Jalan Azam families · premium 3-pax tutor guide
test ideas by asking what else should be true if they are correct
Reliable learners turn methods, interpretations and mechanisms into discriminating predictions, then update conclusions in proportion to the evidence.
Jalan Azam tutors are often searched by families who want precise English, Mathematics or Primary Science support and a truthful account of where lessons happen. This guide develops the learning purpose of how to test ideas by asking what else should be true if they are correct through worked examples, diagnostic choices and independent retries.
A correct answer can hide a fragile route, while a wrong answer can preserve a useful first decision. The chapters make those decisions visible, explain common errors and show how support is reduced. Every task and numerical value is an original editorial illustration, not an examination question or an actual learner result.
Bring the original task, the learner’s untouched attempt and a later correction to a consultation. Premium 3-pax tutorials are normally 1.5 hours weekly at 8 Fourth Avenue, Singapore 268674, near Sixth Avenue MRT in Bukit Timah, subject to suitable subject, level, group fit and current availability. Jalan Azam is the family’s starting locality, not a claimed eduKate branch.
Choose a reading route
Open the full chapter index · Use the diagnostic table · Review current services
Chapter index
FOUNDATION · CHAPTERS 1–4
APPLICATION · CHAPTERS 5–8
PRACTICE AND CONSULTATION · CHAPTERS 9–11
Why this matters
A claim becomes useful when it predicts something beyond the sentence itself. The learner asks: if this were true, what should also be observed, and what result would make me revise it? In counterfactual testing: asking what else should change, remain or become impossible if a proposed method, interpretation or explanation is true, the learner states the decision before acting. That short pause reduces avoidable drift while preserving the need to reason, choose and justify.
Worked teaching illustration
Claim: a table represents a constant rate. If true, equal increases in the input should create equal increases in the output, and output/input should remain constant for non-zero inputs. Values (1,4),(2,8),(3,12) fit; (4,17) would challenge it. This is an original editorial teaching illustration, not an examination question or an actual learner result. Its values and wording are deliberately visible so the reasoning can be checked line by line.
What the tutor diagnoses
The tutor distinguishes a prediction from a restatement. “The rate stays constant” repeats the claim; “doubling input doubles output” gives a consequence that can be checked. The tutor tests the provisional diagnosis with a nearby contrast rather than turning one error into a label. A check must be capable of producing a different result when the claim is wrong. A prompted correction and a fresh independent response are recorded separately.
Independent retry
Try this: Write two consequences of the claim that a sequence is arithmetic, then test 5,8,11,14 and 5,8,12,15. Close the model, name the governing decision, complete the new task on a clean page and use a check capable of exposing the expected error instead of merely repeating the same method.
Transfer beyond this task
Interpretations predict which later clues would fit; mechanisms predict how an outcome responds when a causal condition changes. Transfer keeps subject standards distinct: Mathematics needs valid relationships and calculation, English needs faithful meaning and evidence, and Science needs observable comparison, mechanism and appropriately bounded conclusions.
A calm parent check
Parent prompt: Ask, “What new observation should we expect if your explanation is right?” Keep the untouched attempt, smallest useful prompt, independent retry and later return together. That four-part record is more informative than a broad judgment such as careless or weak.
CHAPTER 2 OF 11 · MAKE A PREDICTION
2. Separate impossible, unlikely and unsupported
Why this matters
Counterfactual reasoning needs calibrated categories. Evidence may rule a claim out, make it less likely or simply fail to support it. Treating these as the same produces overconfident conclusions. In counterfactual testing: asking what else should change, remain or become impossible if a proposed method, interpretation or explanation is true, the learner states the decision before acting. That short pause reduces avoidable drift while preserving the need to reason, choose and justify.
Worked teaching illustration
A bag is defined to contain only red and blue counters. Drawing a green counter is impossible under the definition; drawing five reds in a row may be unlikely depending on proportions but remains possible; predicting the next colour without composition data is unsupported. This is an original editorial teaching illustration, not an examination question or an actual learner result. Its values and wording are deliberately visible so the reasoning can be checked line by line.
What the tutor diagnoses
The tutor asks which rule creates impossibility and which probability evidence changes likelihood. A surprised feeling is not a logical contradiction. The tutor tests the provisional diagnosis with a nearby contrast rather than turning one error into a label. A check must be capable of producing a different result when the claim is wrong. A prompted correction and a fresh independent response are recorded separately.
Independent retry
Try this: Classify three claims about a six-sided die: rolling 7, rolling 6 twice, and claiming 6 is most likely without evidence. Explain the category for each. Close the model, name the governing decision, complete the new task on a clean page and use a check capable of exposing the expected error instead of merely repeating the same method.
Transfer beyond this task
English clues can weaken a motive without disproving it; Science data can fail to distinguish two explanations. Transfer keeps subject standards distinct: Mathematics needs valid relationships and calculation, English needs faithful meaning and evidence, and Science needs observable comparison, mechanism and appropriately bounded conclusions.
A calm parent check
Parent prompt: Ask whether the evidence contradicts, reduces confidence or merely leaves the question open. Keep the untouched attempt, smallest useful prompt, independent retry and later return together. That four-part record is more informative than a broad judgment such as careless or weak.
CHAPTER 3 OF 11 · TEST MATHEMATICS STRUCTURES
3. Use substitution as a counterfactual check
Why this matters
Solving an equation proposes a value that should make the original relationship true. Substitution asks what the original would display if the proposal were correct. In counterfactual testing: asking what else should change, remain or become impossible if a proposed method, interpretation or explanation is true, the learner states the decision before acting. That short pause reduces avoidable drift while preserving the need to reason, choose and justify.
Worked teaching illustration
For 4x−7=21, solving gives x=7. Substitution yields 4(7)−7=28−7=21. A proposed x=6 would yield 17, directly contradicting the original equality. This is an original editorial teaching illustration, not an examination question or an actual learner result. Its values and wording are deliberately visible so the reasoning can be checked line by line.
What the tutor diagnoses
The tutor checks the original equation, not only a repeated inverse operation that may reproduce the same mistake. An independent check should have power to expose the expected error. The tutor tests the provisional diagnosis with a nearby contrast rather than turning one error into a label. A check must be capable of producing a different result when the claim is wrong. A prompted correction and a fresh independent response are recorded separately.
Independent retry
Try this: Solve 5y+3=38 to obtain y=7, substitute, then test the near-miss y=5 and explain which original condition it violates. Close the model, name the governing decision, complete the new task on a clean page and use a check capable of exposing the expected error instead of merely repeating the same method.
Transfer beyond this task
A source quotation checks an English interpretation against the passage; a predicted observation checks a Science mechanism against evidence. Transfer keeps subject standards distinct: Mathematics needs valid relationships and calculation, English needs faithful meaning and evidence, and Science needs observable comparison, mechanism and appropriately bounded conclusions.
A calm parent check
Parent prompt: Ask what the original statement would look like if the candidate were true. Keep the untouched attempt, smallest useful prompt, independent retry and later return together. That four-part record is more informative than a broad judgment such as careless or weak.
CHAPTER 4 OF 11 · TEST MATHEMATICS STRUCTURES
4. Test proportionality by changing scale
Why this matters
A proportional model predicts multiplicative behaviour. Adding a fixed amount or changing only one quantity may reveal that the relation is not proportional even when one pair fits. In counterfactual testing: asking what else should change, remain or become impossible if a proposed method, interpretation or explanation is true, the learner states the decision before acting. That short pause reduces avoidable drift while preserving the need to reason, choose and justify.
Worked teaching illustration
A recipe uses 300 g flour for 12 portions. Under proportional scaling, 18 portions need 300×18/12=450 g. Doubling portions would double flour; adding six portions does not mean adding six grams. This is an original editorial teaching illustration, not an examination question or an actual learner result. Its values and wording are deliberately visible so the reasoning can be checked line by line.
What the tutor diagnoses
The tutor contrasts a direct proportion with a fixed-fee model. If the learner uses cross-multiplication without naming the structure, one matching case can create false confidence. The tutor tests the provisional diagnosis with a nearby contrast rather than turning one error into a label. A check must be capable of producing a different result when the claim is wrong. A prompted correction and a fresh independent response are recorded separately.
Independent retry
Try this: Scale 250 mL of juice for 5 servings to 8 servings, obtaining 400 mL, then explain what observation would show a fixed amount is also present. Close the model, name the governing decision, complete the new task on a clean page and use a check capable of exposing the expected error instead of merely repeating the same method.
Transfer beyond this task
English frequency claims and Science dose–response models also need scale tests before extrapolation. Transfer keeps subject standards distinct: Mathematics needs valid relationships and calculation, English needs faithful meaning and evidence, and Science needs observable comparison, mechanism and appropriately bounded conclusions.
A calm parent check
Parent prompt: Ask what doubling one quantity should do to the other under the proposed model. Keep the untouched attempt, smallest useful prompt, independent retry and later return together. That four-part record is more informative than a broad judgment such as careless or weak.
CHAPTER 5 OF 11 · TEST MATHEMATICS STRUCTURES
5. Find a counterexample to an overbroad rule
Why this matters
Several confirming examples cannot prove a universal statement, but one valid counterexample can disprove it. Searching near boundaries or special values is often more efficient than trying more ordinary cases. In counterfactual testing: asking what else should change, remain or become impossible if a proposed method, interpretation or explanation is true, the learner states the decision before acting. That short pause reduces avoidable drift while preserving the need to reason, choose and justify.
Worked teaching illustration
Claim: squaring a number always makes it larger. The counterexample 1/2 gives (1/2)²=1/4, which is smaller; 0 and 1 are unchanged. The rule needs a restricted domain, such as numbers greater than 1. This is an original editorial teaching illustration, not an examination question or an actual learner result. Its values and wording are deliberately visible so the reasoning can be checked line by line.
What the tutor diagnoses
The tutor asks the learner to repair rather than merely reject the claim. Domain-aware revision turns a failed generalisation into precise understanding. The tutor tests the provisional diagnosis with a nearby contrast rather than turning one error into a label. A check must be capable of producing a different result when the claim is wrong. A prompted correction and a fresh independent response are recorded separately.
Independent retry
Try this: Test “multiplying by 3 always increases a number” with negative and zero values, then write a correct restricted version. Close the model, name the governing decision, complete the new task on a clean page and use a check capable of exposing the expected error instead of merely repeating the same method.
Transfer beyond this task
A single passage clue can challenge a universal interpretation; one controlled Science result can challenge a claimed mechanism. Transfer keeps subject standards distinct: Mathematics needs valid relationships and calculation, English needs faithful meaning and evidence, and Science needs observable comparison, mechanism and appropriately bounded conclusions.
A calm parent check
Parent prompt: Ask which edge case is most likely to break the word always. Keep the untouched attempt, smallest useful prompt, independent retry and later return together. That four-part record is more informative than a broad judgment such as careless or weak.
CHAPTER 6 OF 11 · TEST ENGLISH INTERPRETATIONS
6. Compare two motives through discriminating clues
Why this matters
Two interpretations may explain the same behaviour. A useful counterfactual asks which additional clue would be more expected under one motive than the other. In counterfactual testing: asking what else should change, remain or become impossible if a proposed method, interpretation or explanation is true, the learner states the decision before acting. That short pause reduces avoidable drift while preserving the need to reason, choose and justify.
Worked teaching illustration
A pupil stays after class. She may be confused or may want to help arrange materials. Repeatedly checking the same worksheet favours confusion; carrying chairs without looking at the work favours helpfulness. Staying alone supports both and does not discriminate. This is an original editorial teaching illustration, not an examination question or an actual learner result. Its values and wording are deliberately visible so the reasoning can be checked line by line.
What the tutor diagnoses
The tutor sorts clues into supports A, supports B and fits both. This prevents evidence volume from being mistaken for evidence quality. The tutor tests the provisional diagnosis with a nearby contrast rather than turning one error into a label. A check must be capable of producing a different result when the claim is wrong. A prompted correction and a fresh independent response are recorded separately.
Independent retry
Try this: For a character arriving early, propose two motives and write one clue that would discriminate between them without claiming certainty. Close the model, name the governing decision, complete the new task on a clean page and use a check capable of exposing the expected error instead of merely repeating the same method.
Transfer beyond this task
Mathematical methods are also compared by predictions; Science explanations compete through observations that would differ under each mechanism. Transfer keeps subject standards distinct: Mathematics needs valid relationships and calculation, English needs faithful meaning and evidence, and Science needs observable comparison, mechanism and appropriately bounded conclusions.
A calm parent check
Parent prompt: Ask which clue would make the alternatives behave differently. Keep the untouched attempt, smallest useful prompt, independent retry and later return together. That four-part record is more informative than a broad judgment such as careless or weak.
CHAPTER 7 OF 11 · TEST ENGLISH INTERPRETATIONS
7. Test an argument with a fair counterexample
Why this matters
A strong argument anticipates a relevant case that appears to resist its claim. The writer addresses the case by revising scope, distinguishing conditions or supplying evidence, not by ignoring it. In counterfactual testing: asking what else should change, remain or become impossible if a proposed method, interpretation or explanation is true, the learner states the decision before acting. That short pause reduces avoidable drift while preserving the need to reason, choose and justify.
Worked teaching illustration
Claim: all homework should be completed digitally because it is more convenient. A counterexample is a pupil with unreliable device access. A stronger claim may concern assignments for which access and accessibility have been secured. This is an original editorial teaching illustration, not an examination question or an actual learner result. Its values and wording are deliberately visible so the reasoning can be checked line by line.
What the tutor diagnoses
The tutor checks whether the counterexample attacks the actual claim rather than an exaggerated version. The revision should narrow only as far as necessary. The tutor tests the provisional diagnosis with a nearby contrast rather than turning one error into a label. A check must be capable of producing a different result when the claim is wrong. A prompted correction and a fresh independent response are recorded separately.
Independent retry
Try this: Write a claim about group study, generate one fair counterexample and revise the claim so both benefit and condition remain visible. Close the model, name the governing decision, complete the new task on a clean page and use a check capable of exposing the expected error instead of merely repeating the same method.
Transfer beyond this task
In Mathematics, domain restrictions repair universal rules; in Science, boundary conditions repair mechanisms and generalisations. Transfer keeps subject standards distinct: Mathematics needs valid relationships and calculation, English needs faithful meaning and evidence, and Science needs observable comparison, mechanism and appropriately bounded conclusions.
A calm parent check
Parent prompt: Ask the learner to strengthen a claim by making it harder to disprove, not louder. Keep the untouched attempt, smallest useful prompt, independent retry and later return together. That four-part record is more informative than a broad judgment such as careless or weak.
CHAPTER 8 OF 11 · TEST SCIENCE MECHANISMS
8. Predict the result of changing one Science condition
Why this matters
A mechanism should support a directional prediction. Changing one relevant condition while keeping others comparable reveals whether the explanation connects to the outcome. In counterfactual testing: asking what else should change, remain or become impossible if a proposed method, interpretation or explanation is true, the learner states the decision before acting. That short pause reduces avoidable drift while preserving the need to reason, choose and justify.
Worked teaching illustration
If greater exposed surface area increases evaporation under comparable conditions, spreading equal water volumes in a wide dish should leave less water after the same time than using a narrow container. This predicts direction, not an exact amount. This is an original editorial teaching illustration, not an examination question or an actual learner result. Its values and wording are deliberately visible so the reasoning can be checked line by line.
What the tutor diagnoses
The tutor asks whether temperature, airflow and starting volume were also changed. If so, the counterfactual no longer isolates surface area. The tutor tests the provisional diagnosis with a nearby contrast rather than turning one error into a label. A check must be capable of producing a different result when the claim is wrong. A prompted correction and a fresh independent response are recorded separately.
Independent retry
Try this: Predict and explain what happens to dissolving time when identical sugar crystals are stirred in one cup but not another, with temperature and volume controlled. Close the model, name the governing decision, complete the new task on a clean page and use a check capable of exposing the expected error instead of merely repeating the same method.
Transfer beyond this task
English interpretations also predict likely behaviour; Mathematics models predict how outputs respond to controlled input changes. Transfer keeps subject standards distinct: Mathematics needs valid relationships and calculation, English needs faithful meaning and evidence, and Science needs observable comparison, mechanism and appropriately bounded conclusions.
A calm parent check
Parent prompt: Ask for the changed condition, expected direction and one important control. Keep the untouched attempt, smallest useful prompt, independent retry and later return together. That four-part record is more informative than a broad judgment such as careless or weak.
Why this matters
A missing predicted observation can weaken an explanation, but absence is meaningful only when the observation should have been detectable. Poor measurement or too short a duration may make the test inconclusive. In counterfactual testing: asking what else should change, remain or become impossible if a proposed method, interpretation or explanation is true, the learner states the decision before acting. That short pause reduces avoidable drift while preserving the need to reason, choose and justify.
Worked teaching illustration
A mechanism predicts a measurable temperature rise, but the thermometer resolution is 1°C and the expected change is smaller. No visible rise does not disprove the mechanism; the measurement may be insensitive. This is an original editorial teaching illustration, not an examination question or an actual learner result. Its values and wording are deliberately visible so the reasoning can be checked line by line.
What the tutor diagnoses
The tutor asks whether the test had enough time, resolution and repetitions. “We did not see it” is not identical to “it did not occur”. The tutor tests the provisional diagnosis with a nearby contrast rather than turning one error into a label. A check must be capable of producing a different result when the claim is wrong. A prompted correction and a fresh independent response are recorded separately.
Independent retry
Try this: Review a plant-growth test lasting one day and explain why no height difference may be inconclusive; redesign the observation period. Close the model, name the governing decision, complete the new task on a clean page and use a check capable of exposing the expected error instead of merely repeating the same method.
Transfer beyond this task
A passage may omit a motive because the narrator lacks access, not because the motive is absent. Transfer keeps subject standards distinct: Mathematics needs valid relationships and calculation, English needs faithful meaning and evidence, and Science needs observable comparison, mechanism and appropriately bounded conclusions.
A calm parent check
Parent prompt: Ask whether the predicted evidence had a fair chance to appear. Keep the untouched attempt, smallest useful prompt, independent retry and later return together. That four-part record is more informative than a broad judgment such as careless or weak.
Why this matters
The table separates restatements, repeated checks, irrelevant counterexamples, uncontrolled Science changes and overinterpretation of absence. Each requires a different next contrast. In counterfactual testing: asking what else should change, remain or become impossible if a proposed method, interpretation or explanation is true, the learner states the decision before acting. That short pause reduces avoidable drift while preserving the need to reason, choose and justify.
Worked teaching illustration
A learner who restates a claim needs consequence practice. One who uses the same algebra twice needs an independent check. One who changes temperature and surface area needs control repair. This is an original editorial teaching illustration, not an examination question or an actual learner result. Its values and wording are deliberately visible so the reasoning can be checked line by line.
What the tutor diagnoses
The tutor verifies the row with a fresh case and updates the diagnosis rather than treating it as permanent. The tutor tests the provisional diagnosis with a nearby contrast rather than turning one error into a label. A check must be capable of producing a different result when the claim is wrong. A prompted correction and a fresh independent response are recorded separately.
Independent retry
Try this: Choose one row, write a discriminating prediction and complete a new task without the table. Close the model, name the governing decision, complete the new task on a clean page and use a check capable of exposing the expected error instead of merely repeating the same method.
Transfer beyond this task
The table makes checking purposeful and finite for families. Transfer keeps subject standards distinct: Mathematics needs valid relationships and calculation, English needs faithful meaning and evidence, and Science needs observable comparison, mechanism and appropriately bounded conclusions.
A calm parent check
Parent prompt: Bring both the claim and the evidence so the predicted consequence can be judged. Keep the untouched attempt, smallest useful prompt, independent retry and later return together. That four-part record is more informative than a broad judgment such as careless or weak.
| Observed pattern | Likely decision point | Useful next task |
|---|---|---|
| Claim is merely repeated | No observable consequence was named | Write an if–then prediction |
| Equation is checked by repeating the same steps | Check is not independent | Substitute into the original |
| Several examples support “always” | Edge cases were not tested | Search for a counterexample and repair scope |
| Two motives fit the same clue | Evidence does not discriminate | Predict a clue that differs between motives |
| No effect is observed once | Detection limits were ignored | Check duration, resolution and repetitions |
CHAPTER 11 OF 11 · UPDATE THE CLAIM
11. Build an independent counterfactual routine
Why this matters
Independence means the learner can propose a claim, predict a consequence, design a fair check, interpret confirming or conflicting evidence and update the claim by the required amount. In counterfactual testing: asking what else should change, remain or become impossible if a proposed method, interpretation or explanation is true, the learner states the decision before acting. That short pause reduces avoidable drift while preserving the need to reason, choose and justify.
Worked teaching illustration
A mixed retry uses an algebra solution, a character interpretation and a Science mechanism. The learner writes “If true, then…” and names one result that would require revision. This is an original editorial teaching illustration, not an examination question or an actual learner result. Its values and wording are deliberately visible so the reasoning can be checked line by line.
What the tutor diagnoses
The tutor scores quality of prediction and update separately from initial correctness. A revised claim can show stronger learning than an untested lucky answer. The tutor tests the provisional diagnosis with a nearby contrast rather than turning one error into a label. A check must be capable of producing a different result when the claim is wrong. A prompted correction and a fresh independent response are recorded separately.
Independent retry
Try this: After a delay, complete three new tests and explain whether each result confirms, weakens, contradicts or leaves the claim unresolved. Close the model, name the governing decision, complete the new task on a clean page and use a check capable of exposing the expected error instead of merely repeating the same method.
Transfer beyond this task
The routine transfers because every subject asks what follows from an idea, while using different evidence rules. Transfer keeps subject standards distinct: Mathematics needs valid relationships and calculation, English needs faithful meaning and evidence, and Science needs observable comparison, mechanism and appropriately bounded conclusions.
A calm parent check
Parent prompt: Look for a learner who can change their mind precisely when the evidence changes. Keep the untouched attempt, smallest useful prompt, independent retry and later return together. That four-part record is more informative than a broad judgment such as careless or weak.
Contents · Previous chapter · Continue to the broad Bedok tutor guide
How the learning cycle develops this skill
Counterfactual testing: asking what else should change, remain or become impossible if a proposed method, interpretation or explanation is true starts with a low-stakes cold attempt. The tutor observes whether the learner can derive a consequence that could genuinely challenge the claim before supplying a route. The untouched page preserves evidence about interpretation, selection, execution and checking that would disappear if a polished answer were copied first.
The first model isolates one claim–prediction–test–update sequence. The tutor explains what the decision protects, where it can fail and what evidence would reveal failure. The learner sees a conditional choice rather than a decorative technique or slogan.
Guided practice changes one surface feature while retaining the same intellectual job. Numbers, wording, representation or evidence strength may change. The learner predicts whether the earlier decision still applies before calculating, drafting or concluding.
Checking is operational: original-equation substitution, boundary counterexamples, discriminating clues and one-change Science comparisons. A useful check can reveal a named kind of drift. A second reading, a neat page or agreement with a peer is not automatically evidence of control.
The independent retry removes the route label and closes the worked page. The learner states the demand, chooses a path, completes the task and explains one rejected option or likely failure. Prompt level is recorded explicitly.
A delayed return mixes the target decision with one secure skill. Interleaving begins only after each component is understood. Selection failure returns to contrasting cases; execution failure returns to the later step.
A three-student group can compare routes, challenge assumptions and propose checks, but every learner then completes a fresh task independently. Discussion supplies useful contrasts; it does not replace personal reconstruction.
For families, the useful record is compact: original demand, first attempt, earliest unreliable decision, support supplied, independent retry and delayed return. This supports diagnosis before tuition and helps confirm subject, level, pace and class fit.
Frequently asked questions
Is there an eduKate branch on Jalan Azam?
No. Jalan Azam identifies the family’s starting locality; teaching described here is at 8 Fourth Avenue near Sixth Avenue MRT.
Is a counterfactual just guessing?
No. It is a structured prediction derived from a claim, method or mechanism and checked against relevant evidence.
Does one failed prediction always disprove an idea?
No. The test may be uncontrolled or insensitive, so design quality and detection limits also matter.
What should families bring?
Bring the prompt, first explanation, check used and a later independent retry.
Jalan Azam location and consultation notes
Jalan Azam appears as entry 102 in the 800 Super PRB Sector (Bedok) Landed Recycling & Garden Waste Collection Schedule. This continuation treats the schedule’s complete 356-entry street list as a bounded authoritative candidate inventory. The schedule confirms the road label used for collection planning; it does not indicate an eduKate teaching branch there.
Related reading: Tutors | Bedok, Tutors | East Coast Road, Tutors | Katong and the preserved Secondary 1 Mathematics Tutor Clementi reference.
To discuss the learner’s current starting point, use the eduKate Singapore consultation gateway or review current eduKateSG services. A conversation should confirm subject, level, group fit, timetable availability and whether travel to 8 Fourth Avenue is practical before enrolment.
Diagnosis before tuition. Start with the actual question, the learner’s working and the first decision that became unreliable. The aim is a calmer, more independent next attempt—not simply more pages of practice.