can spend weeks doing one paper after another.

The same weaknesses are sampled repeatedly.

Sampling is not repair.

If the last three mocks show weak method selection in the same family of questions, the next training block should probably not be another whole paper.

It should be targeted work on the bottleneck.

Simulation generates the training queue.

Targeted practice changes the queue.

Later simulation tests whether the change integrated.

Failure Mode 7: mocks are scheduled back-to-back without a reason

Three full papers in one day can look disciplined.

The second and third papers may mainly measure exhaustion.

If the real examination sequence includes back-to-back papers, the simulation can be valuable.

If it does not, the extra fatigue contaminates the measurement.

Use back-to-back simulation to answer a back-to-back question.

Do not add suffering as proof of seriousness.

Failure Mode 8: the wrong specification is used

An old syllabus, wrong tier, wrong board, wrong calculator policy or obsolete paper structure can make a simulation look authentic while training the wrong event.

Check the current authoritative specification before treating a paper as readiness evidence.

Format drift matters because timing, question mix, answer forms and permitted resources shape performance.

A mock is only as useful as its relationship to the event it claims to model.

Failure Mode 9: all mocks come from one source

Students adapt to authors.

They learn recurring layouts, wording patterns, topic emphases and distractor styles.

This can produce local fluency that feels like broad readiness.

Use enough source diversity to test transfer within the target specification.

The goal is not random novelty.

The goal is to ensure the capability belongs to the learner rather than to familiarity with one paper producer.

Failure Mode 10: the same paper is repeated until memory inflates the score

Retaking a paper can be useful.

It can verify that a correction was understood.

But rapid repeated papers carry answer memory, remembered traps and remembered methods.

The score becomes a mixture of learning and item familiarity.

Use the repeated paper as a repair check.

Use an unseen or sufficiently delayed paper when you need fresh readiness evidence.

Failure Mode 11: the mark scheme leaks into the attempt

The learner checks uncertain answers while still performing.

Independence disappears exactly where uncertainty was highest.

That part of the score is no longer clean readiness evidence.

Preserve the closed performance window first.

Then review deeply afterward.

Learning can be generous after the measurement phase. The measurement phase should remain honest enough to answer its question.

Failure Mode 12: the simulation mode is undefined

A paper is timed for the first hour, untimed for the second, open-book for difficult sections, paused for interruptions and lightly prompted by a tutor.

The final score cannot be interpreted cleanly.

Declare the mode.

Diagnostic paper with resources.

Timed section.

Full realistic simulation.

Supported teaching paper.

Different modes are useful.

The failure is pretending they measure the same state.

Failure Mode 13: timing is approximate

“It took about two hours.”

About is often where pacing failures hide.

If the real paper is one hundred and twenty minutes, a ten-minute overrun matters.

Use a real clock when timing is part of the claim.

Record only enough timing detail to locate meaningful bottlenecks.

The objective is not surveillance.

It is interpretable evidence.

Failure Mode 14: extra time is added invisibly

The official time ends.

The learner keeps going.

The eventual score is recorded as the mock score.

This hides the most important fact: some marks were not reachable under the target constraint.

A better review can record both states.

Official-time score and completion.

Post-time score if useful diagnostically.

The second number can show what knowledge exists. The first shows what performance was available within the event.

Failure Mode 15: the mock ends when the learner feels finished

The student completes all attempted questions with fifteen minutes left and stops.

The final checking phase is never practised.

The real examination may offer those same fifteen minutes.

Use them.

Scan for omissions, known error signatures, answer transfer, units, flags and low-confidence items according to the paper.

Readiness includes how the learner uses the final phase, not only how quickly the first pass ends.

Failure Mode 16: the environment is unrealistically perfect

Absolute silence. Favourite desk. Favourite pen. No movement. Ideal temperature. No waiting.

Real examinations are controlled but not frictionless.

The goal is not to manufacture distraction.

It is to avoid building a performance routine so context-dependent that ordinary examination-room variation feels like failure.

Occasionally vary harmless contextual details.

The routine should belong to the learner, not the room.

Failure Mode 17: the environment is deliberately chaotic

Noise is added. People interrupt. Time is shortened. Pressure comments are made.

This is sometimes called resilience training.

It can easily become invalid simulation.

Train specific resilience capacities separately if needed.

Do not corrupt the readiness instrument merely to prove toughness.

Failure Mode 18: the same seat, room and ritual become hidden cues

Context can support retrieval.

If every full paper occurs in the same environment, performance can become partly tied to that context.

Variation does not need to be dramatic.

Different desk. Different room. Different time. Same task standard.

Small context changes reveal whether the routine is portable.

Failure Mode 19: equipment constraints are ignored

The learner practises with a calculator model or reference sheet different from the real event.

Or uses tools that will not be permitted.

Or avoids permitted tools to make practice harder.

All three distort the task.

High-stakes simulation should use the resource environment the learner will actually face.

Failure Mode 20: answer-booklet mechanics are ignored

Practice papers are solved in notebooks, on loose paper or directly beside the question when the real assessment requires separate answer spaces, numbering, grids or transfer.

Correct reasoning can still lose marks at the interface.

If answer location matters, simulate it.

The reasoning path ends only when the answer is recorded where the assessment can recognise it.

Failure Mode 21: digital navigation is never rehearsed

A digital examination is practised entirely on paper.

The learner never experiences scrolling, timers, flags, answer fields, navigation, autosave or the visibility limits of the digital interface.

Digital operation is part of performance when the medium changes how tasks are found and completed.

Use a representative digital environment when available.

Failure Mode 22: endurance is inferred from short sets

The learner completes twenty-minute drills accurately and is declared ready for a two-hour paper.

Short accuracy does not certify full-duration stability.

Late-stage reading, working memory, checking and handwriting can all degrade.

Use full-duration simulation when endurance is one of the uncertainties that still matters.

Failure Mode 23: endurance is praised while content is missing

The learner finishes every page and receives praise for stamina.

Large conceptual gaps remain.

Endurance is one layer of readiness.

It cannot replace knowledge.

Interpret the full system.

Failure Mode 24: topic-specific warm-up creates false readiness

Immediately before the mock, the learner revises the exact formulas, vocabulary or concepts likely to appear.

Performance benefits from recent activation.

The real examination may begin cold.

Include cold-start simulations where that is representative.

A capability that exists only immediately after priming is not the same as independently available knowledge.

Failure Mode 25: the learner always practises at the best time of day

Every mock occurs on a quiet weekend morning when the student is fresh.

The real examination may happen after travel, waiting, another paper or a different circadian point.

Do not deliberately impair the learner.

But sample enough representative states to know whether readiness depends on ideal timing.

Failure Mode 26: one exhausted mock becomes the readiness verdict

A student sleeps badly, sits a paper, scores weakly and concludes that preparation has failed.

The result is real evidence.

It is evidence from a noisy state.

Record the state and avoid letting one atypical low-performance condition become the whole learner model.

Failure Mode 27: one strong mock becomes proof of readiness

A ninety appears.

Confidence surges.

One paper can align with strengths, recent revision or a favourable question mix.

Peak performance proves possibility.

Readiness needs repeatability.

Use enough representative evidence to show that strong performance is not a one-paper accident.

Failure Mode 28: one weak mock becomes proof of failure

A low score causes panic.

The whole plan is rewritten.

New resources are bought. Every topic is reopened. Confidence collapses.

Diagnose first.

Was the loss diffuse or concentrated? Was one early error responsible for several later losses? Was the paper unusually difficult? Did timing or state dominate?

A weak mock should change the plan only as far as the evidence justifies.

Failure Mode 29: average score hides volatility

A learner averages eighty per cent.

The individual papers are fifty-eight, ninety-four, seventy-one and ninety-seven.

Another learner averages seventy-nine with scores of seventy-seven, eighty, eighty-one and seventy-eight.

The averages are similar.

The reliability is not.

High-stakes readiness cares about spread as well as centre.

Ask what happens on an ordinary bad day, not only what the average says.

Failure Mode 30: peak score becomes the forecast

Students remember their best mark.

Parents remember it too.

The highest score is emotionally attractive because it shows what is possible.

Expected performance needs a more representative estimate.

Use recent ranges, conditions, and the stability of underlying mechanisms.

Failure Mode 31: the lowest score becomes identity

The worst mock often feels more truthful than the best because failure is emotionally loud.

Use the low paper diagnostically.

Do not automatically turn it into a global identity.

Ask which mechanisms produced it and whether they recur.

Failure Mode 32: paper difficulty context is ignored

Seventy on a familiar accessible paper is treated as identical to seventy on a demanding transfer-heavy paper.

Scores need task context.

This does not require complicated statistical modelling.

It requires enough judgement to avoid false equivalence.

Failure Mode 33: score improvement becomes the only goal

The learner expects every mock to beat the previous one.

Natural variation makes the training process emotionally unstable.

A lower score can occur alongside better timing, fewer recurring errors and stronger transfer if the paper is harder.

Track mechanisms.

Let the score remain important without becoming the only signal.

Failure Mode 34: a useful repair is abandoned because the total score falls

The learner introduces a better checking routine.

The next paper is harder and the total mark drops.

The checking routine is blamed.

Look at the target mechanism.

Did preventable errors decrease?

Did timing remain acceptable?

One aggregate outcome should not automatically overturn mechanism-level evidence.

Failure Mode 35: support-inflated scores are compared with independent scores

One paper is completed with hints. Another is completed cold.

The marks are plotted on the same line as though they represent the same state.

Label the mode.

Support level is part of the measurement.

Failure Mode 36: self-marking is generous

The learner awards benefit of doubt repeatedly.

Close answers become correct. Incomplete explanations receive full credit. Ambiguous working is assumed to deserve method marks.

Calibration drifts upward.

Use official criteria where available and periodically compare self-marking with expert marking.

Failure Mode 37: marking is deliberately harsh

Adults sometimes remove marks to keep a learner humble or motivated.

The score stops being a measurement instrument.

Use the real standard.

Motivation should not be manufactured by corrupting evidence.

Failure Mode 38: the learner marks from intuition instead of criteria

“This sounds right.”

That is not always the same as “This earns credit under this assessment.”

Practise the actual scoring interface.

Marks are not merely about truth. They are about demonstrated fulfilment of the task.

Failure Mode 39: corrections become mark-scheme copying

The student rewrites official language word for word.

The page looks perfect.

The underlying model may still be weak.

Ask the learner to close the mark scheme and reconstruct the answer.

Then vary the question.

Recognition is not independent generation.

Failure Mode 40: the paper is marked but not diagnosed

Every wrong answer receives a cross.

The student knows where marks were lost.

The learner still does not know what to train.

At least the major errors should be classified: knowledge, retrieval, selection, execution, expression, control, checking or state.

Diagnosis turns the paper into a plan.

Failure Mode 41: diagnosis produces no repair

A detailed post-mortem is completed.

Everything is understood.

No practice changes.

Insight is not performance change.

Every important diagnosis needs a next action.

Failure Mode 42: the mock generates too many repair targets

A paper loses thirty marks in fifteen different ways.

All fifteen become homework.

The learning system fragments.

Prioritise.

Which failures recur? Which are upstream? Which propagate? Which cost many marks? Which can realistically improve now?

The paper is a field of evidence, not a command to repair everything at once.

Failure Mode 43: the mock generates too few repair targets

The review ends with “revise more” and “be careful.”

The simulation’s diagnostic resolution is discarded.

Extract enough specificity to alter the next training block.

Failure Mode 44: every wrong answer is corrected immediately

Immediate correction is useful for understanding.

It can create a false sense of repair because the right route is still active in memory.

Use immediate correction as stage one.

Then reattempt later without support.

Failure Mode 45: failed questions are never reattempted

The learner reads the explanation and moves on.

No evidence shows that the corrected process can now be generated.

Close the solution.

Try again.

Failure Mode 46: only identical questions are reattempted

Memory of the original item can carry success.

Use a near variant next.

Then mix it with neighbours.

The target is the rule, not the historical page.

Failure Mode 47: targeted repair is never reintegrated into a full mock

The weak component improves in drills.

Does it survive full-paper timing, switching and fatigue?

That remains unknown.

Return to integration after local repair.

Failure Mode 48: another full mock is taken before the last mock is repaired

The same bottleneck appears again.

The student calls this consistency.

It is repeated measurement.

When the diagnosis is already clear, repair before resampling.

Failure Mode 49: mocks become punishment

A weak result triggers another full paper.

The message becomes: failure earns more labour.

Simulation should answer a readiness question.

Do not use it as discipline.

Failure Mode 50: mocks become reassurance

After a disappointing paper, an easy familiar paper is selected to restore confidence.

The score rises.

Calibration may worsen.

Confidence should recover from representative evidence, not protected measurement.

Failure Mode 51: mocks are used to scare the learner

A deliberately brutal paper is used to create urgency.

Fear increases.

The plan may not improve.

Use evidence to motivate change, not theatrical difficulty.

Failure Mode 52: mock success depends on tutor presence

The tutor sits nearby, occasionally redirects attention, answers procedural questions and gives reassurance.

The paper looks independent because no answers are supplied.

The learner is still receiving control support.

Independent readiness requires enough simulation without that external controller.

Failure Mode 53: all support is removed too early

A learner in active repair is thrown into a full strict mock to “see what happens.”

Many systems fail at once.

Diagnosis becomes noisy and confidence collapses.

Fade support progressively.

Independence is a destination, not a stunt.

Failure Mode 54: the learner changes strategy after every mock

Question order changes.

Timing changes.

Checking changes.

Planning changes.

The next mock contains a different strategy stack.

No approach remains stable long enough to evaluate.

Change when evidence points to a specific weakness.

Do not chase comfort after each paper.

Failure Mode 55: the learner refuses to change a failing strategy

“This is how I always do it.”

Repeated evidence shows the routine creates lost marks.

Consistency becomes stubbornness.

Preserve what works. Update what fails.

Failure Mode 56: the mock has no declared purpose

The learner sits a paper because the timetable says “Mock.”

Afterward, every interpretation becomes possible.

State the main readiness question first.

Are we testing timing? endurance? full-paper transfer? digital interface? recovery? score stability?

One paper can generate many observations, but a declared purpose protects interpretation.

Failure Mode 57: the mock tries to measure everything at once

Full-paper performance is a complex system.

A poor result can have interacting causes.

Use the mock to reveal the system.

Then use smaller tests to isolate causes.

Do not expect one full paper to tell you with certainty which hidden mechanism failed.

Failure Mode 58: a targeted test is called a readiness mock

One section performs strongly.

The learner generalises to the whole examination.

Keep claims proportional to what was tested.

Failure Mode 59: a full mock is treated as a pure knowledge test

Every wrong answer becomes a content-revision task.

Technique and execution disappear from diagnosis.

Follow up selectively.

Can the learner answer correctly untimed? with a prompt? after the paper? If yes, the weakness may sit elsewhere.

Failure Mode 60: a full mock is treated as a pure technique test

Weak knowledge is blamed on timing, nerves or strategy.

Technique becomes a convenient explanation.

Sometimes the learner simply does not know enough.

Use follow-up evidence to separate missing knowledge from poor performance control.

Failure Mode 61: question-order strategy becomes dogma

“Always do easy questions first.”

“Always start with the hardest.”

“Always follow the paper.”

No universal order fits every assessment.

Evaluate navigation cost, paper structure, learner strengths, compulsory sections and return reliability.

A strategy is good because it works under the actual constraints, not because it sounds clever.

Failure Mode 62: the learner practises an order the real examination does not permit

Some assessments constrain navigation or section order.

A practice strategy that cannot be executed legally has no transfer value.

Train inside the rules.

Failure Mode 63: move-on decisions are never rehearsed

Untimed practice teaches persistence without opportunity cost.

The real paper requires containment.

Use checkpoints and move-on triggers for questions that can consume the paper.

Failure Mode 64: the learner moves on but never returns

A fallback without a return path is abandonment.

Flag clearly, preserve partial working and define when unresolved items re-enter the queue.

Failure Mode 65: return order is random

Every skipped question receives equal late attention.

A low-probability rescue absorbs time while cheap marks remain elsewhere.

Triage by recoverability, mark value and verification cost.

Failure Mode 66: final checking is never practised

The learner finishes the last question as time expires in every mock.

No final quality-control routine exists.

Either pacing must change or local checking must become stronger.

Do not simply tell the student to “check at the end” when no time budget exists.

Failure Mode 67: checking consumes too much of the mock

The learner repeatedly re-solves already strong questions.

Completion suffers.

Verification has an opportunity cost.

Use targeted checks and stopping rules.

Failure Mode 68: review creates harmful answer changes

Correct answers become uncertain after repeated inspection.

The student changes without new evidence.

Track answer-change outcomes in practice.

Use a simple rule: uncertainty triggers checking; evidence triggers change.

Failure Mode 69: the first answer is treated as sacred

The opposite superstition also fails.

If checking reveals a genuine error, change it.

The policy is evidence, not loyalty to chronology.

Failure Mode 70: confidence is never recorded when it would help diagnosis

A correct guess looks identical to a correct known answer.

A high-confidence misconception looks like an ordinary mistake.

Sample confidence selectively during training to improve calibration and diagnosis.

Failure Mode 71: confidence is recorded on every item forever

Metacognitive measurement becomes another cognitive task.

Use it strategically.

Fade it when the internal gauge improves.

Failure Mode 72: anxiety is treated as unreadiness

The learner feels nervous during a mock and concludes that the preparation is failing.

Nerves are a state.

Performance evidence is separate.

A student can feel anxious and still execute strong procedures.

Train action under ordinary uncertainty instead of demanding perfect calm.

Failure Mode 73: calmness is treated as readiness

A relaxed student may be underprepared.

Use performance evidence.

Emotion is informative but not a substitute for measurement.

Failure Mode 74: pressure is simulated through threats

Adults add criticism, punishment or humiliation to make the mock “real.”

The interpersonal threat is not the examination.

Authentic demand already exists in time, independence, uncertainty and consequence.

Do not manufacture a different stressor and call it fidelity.

Failure Mode 75: every question feels familiar

The learner has practised the same shapes repeatedly.

The mock produces high confidence.

Unseen wording later reveals brittle transfer.

Use representative variation within the real problem space.

Failure Mode 76: novelty is exaggerated beyond relevance

Extreme questions are inserted to prove robustness.

The mock no longer resembles the target distribution.

Novelty should test transferable structure, not create arbitrary difficulty.

Failure Mode 77: cold starts are never tested

Every mock begins after subject-specific revision.

The learner never practises retrieving core knowledge from a cold state.

Include cold starts where they represent the real event.

Failure Mode 78: the learner over-warms up before every mock

A long revision session immediately precedes simulation.

The learner begins mentally tired and unusually primed.

Practise a realistic pre-paper routine.

Failure Mode 79: hidden feedback enters between sections

The learner checks answers during a break that would not exist in the real event.

Later sections benefit from information unavailable under target conditions.

Preserve authentic feedback boundaries when integrated readiness is the question.

Failure Mode 80: switching between papers is never rehearsed

One subject ends. Another begins.

The learner carries frustration, triumph or unresolved thinking into the next paper.

If the real schedule contains close sequencing, train a compact reset routine.

Failure Mode 81: unnecessary back-to-back simulation is added

When the actual schedule does not require it, serial papers may only create artificial fatigue.

Fidelity means reproducing relevant constraints, not maximising hardship.

Failure Mode 82: sleep context is ignored

Some mocks follow good sleep. Others follow late-night study.

The scores are compared without context.

Where state plausibly matters, record it lightly enough to aid interpretation.

Failure Mode 83: every low-sleep day is cancelled

The learner only simulates under ideal conditions.

Readiness may be fragile to ordinary imperfection.

Do not deliberately sleep deprive.

But if an ordinary suboptimal day occurs, the result can provide robustness evidence when interpreted carefully.

Failure Mode 84: food, hydration and break routines do not match the event

Minor physical routines can become major distractions when unfamiliar.

Where rules permit specific hydration, breaks or reporting procedures, rehearse them at least enough that they are not novel on the day.

Failure Mode 85: reporting and setup are ignored

Simulation starts exactly when the paper begins.

Real high-stakes events may involve travel, waiting, seating, equipment checks and pre-paper arousal.

Occasionally rehearse the broader event when those transitions are likely to matter.

Failure Mode 86: the mock ends at submission and ignores recovery

One exam finishes. The learner immediately analyses mistakes, argues about answers, or scrolls messages.

If another paper follows soon, this can carry cognitive residue.

Train post-paper disengagement when the schedule demands it.

Failure Mode 87: post-mock review becomes rumination

The learner revisits the same lost marks repeatedly.

No new decision emerges.

Analysis has ended; rumination continues.

Extract the operational lessons and close the review.

Failure Mode 88: review happens while emotion is too high

An angry or frightened learner may not process feedback accurately.

Preserve the script and evidence.

Use enough decompression to restore diagnostic quality.

Failure Mode 89: review is delayed until process memory disappears

Weeks later, the learner remembers the wrong answer but not why it seemed sensible.

Review within a useful window or preserve working, flags, timing and confidence traces.

Failure Mode 90: total time is recorded but bottlenecks remain invisible

“Finished in two hours.”

Where did the time go?

One extended response may have consumed twenty unnecessary minutes.

Sample timing at meaningful boundaries during training.

Failure Mode 91: timing data becomes obsessive

Every minute is logged.

The learner thinks about measurement more than the paper.

Use the minimum timing resolution needed to reveal the bottleneck.

Failure Mode 92: minutes-per-mark becomes a rigid law

Heuristics help allocate time.

They are not physical laws.

Some questions require setup. Some marks are faster. Some papers have uneven structures.

Use pacing rules as controls, not shackles.

Failure Mode 93: mark value is ignored

Ten minutes disappear into one low-value item.

Opportunity cost grows.

Mark value should influence persistence, especially when time is scarce.

Failure Mode 94: partial-credit strategy is absent

A difficult question is abandoned completely.

The assessment would have rewarded method, working, evidence or partial reasoning.

Train how to preserve recoverable marks where the marking system allows.

Failure Mode 95: practice working is longer than exam working

Teaching-mode solutions are verbose.

Mock answers copy that style.

Timing becomes invalid.

Separate explanatory learning work from examination sufficiency.

Failure Mode 96: mock working is too compressed

The learner removes useful steps to save time.

Traceability disappears and partial credit is threatened.

Train efficient sufficiency, not minimalism.

Failure Mode 97: templates survive only familiar prompts

A writing or problem-solving template performs well on known forms.

One changed prompt causes collapse.

Test whether the learner understands the function under the template.

Failure Mode 98: one awkward question causes useful structure to be abandoned

A template does not fit perfectly.

The learner discards all planning or structure.

Adapt rather than reject when the underlying function still helps.

Failure Mode 99: mocks are used to predict exact grades

Two papers become a precise forecast.

Measurement uncertainty disappears from the conversation.

Use score ranges, readiness states and mechanism stability rather than false precision.

Failure Mode 100: predictions never update

The student expects the same grade despite repeated contradictory evidence.

Confidence becomes detached from data.

Use mock outcomes to update forecasts proportionally.

Failure Mode 101: predictions are made only after seeing the result

Hindsight feels accurate.

When useful, predict a range before marking or before starting.

Then compare.

This teaches calibration.

Failure Mode 102: self-marking is never calibrated

The learner becomes increasingly confident in scores that do not match external marking.

Periodically compare with teacher or authoritative standards.

Self-marking is a skill.

Failure Mode 103: teacher marking is never calibrated

Different tasks receive different standards.

The learner sees noise as progress or decline.

Use rubrics, moderation and consistent examples where relevant.

Failure Mode 104: ambiguous items are automatically counted as learner failure

Practice materials can be flawed.

Verify surprising items before constructing a major intervention around them.

Measurement instruments need quality control too.

Failure Mode 105: ambiguity is blamed whenever an item is uncomfortable

Unfamiliarity is not invalidity.

Require a specific reason and compare with authoritative guidance.

Failure Mode 106: unofficial papers are treated as equivalent to official style

Third-party materials vary in quality, difficulty and alignment.

Use them knowingly.

Calibrate against authoritative samples.

Failure Mode 107: official past papers become predictions of future content

Students infer that frequently tested topics will recur.

Simulation turns into question spotting.

Use past papers to learn task demands, not to guarantee future sampling.

Failure Mode 108: question spotting narrows preparation

Predicted topics receive heavy attention.

Unseen sampling punishes the neglected regions.

Keep the full target domain visible.

Failure Mode 109: strong topics dominate the mock set

Papers are unconsciously selected to reassure.

Weakness remains unmeasured.

Representative simulation needs representative coverage.

Failure Mode 110: weak topics dominate every mock

Every paper is selected to expose known weaknesses.

Overall readiness is underestimated and stable strengths receive no maintenance.

Use targeted weak-topic work separately.

Failure Mode 111: no balanced paper exists in the training set

The learner alternates between easy confidence papers and extreme challenge papers.

Calibration lacks a realistic middle.

Include representative tasks.

Failure Mode 112: no repair capacity exists after the mock

The calendar schedules a full paper every Saturday and fills every weekday with fixed content.

The mock reveals problems but there is nowhere to insert repair.

Reserve adjustment capacity.

A diagnostic system needs space to respond.

Failure Mode 113: the mock is scheduled because the calendar says so

The learner is in the middle of rebuilding a major prerequisite.

The timetable demands a mock anyway.

Evidence should influence schedule.

Calendars structure learning; they should not override obvious diagnostic value.

Failure Mode 114: too many mocks cluster near the examination

Full-paper volume increases while recovery, sleep and targeted repair shrink.

More simulation can reduce performance when the readiness question is already answered.

Taper when evidence is sufficient.

Failure Mode 115: one good mock stops simulation too early

Readiness is declared after a single success.

Stability remains unknown.

Use enough evidence proportional to stakes.

Failure Mode 116: mocks continue after readiness is stable

More measurement adds little information.

The learner could instead preserve sleep, confidence and maintenance.

Stop when the readiness question is sufficiently answered.

Failure Mode 117: full mocks replace ordinary study too early in the year

The learner repeatedly samples material that has not yet been built.

Low scores provide obvious information.

Foundations first. Integration later.

Failure Mode 118: mocks are avoided because low scores feel discouraging

The learner postpones realistic evidence until confidence feels safer.

False readiness survives.

Introduce simulation progressively and frame it as measurement rather than judgement.

Failure Mode 119: mocks are postponed until “revision is finished”

Revision completion becomes an endless prerequisite.

No integrated performance evidence appears.

Use simulation to reveal what revision still needs to accomplish.

Failure Mode 120: the whole syllabus is revised after every mock

One paper reopens everything.

Targeting disappears.

Repair the failures the mock actually exposed.

Failure Mode 121: nothing changes after the mock

The paper is another score in a sequence.

No learning loop closes.

At minimum, major failures should alter the next training block.

Failure Mode 122: mocks are used to decide ability rather than readiness

A temporary state becomes a statement about intelligence.

“I am bad at mathematics.”

“I am not an exam person.”

Keep the inference local.

What is reliable now? What fails? What can change?

Failure Mode 123: readiness is interpreted without considering remaining time

The same score means different things two months and two days before the examination.

Repair options shrink with time horizon.

Interpret evidence relative to the calendar.

Failure Mode 124: readiness is interpreted without a target standard

“That was a good mark.”

Good for what objective?

Readiness requires a criterion.

Use the actual goal and the margin needed for reliability.

Failure Mode 125: no score margin is considered

The learner repeatedly reaches exactly the desired threshold.

Ordinary variation could push performance below it.

Robust readiness needs some headroom where stakes justify it.

Failure Mode 126: headroom training destabilises the core

Advanced challenge consumes all preparation.

Routine marks become rusty.

Secure the core before using stretch work to build reserve.

Failure Mode 127: the mock becomes a performance for parents or tutors

The student wants to prove competence.

Weakness is hidden, hints are welcomed, and difficult questions may be avoided.

Diagnostic honesty falls.

Create enough psychological safety that a mock can reveal failure without becoming humiliation.

Failure Mode 128: the learner performs for their own self-image

Easy papers are chosen. Conditions are softened. Disputed marks are awarded generously.

Measurement protects identity.

Predeclare conditions and use representative tasks.

Failure Mode 129: each mock is isolated from history

One paper is reviewed without comparison to prior patterns.

Recurrence and improvement disappear.

Track a small set of longitudinal signals: score range, completion, major error families, timing, support and known risks.

Failure Mode 130: the history becomes a giant dashboard

Every metric is tracked.

No decision changes.

Data collection becomes theatre.

Keep only measures that influence action.

Failure Mode 131: the mock produces data but no action thresholds

The learner knows the score, timing, error rate and confidence.

What happens next?

Define what evidence triggers targeted repair, another mock, maintenance, taper or rest.

Failure Mode 132: action thresholds are rigid

One point below target triggers a total plan change.

Natural noise creates instability.

Use ranges and mechanism evidence.

Failure Mode 133: readiness is stable but new techniques keep being added

Last-minute novelty enters a working system.

Reliability can fall.

Freeze what works near the event unless strong evidence demands change.

Failure Mode 134: a weak mock causes everything to change at once

Question order, timing, checking, study methods and sleep routine all change.

The next result cannot reveal which intervention mattered.

Prioritise and change the smallest set that addresses the main bottleneck.

Failure Mode 135: simulation fidelity becomes exact replication theatre

The desk is arranged perfectly. The same pen is used. A fake invigilator reads instructions.

Meanwhile timing, answer transfer and realistic question mix are poorly matched.

Copy the constraints that materially affect performance.

Do not fetishise superficial resemblance.

Failure Mode 136: fidelity is ignored because “knowledge is knowledge”

No timing. No interface. No endurance. No independent performance.

Transfer is simply assumed.

If the target is high-stakes performance, rehearse the constraints that change output.

Failure Mode 137: debrief remains question by question

Every item is reviewed locally.

System effects remain invisible.

Did fatigue rise late? Did one hard question cause a pacing cascade? Did checking disappear? Did confidence collapse?

Add a whole-paper debrief.

Failure Mode 138: debrief remains only at system level

“Timing was bad.”

“Confidence fell.”

“Need more endurance.”

Which questions reveal the mechanism?

Use both macro and micro views.

Failure Mode 139: knowledge failure and retrieval failure are confused

A blank answer automatically triggers reteaching.

After the mock, the learner answers correctly with a cue.

The knowledge may exist but be inaccessible.

Retest in a simpler condition before choosing the repair.

Failure Mode 140: retrieval failure and selection failure are confused

The student knows several methods when prompted but chooses the wrong one independently.

More memorisation will not necessarily solve the decision problem.

Use mixed classification and trigger contrast.

Failure Mode 141: selection failure and execution failure are confused

The right method was chosen but carried out badly.

Or the wrong method was executed perfectly.

These need different repairs.

Find the first divergence.

Failure Mode 142: execution failure and checking failure are confused

A sign error occurs.

The learner knows the method.

Could a cheap realistic check have caught it?

If yes, the quality-control layer matters too.

Failure Mode 143: checking failure and time-allocation failure are confused

The student never checks because the paper is unfinished.

Adding a bigger checklist cannot fit into the current system.

Repair pacing first.

Failure Mode 144: time-allocation failure and knowledge latency are confused

The learner is slow because retrieval and method selection take too long.

“Manage time better” is too shallow.

Measure where latency lives.

Failure Mode 145: fatigue and weak content are confused

Late errors trigger extra topic revision.

The same questions are correct when fresh.

Test state dependence.

Failure Mode 146: weak content is excused as fatigue

Every late error is blamed on tiredness.

A real concept gap survives.

Retest fresh.

Failure Mode 147: support creates false readiness

Hints, prompts, reassurance and procedural rescue occur during the paper.

The learner’s score contains external capability.

Record support explicitly and remove it before certifying independence.

Failure Mode 148: unnecessary isolation creates false unreadiness

An open-book or tool-permitted assessment is practised under harsher closed conditions.

The learner is measured against the wrong task.

Use the actual resource environment.

Failure Mode 149: unauthorised tools enter the mock

Spellcheck, calculator functions, notes or AI assistance exceed the real rules.

Performance cannot transfer legally.

Use only permitted support in readiness simulations.

Failure Mode 150: permitted tools are avoided to prove toughness

The learner refuses the calculator or formula sheet that the real assessment permits.

The mock measures an irrelevant harder task.

Train the event you will actually perform.

Failure Mode 151: contingency plans are never tested

The learner has never rehearsed what to do after blanking, a hard opening question, a time slip or a failed method.

Exam-day disruption feels catastrophic.

Build compact fallbacks.

You do not need to stage disasters constantly.

Failure Mode 152: contingency training dominates every mock

Adults inject problems into every simulation.

The learner practises crisis more than normal examination performance.

Test robustness occasionally.

Keep most mocks representative.

Failure Mode 153: a good mock is treated as a guarantee

The unseen examination still contains uncertainty.

Strong evidence should increase confidence, not eliminate uncertainty.

Readiness is probabilistic.

Failure Mode 154: a bad mock is treated as destiny

A poor simulation can be repaired.

Change the system, then gather new evidence.

The mock is information, not prophecy.

Failure Mode 155: the mock is emotionally treated as the real examination

Practice becomes irreversible in the learner’s mind.

Failure feels final.

Diagnostic honesty falls.

Keep enough seriousness for valid effort while preserving the purpose of practice: this is where failure is still allowed to teach.

Failure Mode 156: the mock is treated as meaningless because it “doesn’t count”

Effort drops.

The score becomes invalid evidence.

Create practice stakes around truthful effort and useful review rather than punishment.

Failure Mode 157: the learner cannot enter performance mode

Practice remains conversational, stoppable and teacher-supported.

The real transition to sustained independent execution feels unfamiliar.

Use a consistent start routine for full simulations.

Failure Mode 158: the learner cannot leave performance mode

The mock ends but rumination continues for hours.

Recovery and later learning suffer.

Use a deliberate debrief boundary.

Failure Mode 159: the simulation is developmentally inappropriate

Young learners are given marathon mock schedules that exceed the real assessment demands.

The simulation measures tolerance for artificial burden.

Match the design to the actual event and learner stage.

Failure Mode 160: approved accommodations are omitted

The learner rehearses without support that will legitimately exist in the real event.

The mock is misaligned.

Use the approved conditions.

Failure Mode 161: unconfirmed accommodations are added

The learner becomes dependent on extra support that may not be available.

Clarify actual permitted conditions before high-stakes simulation.

Failure Mode 162: the mock system is never revalidated against the official event

Local routines evolve.

The real format changes.

Simulation drift accumulates.

Periodically recheck timing, resources, paper structure and answer requirements against current authoritative information.

Failure Mode 163: validation happens once and then freezes

Examination systems can change.

Contemporary readiness requires contemporary conditions.

Recheck when preparing for the actual cycle.

Failure Mode 164: the learner overfits to mark schemes

Answers become formulaic.

Unseen wording breaks the script.

Use mark schemes to learn scoring functions, then vary the surface.

Failure Mode 165: the learner underuses mark schemes

The answer interface remains mysterious.

Known content fails to earn credit.

Use official criteria to understand sufficiency.

Failure Mode 166: examiner comments become decorative knowledge

The learner can recite common mistakes but still makes them.

Translate comments into drills, triggers, checking routines and retests.

Failure Mode 167: examiner reports are ignored because they feel generic

Population-level patterns can still identify common traps.

Use them selectively where they match the target assessment and the learner’s evidence.

Failure Mode 168: examiner reports are treated as personal diagnosis

Generic population commentary is applied blindly.

Verify it against the learner’s own scripts.

Failure Mode 169: mocks are never compared with authentic school performance

Practice and school tests may diverge.

A pattern visible in one source can challenge assumptions from another.

Compare while respecting differences in purpose and difficulty.

Failure Mode 170: school tests are treated as perfect final-exam simulations

Local assessments can emphasise different content or construction.

Use multiple evidence sources.

Failure Mode 171: only full-paper practice is used

Integrated performance is sampled repeatedly.

Weak components remain weak.

Decompose between simulations.

Failure Mode 172: only subskill practice is used

Every component looks good alone.

Interactions fail under the clock.

Reintegrate before the event.

Failure Mode 173: completion is undefined

A paper with skipped sections is casually called “done.”

Define completion under the real rules so attempts are comparable.

Failure Mode 174: readiness is undefined

A score target exists without timing, independence, completion, recovery or stability criteria.

One high mark can then certify too much.

Define readiness as a bundle of evidence appropriate to the stakes.

Failure Mode 175: readiness criteria demand perfection

Zero errors. Total calm. Exact score certainty.

Preparation never ends.

Use realistic thresholds and accept residual uncertainty.

Failure Mode 176: readiness criteria are too loose

One success qualifies.

Stability remains unknown.

Require enough evidence proportional to consequence.

Failure Mode 177: mock quantity becomes a badge of seriousness

Thirty papers sound more impressive than ten.

Volume says little about repair quality.

Judge the loop by what each simulation changes.

Failure Mode 178: fewer mocks are interpreted as laziness

Targeted repair can look less dramatic than full-paper effort.

Do the minimum simulations needed to answer important readiness questions.

Spend the rest of the time changing capability.

Failure Mode 179: a quiet room is mistaken for valid simulation

Orderliness is visible.

Fidelity is deeper.

Timing, interface, independence, question mix and answer demands may matter more.

Failure Mode 180: stress is mistaken for authenticity

Harsh atmosphere is added because real exams feel stressful.

The training manipulates interpersonal threat instead of exam constraints.

Use authentic demand rather than theatrical intimidation.

Failure Mode 181: instructions are never practised

The learner starts at question one instantly.

Global paper rules remain unexamined.

Include instruction reading and paper orientation where the real event requires it.

Failure Mode 182: instructions consume too much opening time

Over-caution becomes its own pacing problem.

Train a bounded orientation routine.

Failure Mode 183: page scanning is absent

The learner discovers paper structure reactively.

Unexpected sections disrupt pacing.

Use a brief allowed orientation when appropriate.

Failure Mode 184: the whole paper is over-planned before starting

Too much opening time is spent predicting every difficulty.

Execution is delayed.

Orientation should reduce uncertainty, not become procrastination.

Failure Mode 185: extended responses are never planned

The learner begins writing immediately.

Structure breaks halfway through.

Train bounded planning when the task benefits from it.

Failure Mode 186: extended responses are over-planned

Beautiful plans consume the clock.

Use a stopping rule for planning.

Failure Mode 187: recovery after a wrong start is never practised

One mistaken opening causes total abandonment.

Train local repair, partial salvage and re-entry.

Failure Mode 188: entire answers are rewritten after local errors

A small mistake consumes huge time.

Practise minimal sufficient correction.

Failure Mode 189: the learner never writes through uncertainty

Questions are either known or skipped.

Partial-reasoning skills remain weak.

Where the assessment rewards it, practise structured best effort.

Failure Mode 190: the learner persists through uncertainty without a time boundary

Persistence becomes stubbornness.

Pair partial reasoning with move-on rules.

Failure Mode 191: preventable and non-preventable errors are not separated

Every wrong answer becomes “careless.”

Missing knowledge is treated as moral failure.

Separate capability gaps from execution losses.

Failure Mode 192: everything is declared non-preventable

“I just made a mistake.”

Quality control never improves.

Ask whether a realistic procedure or check could have caught the failure.

Failure Mode 193: felt difficulty is never compared with actual performance

Some questions feel hard and score well.

Others feel easy and leak marks.

Compare occasionally to improve calibration.

Failure Mode 194: personal difficulty is confused with paper difficulty

“That paper was impossible.”

Perhaps it exposed a specific personal bottleneck.

Separate subjective experience from broader task evidence.

Failure Mode 195: difficulty is equated with unfairness

Demanding questions are rejected instead of analysed.

Distinguish hard from invalid.

Failure Mode 196: an easy paper is treated as proof of readiness

A strong score on low-demand material creates complacency.

Use a representative difficulty range.

Failure Mode 197: the mock system has no exit condition

Papers continue until the examination arrives.

Measurement never stops.

Define when enough evidence exists to taper.

Failure Mode 198: the mock system has no restart condition

Mocks are reduced and then never reintroduced despite new warning signs.

Define triggers for renewed representative simulation if readiness evidence materially changes.

Failure Mode 199: nobody knows what evidence would change the readiness judgement

Confidence becomes ideological.

State what pattern would raise or lower readiness.

A model that cannot update is not a useful model.

Failure Mode 200: parents interpret mock marks as final grades

Practice scores trigger praise, punishment and comparison.

The diagnostic function becomes emotionally expensive.

Frame mock results as evidence for the next training decision while still respecting the target standard.

Failure Mode 201: tutors protect students from poor mock scores

Hints, selective marking or softened criteria preserve confidence.

The signal is weakened.

Protect the learner emotionally without corrupting measurement.

Failure Mode 202: tutors dramatise poor mock scores

The result is used to create fear or urgency.

Trust and calibration suffer.

Keep diagnostic evidence honest.

Failure Mode 203: every mock becomes a referendum on self-worth

The learner cannot experiment or expose weakness safely.

Separate current performance from identity.

The mock is allowed to fail because the real examination has not happened yet.

Failure Mode 204: every mock is treated as disposable

No seriousness. No review. No repair.

The time cost is wasted.

Give each simulation a purpose and a return path.

Failure Mode 205: the system never measures whether the learner is getting better at learning from mocks

The same post-paper process repeats forever.

Meta-learning stalls.

Over time, diagnosis should become faster, repairs more targeted, self-marking more calibrated and recurrence lower.

The mock process itself should improve.

Failure Mode 206: growing independence is not recognised

The learner needs less feedback but continues receiving more.

External control persists.

Fade support as self-diagnosis improves.

Failure Mode 207: novices receive too little feedback because mocks are assumed to be independent

The student gets a score and is left alone to interpret it.

Weak models persist.

Independence should match capability.

Failure Mode 208: readiness has no margin

The learner scores exactly at the threshold repeatedly.

There is no reserve for ordinary variation.

Build reasonable headroom where stakes justify it.

Failure Mode 209: reserve building weakens the core

Advanced challenge dominates preparation.

Routine marks decay.

Maintain the baseline cheaply while extending headroom selectively.

Failure Mode 210: the mock reveals a problem too late to change anything useful

Late information can increase anxiety without creating intervention capacity.

Near the event, choose simulations only when the result can still influence a worthwhile action.

Failure Mode 211: the learner keeps measuring after the decision is already clear

Another mock is taken to reduce residual uncertainty.

No realistic decision would change.

Stop measuring and protect recovery.

Failure Mode 212: the learner stops measuring while important uncertainty remains

Revision feels complete.

Integrated performance is still unknown.

Use a representative simulation to close the highest-value uncertainty.

Failure Mode 213: the final mock is too close to the examination

One atypical result becomes emotionally dominant.

There is no time to gather counter-evidence or complete meaningful repair.

Place the last major simulation with enough space for interpretation and recovery.

Failure Mode 214: the final mock is too early

Weeks of learning occur afterward.

The evidence becomes stale.

Use a lighter later confirmation if important readiness uncertainty remains.

Failure Mode 215: the learner tries to peak on every mock

Sleep, taper, diet, motivation and ritual are manipulated repeatedly.

Training becomes unsustainable.

Use some mocks as ordinary measurements. Reserve full peaking strategy for the actual performance cycle.

Failure Mode 216: the learner underperforms deliberately to protect against disappointment

Effort is withheld.

The evidence becomes unusable.

Psychological safety should make honest effort tolerable.

Failure Mode 217: the learner overextends mocks to prove toughness

Breaks, food or recovery are denied beyond exam requirements.

The simulation measures unnecessary suffering.

Fidelity beats heroics.

Failure Mode 218: the mock becomes theatre

The room looks official. The papers are printed beautifully. A timer is projected. Everyone sits in silence.

No diagnosis follows.

No repair follows.

No retest follows.

Surface realism cannot rescue a broken learning loop.

A simulation earns its cost when it changes preparation or confirms readiness with credible evidence.

Failure Mode 219: the mock becomes ritual

“It is Saturday, therefore we do a mock.”

Purpose disappears.

Every simulation should answer a question.

Failure Mode 220: the mock becomes the curriculum

Preparation collapses into papers.

Underlying knowledge and skill development slow.

Mocks are system tests inside a broader learning architecture.

They are not the entire architecture.

Failure Mode 221: the mock is never allowed to certify success

Adults always find another weakness.

Confidence cannot update upward.

When evidence is stable, say so.

Readiness is allowed to improve.

Failure Mode 222: the mock certifies success too easily

One strong paper produces “You’re ready.”

Confidence outruns evidence.

Require enough stability for the stakes.

Failure Mode 223: the learner cannot explain what changed between early and late mocks

Progress is experienced only as a bigger number.

Name the mechanisms.

Retrieval became faster.

Method selection improved.

Checking became targeted.

Endurance stabilised.

Knowing what changed helps preserve it.

Failure Mode 224: the learner can explain improvement but cannot reproduce it

Metacognitive language outruns performance.

Talking about the right strategy is not the same as executing it.

Retest behaviour.

Failure Mode 225: readiness is stable but the plan keeps intensifying

Fear drives more papers, more hours and more novelty.

Fatigue and interference threaten the stable system.

When building is done, shift to preserving.

Failure Mode 226: unreadiness is clear but the plan remains unchanged

Evidence is collected and ignored.

The simulation becomes ceremonial.

A mock that cannot change the next move has little diagnostic value.

The mock-exam failure map

• Fidelity failure: the simulation does not preserve the constraints that materially affect performance.
• Mode failure: supported training, stretch work and readiness measurement are mixed together.
• Sampling failure: one paper, one source or one difficulty band receives too much weight.
• Timing failure: time is measured poorly, added invisibly, or interpreted at the wrong resolution.
• Interface failure: answer booklets, digital navigation, transfer, equipment and permitted resources are omitted.
• Endurance failure: short practice is mistaken for full-paper stability or fatigue is misread as content weakness.
• Calibration failure: peak scores, worst scores, emotions or support-inflated marks distort readiness estimates.
• Marking failure: self-marking, harshness, generosity or weak criteria corrupt the evidence.
• Diagnosis failure: wrong answers are counted without locating mechanisms.
• Repair failure: the mock generates insight but practice does not change.
• Retest failure: corrections are never independently re-performed or transferred.
• Integration failure: targeted repairs are never returned to full-paper conditions.
• Strategy failure: question order, persistence, checking and return rules are untested or constantly changed.
• State failure: anxiety, sleep, fatigue and transitions are ignored or over-weighted.
• Volume failure: too many mocks crowd out the targeted work they were supposed to inform.
• Stopping failure: the system has no rule for when enough simulation evidence exists.
• Identity failure: temporary performance becomes a theory of the learner.
• Theatre failure: surface realism substitutes for diagnosis, repair and retesting.

The Alicia test: where did the full-paper failure begin?

Alicia refuses to treat a mock as one indivisible object.

Suppose the learner loses twenty marks late in the paper.

Was the late section weak?

Perhaps.

Or perhaps an earlier hard question consumed twelve extra minutes.

Why did that happen?

Two methods competed.

Why could the learner not choose?

The relevant distinction had never been trained in mixed conditions.

Now the mock is no longer saying “weak at the final section.”

It is saying “method-selection ambiguity earlier in the paper created a pacing cascade that later looked like fatigue.”

That is a much better repair target.

Alicia’s rule is simple:

Trace the first consequential divergence before repairing the last visible symptom.

The Tricia test: how representative is this evidence?

Tricia asks about sample quality.

Correct specification?

Representative difficulty?

Unseen enough?

Independent enough?

Timed honestly?

Marked credibly?

Normal enough state?

One paper can still be useful even when several answers are “not perfectly.”

The key is proportional inference.

If the mock was heavily supported, it can tell us about supported capability.

If it was unusually difficult, it can tell us about stretch tolerance.

If it was a clean representative simulation, it can tell us more about likely examination performance.

Tricia prevents evidence from claiming more than the conditions allow.

The Kai Kai test: what changed because we ran this mock?

Kai Kai asks the question that decides whether the simulation was worth the time.

What changed?

Did we discover the first weak link?

Did the revision queue change?

Did we remove an unnecessary check?

Did we add a pacing checkpoint?

Did confidence rise because stability was demonstrated?

Did confidence fall because support dependence became visible?

Did we decide to stop full mocks and taper?

If nothing changed and no important uncertainty was resolved, the mock may have been mostly theatre.

A five-level simulation ladder

Not every practice event should jump directly to a full mock.

A useful progression is:

Level 1: component test. One skill or question family, often with support.

Level 2: timed mini-set. Several related items under a clock.

Level 3: mixed section. Method selection and switching become necessary.

Level 4: full paper. Timing, endurance, checking and integration are tested.

Level 5: event simulation. The broader sequence, interface, reporting routine or back-to-back conditions are rehearsed when relevant.

Move up when the lower level is stable enough that the next level will reveal something new.

Move down when a full simulation exposes a component that needs repair.

This ladder prevents the false choice between drills and mocks.

Both belong to the same architecture.

A simulation-fidelity checklist

Before a readiness mock, verify the conditions that materially affect performance:

• current paper format or credible representative format;
• correct official time limit;
• permitted calculator, formula sheet, notes or other resources;
• realistic answer booklet or digital interface where relevant;
• independent performance without unauthorised prompting;
• representative question mix and difficulty;
• normal enough physical state;
• appropriate sequence if several papers occur close together;
• realistic final checking and submission procedure.

Do not confuse this list with a demand for theatrical exactness.

Fidelity means preserving the constraints that change performance.

A mock evidence card

A useful post-mock record can stay compact.

• Score range and completion: what was achieved within official time?
• Timing: where did pace drift?
• Major error families: knowledge, selection, execution, expression, checking, state.
• Support: what external help entered, if any?
• Confidence: which high-confidence errors or low-confidence correct answers matter?
• First weak link: what upstream mechanism deserves repair?
• Next action: what changes before the next full simulation?

If the record grows into a page of bureaucracy, compress it.

The purpose is routing.

A mock-repair matrix

Different mock failures need different responses.

• Missing knowledge → reteach, retrieve, rebuild prerequisites.
• Slow retrieval → spaced cold recall and fluency practice.
• Method-selection errors → contrast, mixed classification, trigger training.
• Execution errors → guided practice, fading, local verification.
• Question-reading errors → command-and-constraint gate.
• Incomplete answers → answer-function modelling and sufficiency checks.
• Timing drift → identify the latency source, then train checkpoints or move-on rules.
• Checking failure → targeted verification, omission scan, evidence-before-change rule.
• Endurance failure → longer integrated sets, pacing and recovery.
• Confidence miscalibration → prediction versus outcome review.
• Interface failure → practise transfer, numbering, navigation or permitted tools.
• Recovery failure → fallback and return-path rehearsal.

The matrix prevents the universal prescription of “do another mock.”

A worked example: the strong score that is not readiness

Alicia scores eighty-eight per cent.

Everybody is pleased.

Tricia looks at the conditions.

The paper was familiar. Alicia had revised the two largest topics that morning. She paused the clock twice. One difficult question was discussed briefly with a tutor. The paper finished ten minutes beyond official time.

The eighty-eight is not worthless.

It demonstrates strong supported knowledge.

It does not yet certify independent timed readiness.

The next step is not panic or celebration.

The next step is a representative cold paper with no prompts and strict official timing.

That mock produces eighty-one within time.

Now the evidence is stronger because the claim and conditions match more closely.

A worked example: the weak score that contains good news

Tricia scores sixty-nine, down from seventy-five.

The new paper is more difficult.

The old timing collapse is gone. Every section is reached. The repeated unit errors disappear. One advanced question creates a large mark loss.

The total falls.

The system improves.

Feedback should preserve that distinction.

The learner still has work to do.

But abandoning the successful pacing and unit-control repairs because the headline score fell would be irrational.

A worked example: the student who does fifteen mocks and remains stuck

Kai Kai has completed fifteen full papers.

The score range barely moves.

Each review identifies weak inference questions, slow long responses and late checking failure.

Then the next paper begins.

The mocks are not failing because fifteen is too few.

They are failing because the middle of the loop is missing.

The new schedule changes:

1. Three days of targeted inference work.
2. Two timed long-response mini-sets.
3. One checking drill on omissions and evidence-before-change.
4. One delayed mixed retest.
5. Then another full paper.

The sixteenth mock is valuable because the learner is no longer merely resampling the same system.

A worked example: the learner who panics after one bad mock

Alicia has been stable around eighty.

One mock produces sixty-one.

The first reaction is global: everything is going wrong.

The script says otherwise.

One early misread sends a multi-part question down the wrong route. The resulting stall consumes fifteen minutes. The final section is rushed. Four late errors follow.

The score is poor.

The failure chain is concentrated.

The repair is not “revise the whole syllabus.”

It is question-reading control, a move-on rule and recovery after one bad section.

The next representative paper returns to the normal range.

One bad mock changed the plan without changing the person.

A worked example: the learner whose mocks are always excellent at home

Kai Kai performs brilliantly in home simulations.

School examinations are much weaker.

The gap persists.

The home environment contains invisible support: flexible starts, familiar room, silent conditions, easy access to stationery, topic-specific warm-up and a parent nearby.

None is individually dramatic.

Together they reduce friction.

The repair is not to make home hostile.

It is to remove the supports that matter for the readiness claim and occasionally simulate a more representative event.

The gap narrows.

A worked example: the learner whose mocks are always worse than school

Tricia’s home mocks are deliberately difficult.

Parents believe this will build toughness.

The papers use extra-hard sources, harsh marking and shortened time.

School results are consistently better.

The mocks are not predicting readiness.

They are stretch tests.

Once renamed, the confusion disappears.

Representative mocks are added for calibration.

Stretch work remains, but it no longer controls confidence.

A worked example: the final mock that should not have happened

Three days before the examination, the learner already has stable evidence.

Recent papers sit comfortably above the target. Timing is controlled. Major errors are quiet.

Another full mock is scheduled because “one more cannot hurt.”

The learner sleeps badly the night before, sits an unusually difficult paper and scores far below normal.

Confidence collapses.

No meaningful repair can be completed in the remaining time.

The simulation created information with almost no decision value and large emotional cost.

A stopping rule would have prevented it.

Mocks in mathematics

Mathematics mocks are especially useful for separating selection, execution and checking.

When an answer is wrong, locate the first invalid line.

Was the method inappropriate?

Were values copied incorrectly?

Did algebra fail?

Did calculator input fail?

Was the final answer plausible?

Could substitution or estimation have caught it?

Timing data is also powerful.

A correct ten-minute solution can still be a performance problem.

Full papers reveal whether routine questions remain efficient after difficult ones and whether checking survives fatigue.

The key is to avoid responding to every mathematical error with another full mathematics paper.

Decompose. Repair. Reintegrate.

Mocks in science

Science simulations can reveal whether knowledge survives unfamiliar context.

Look beyond keywords.

Did the learner interpret variables correctly?

Use data?

Distinguish observation from explanation?

State causal mechanisms?

Respect conditions?

Longer science papers also expose whether precise explanation deteriorates late.

If the learner knows the science when discussing the paper afterward but writes incomplete answers under time, the repair should target answer construction and performance control, not only content revision.

Mocks in English and writing

Writing mocks reveal interactions that ordinary drafting hides.

Planning consumes time. Ideas compete. Paragraphs drift. Language quality changes under speed. Proofreading may disappear.

Review at several levels:

• Did the response answer the task?
• Was the structure controlled?
• Was evidence or detail sufficiently developed?
• Did sentence quality deteriorate late?
• Did planning help or overrun?
• Did final checking protect high-value errors?

A weak mock composition does not automatically mean “write more compositions.”

The first weak link may be planning, relevance, development, sentence control or time allocation.

Mocks in comprehension

Comprehension simulation tests more than reading ability.

It tests evidence location, inference, answer scope, pacing and the ability to move between text and questions efficiently.

A full-paper review should ask where the chain failed:

text → question interpretation → evidence selection → inference → answer form.

Then repair that transition.

Repeated full passages are not always the fastest route.

Mocks in multiple-choice examinations

Multiple-choice mocks offer rich calibration data.

Track selected high-confidence wrong answers.

Inspect low-confidence correct answers.

Look for negative wording, distractor families and answer-changing behaviour.

If the scoring system includes penalties, simulation must respect them.

The aim is not merely the number correct.

It is decision quality under the real rules.

Mocks in open-book examinations

Open-book does not mean unprepared.

Simulation should test retrieval, navigation, source selection, search cost and integration.

A learner who can find everything eventually may still be unready for a timed open-resource assessment.

Use the actual resource limits.

Measure whether searching supports thinking or replaces it.

Mocks in digital examinations

Digital simulations must include enough interface realism to expose operational risk.

Flags.

Scrolling.

Answer fields.

Timers.

Navigation.

Submission.

Practise with representative systems where possible.

A technically knowledgeable learner can still lose marks through interface error.

Mocks across multiple papers

Some examination periods are portfolios of events rather than one paper.

Readiness includes switching, recovery and allocation across days.

Where schedules are known and relevant, simulate selected sequences.

But do not turn every practice week into a marathon.

Sequence simulation should answer sequence questions.

Mock examinations and parents

Parents often want certainty from mocks.

“What grade will my child get?”

A useful answer is usually a range plus conditions.

Recent representative evidence suggests this level, with these active risks.

Parents can help by treating the mock as training evidence rather than a moral judgement.

Ask:

What did we learn?

What is the next repair?

When will it be retested?

Is another full paper actually needed now?

A calm diagnostic culture improves the quality of the signal.

Mock examinations and teachers

Teachers need to protect both validity and educational value.

A mock can certify, diagnose and train, but those functions should not be confused.

Where marks matter, use defensible standards.

Where learning matters, provide enough mechanism-level feedback to change the next attempt.

Where the learner is already stable, reduce unnecessary commentary and let success count.

Where the learner is not ready, route the smallest high-value repair rather than assigning another full paper automatically.

Mock examinations and tutors

Tutors have a particular risk: invisible assistance.

A hint, facial cue, reassurance or timing reminder can quietly improve the paper.

That support may be pedagogically useful.

It should not be mistaken for learner independence.

A strong tutoring system therefore has distinct modes:

teaching;

guided practice;

independent attempt;

mock simulation;

post-mock diagnosis.

The tutor’s job during a true mock is often to become deliberately less useful until the paper ends.

Mock examinations and confidence

Mocks can calibrate confidence better than almost any other practice tool because they place many skills under one realistic constraint set.

They can also distort confidence dramatically.

Easy paper, high confidence.

Brutal paper, low confidence.

Support-inflated paper, false confidence.

One bad-state paper, false pessimism.

Confidence should follow representative repeated evidence, not the emotional loudness of the most recent mock.

See How Confidence Fails.

Mock examinations and feedback

A mock is a high-resolution feedback event.

That value is lost if the learner receives only a score or a page of comments that never changes practice.

Feedback should identify a small number of mechanisms, translate them into action and create a return date.

See How Feedback Fails.

Mock examinations and error analysis

The full paper reveals interactions.

Error analysis decides whether the visible failure is local or systemic.

A blank final page might be a pacing failure, method-selection delay, fatigue, overchecking or weak content.

Do not repair the blankness.

Repair the cause.

See How Error Analysis Fails.

Mock examinations and checking

Mocks are where checking becomes measurable.

Does the learner have time for it?

Does it catch known errors?

Does it create harmful answer changes?

Does the routine survive fatigue?

Use the paper to refine checking until it earns its time.

See How Checking Fails.

Mock examinations and practice

Mocks should change practice.

Practice should change later mocks.

This reciprocal loop is the core.

If mock performance is weak but practice stays the same, simulation has become observation without control.

If practice changes but later mocks never test integration, repair has become local and uncertified.

See How Practice Fails.

Mock examinations and revision

Revision should not be completed first and then tested once.

Mocks can help allocate revision.

Stable topics move to maintenance.

Fragile topics return to retrieval.

Control failures receive performance training.

Mocks are one way revision learns where its next hour should go.

Mock examinations and exam technique

Technique is visible only when the clock and whole-paper constraints exist.

Reading gates, move-on rules, answer transfer, checking, return paths and recovery can all look unnecessary in short drills.

A full simulation reveals whether procedures actually run.

See How Exam Technique Fails.

Mock examinations and examination preparation

Preparation is the larger architecture.

Mocks are sensors inside it.

A sensor that reports but does not change control is wasted.

A preparation plan should have explicit places where mock evidence can alter priority, volume, timing or taper.

See How Exam Preparation Fails.

Mock examinations and performance under pressure

Pressure should not be added theatrically.

It already emerges from finite time, unseen questions, independence, consequence and uncertainty.

Good simulation lets the learner practise procedures while those constraints are active.

The goal is not to feel no pressure.

The goal is to preserve control under ordinary pressure.

Mock examinations and fault tolerance

A robust mock system asks not only whether the learner can perform when everything goes well.

It asks whether one local failure remains local.

Can the student recover from a difficult opening?

Can a lost method be contained?

Can one wrong section end without poisoning the next?

Can time drift be corrected?

Can checking continue after fatigue?

Fault tolerance is one of the differences between fragile readiness and real readiness.

The mock examination readiness ladder

A useful readiness ladder can look like this:

Stage 1: knowledge available with support.

Stage 2: knowledge available independently in topical work.

Stage 3: mixed sections completed within reasonable time.

Stage 4: full paper completed under representative conditions.

Stage 5: performance repeats across several representative samples.

Stage 6: known failure modes are contained and recovery works.

Stage 7: the learner has enough headroom and stability to taper rather than keep proving readiness.

The stages are not official categories.

They are a way to prevent one strong paper from carrying an oversized claim.

The mock examination stopping rule

Stop full simulations when:

• the major readiness questions have been answered;
• score range and completion are sufficiently stable for the target;
• known high-cost failure modes are controlled;
• another full paper is unlikely to change the training plan;
• targeted maintenance, sleep and recovery now have higher expected value.

Stopping does not mean abandoning practice.

It means shifting from measurement to preservation.

The mock examination restart rule

Reopen simulation when:

• a major new weakness appears;
• format or examination conditions change;
• performance becomes volatile;
• a large repair has been completed and needs reintegration;
• evidence has become stale enough that readiness is again uncertain.

A strong system can stop and restart according to evidence rather than fear.

The final fourteen-day mock strategy

The final two weeks are not the time for maximal mock volume.

They are the time for enough simulation to confirm the system, enough targeted repair to address remaining high-value weaknesses, and enough recovery to protect performance.

Where readiness is unstable, a representative paper can reveal what still needs work.

Where readiness is stable, additional full papers should justify their cost.

Use shorter confirmations, targeted sets and maintenance when full simulation no longer produces new information.

The exact schedule depends on the examination and learner. The principle is to stop confusing activity with readiness.

The final seven-day mock strategy

Nearer the examination, the question changes.

Can this simulation still change something useful?

If yes, run it with a clear purpose.

If no, do not create new emotional volatility merely to prove seriousness.

Use the final week to stabilise, maintain retrieval, protect sleep, rehearse compact procedures and preserve confidence that is already supported by evidence.

The final twenty-four-hour rule

A full mock immediately before a major examination often has poor expected value.

There may be little time to repair what it reveals, while fatigue and confidence effects remain large.

Use light retrieval, orientation, equipment checks and calm procedure review instead unless a very specific circumstance justifies otherwise.

The final day is usually for preservation, not measurement.

A parent interpretation guide

When a mock result arrives, parents can ask five questions instead of reacting to the number alone:

1. Was the paper representative and honestly timed?
2. Where were marks actually lost?
3. Which losses repeat?
4. What is the first weak link?
5. What changes before the next full paper?

This reframes the mock from judgement to control.

A low score can contain a small repairable bottleneck.

A high score can contain hidden instability.

The value is in the mechanism.

A tutor interpretation guide

After a mock, the tutor should resist two temptations.

First, explaining everything.

Second, assigning another mock immediately.

Instead:

1. Preserve the script.
2. Identify concentration of mark loss.
3. Trace the first weak link.
4. Choose one to three high-value repairs.
5. Train them at the smallest useful scale.
6. Retest independently.
7. Reintegrate later.

This is how small-group or individual teaching converts a full paper into targeted instruction rather than simply more homework.

A student self-review guide

The student can ask:

What surprised me?

Where did time disappear?

Which answer was wrong for a reason I did not understand?

Which error could a realistic check have caught?

Which correct answer was a guess?

Which topic looked weak only because I was tired?

What one change would most improve the next paper?

Self-review becomes more accurate over time when compared with teacher or tutor diagnosis.

The difference between a mock and a rehearsal

A mock is usually a measurement-heavy simulation.

A rehearsal can be more procedural.

You can rehearse opening routines, page navigation, checking, answer transfer, or transitions without sitting the entire paper.

This distinction matters because many students use a full mock to train something that could be trained more efficiently in ten minutes.

Use full simulation for full-system questions.

Use rehearsal for specific procedures.

The difference between a mock and a stress test

A stress test deliberately pushes beyond normal conditions to reveal limits.

A mock should usually remain representative if the purpose is readiness estimation.

Both can be useful.

Do not let a stress-test score control confidence about the normal event without interpretation.

The difference between a mock and a diagnostic paper

A diagnostic paper may be open-resource, untimed or interrupted for questioning because the aim is to expose thinking.

A readiness mock may need strict conditions because the aim is to estimate independent performance.

Again, mode clarity prevents confusion.

The difference between a mock and a past paper

A past paper is a source.

A mock is a use mode.

The same past paper can be used open-book for learning, section-by-section for practice, or under strict conditions as a mock.

Do not confuse the material with the procedure.

The difference between a mock and an examination

The examination is irreversible.

The mock is valuable precisely because it is not.

A failed mock can still improve the system.

That difference should reduce shame, not reduce effort.

Practice seriousness means giving the mock enough truthful effort to produce useful evidence while remembering that its purpose is to make later irreversible performance better.

The complete mock-examination operating protocol

For practical use, the entire article can be compressed into one sequence:

1. Choose the readiness question. Do not simulate without a reason.
2. Verify the target event. Current format, time, tools, interface, sequence and standards.
3. Declare the mode. Diagnostic, stretch, timed section, full readiness mock or event rehearsal.
4. Set conditions that match the claim.
5. Perform with honest independence.
6. Capture only useful evidence. Score, completion, timing, support, major error patterns and state where relevant.
7. Separate outcome from mechanism.
8. Trace the first weak link.
9. Prioritise one to three high-value repairs.
10. Train at the smallest useful scale.
11. Reattempt without support.
12. Vary and delay.
13. Restore realistic conditions.
14. Reintegrate in a later representative mock.
15. Update confidence and readiness.
16. Stop full simulations when new papers stop changing useful decisions.
17. Restart only when meaningful uncertainty returns.

This is how simulation becomes a learning instrument instead of examination theatre.

A final scene: the mock that did its job

The paper ends.

Kai Kai puts the pen down.

The score is seventy-six.

Last month it was seventy-nine.

The old version of the conversation would begin immediately.

“Why did the score drop?”

Alicia looks at the paper instead.

Every section was completed.

The final page, previously blank, is now finished.

The old pacing collapse did not return.

The repeated sign error did not return.

One unfamiliar data interpretation question caused six lost marks.

Tricia checks the paper conditions.

Strict time.

No prompts.

Representative source.

Normal sleep.

The seventy-six is credible evidence.

Not evidence that the learner has become worse.

Evidence that two old failure modes are now controlled and one new transfer weakness has appeared.

Kai Kai asks the question.

“So do we do another mock tomorrow?”

No.

They spend two days on unfamiliar data interpretation.

First with contrast.

Then mixed.

Then timed.

Then delayed.

A week later, another representative paper arrives.

The old pacing problem remains quiet.

The sign problem remains quiet.

The data interpretation holds.

The score rises.

More importantly, the system is different.

That is what the earlier mock was for.

Not to produce a number.

Not to scare the learner.

Not to prove seriousness.

Not to imitate an examination aesthetically.

The mock succeeded because it gave the learning system reliable enough evidence to make the next decision better.

That is the standard.

Fidelity where fidelity matters.

Diagnosis where diagnosis matters.

Repair before repetition.

Retesting before confidence.

Stopping before simulation becomes theatre.

A mock examination is useful when it makes the real examination less surprising and the learner more capable.

Everything else is paper, clocks and costume.