How to prepare for problem-solving exams is not to memorize worked solutions. Strong preparation teaches students to recognize structure, choose methods, remain productive when stuck, show clear working and verify whether an answer makes sense.
Problem-based papers often remove the chapter labels and familiar sequences found in worksheets. The learner must decide what kind of problem is present before any procedure begins, then execute under time without the solution visible.
This guide turns problem-solving exam preparation into an operating system for students, parents and tutors: interpret the task, make the decision visible, practise under realistic constraints, check the result, repair the first meaningful weakness and retest later.
The 60-Second Answer
Practise new and mixed problems. Before solving, identify what is known, what is unknown and which relationships may apply. Try before checking. Estimate where useful. Show clear working. Check units and reasonableness. Analyse the first wrong decision and retest on a fresh variant.
Definition: Problem-solving exam preparation is deliberate rehearsal of interpreting unfamiliar problems, selecting appropriate methods, carrying out valid procedures and checking solutions under realistic assessment conditions.
Why Problem-Solving Exam Preparation Matters
The exam tests method selection as well as execution.
Blocked worksheets can hide the selection problem because the heading already reveals the technique.
New and mixed problems force the learner to recognise structure and transfer methods rather than reproduce one surface pattern.
Strong exam performance is not produced by content knowledge alone. The learner must also express that knowledge in the form, scope, sequence and time the paper demands.
The Operating Model
Understand → Classify → Represent → Plan → Execute → Estimate → Check → Diagnose.
The model deliberately separates interpretation, strategy choice, execution, checking and later diagnosis. When those stages are collapsed, students often describe every failure as “careless” or “I did not know it”, which hides the actual teaching target.
How to prepare for problem-solving exams: Step by Step
1. Understand the problem
Identify what is being asked, what information is given and which parts must be answered.
Make this step observable. Ask the learner what evidence shows the step was completed correctly, what a plausible wrong version would look like, and what they would do if uncertainty remained.
2. Classify the structure
Identify the concept or method family and distinguish it from nearby alternatives.
Make this step observable. Ask the learner what evidence shows the step was completed correctly, what a plausible wrong version would look like, and what they would do if uncertainty remained.
3. Represent the information
Use a diagram, equation, table or labelled quantities when useful.
Make this step observable. Ask the learner what evidence shows the step was completed correctly, what a plausible wrong version would look like, and what they would do if uncertainty remained.
4. Attempt before checking
Delay answer-key access until a real attempt exists.
Make this step observable. Ask the learner what evidence shows the step was completed correctly, what a plausible wrong version would look like, and what they would do if uncertainty remained.
5. Estimate or predict
Predict sign, range or rough magnitude where possible.
Make this step observable. Ask the learner what evidence shows the step was completed correctly, what a plausible wrong version would look like, and what they would do if uncertainty remained.
6. Execute clearly
Use consistent notation, logical steps and units.
Make this step observable. Ask the learner what evidence shows the step was completed correctly, what a plausible wrong version would look like, and what they would do if uncertainty remained.
7. Check reasonableness
Ask whether the result fits the units, context and expected scale.
Make this step observable. Ask the learner what evidence shows the step was completed correctly, what a plausible wrong version would look like, and what they would do if uncertainty remained.
8. Analyse the first wrong step
Trace errors to the earliest unsupported decision.
Make this step observable. Ask the learner what evidence shows the step was completed correctly, what a plausible wrong version would look like, and what they would do if uncertainty remained.
What Strong Performance Looks Like
- Methods are selected deliberately.
- New problems are attempted.
- Answer keys stay hidden during first attempts.
- Working is readable.
- Units are shown.
- Estimates support checking.
- Mixed problem types are normal.
- Wrong-method errors are separated from calculation errors.
These are behaviours, not personality labels. They can be rehearsed repeatedly until the student can perform them with less prompting and under greater time pressure.
Common Failure Modes
1. Memorized pattern matching
The student recognizes layout rather than structure.
Repair the failure by isolating one clean instance, requiring an independent attempt, comparing with a reliable standard and then testing the same underlying decision on a fresh item.
2. Solution peeking
The answer appears before the learner has generated a route.
Repair the failure by isolating one clean instance, requiring an independent attempt, comparing with a reliable standard and then testing the same underlying decision on a fresh item.
3. Blocked practice only
Every item uses one method.
Repair the failure by isolating one clean instance, requiring an independent attempt, comparing with a reliable standard and then testing the same underlying decision on a fresh item.
4. No estimation
Calculator output is accepted automatically.
Repair the failure by isolating one clean instance, requiring an independent attempt, comparing with a reliable standard and then testing the same underlying decision on a fresh item.
5. Messy working
Errors become difficult to locate.
Repair the failure by isolating one clean instance, requiring an independent attempt, comparing with a reliable standard and then testing the same underlying decision on a fresh item.
6. Stuck equals failure
The learner quits too early.
Repair the failure by isolating one clean instance, requiring an independent attempt, comparing with a reliable standard and then testing the same underlying decision on a fresh item.
7. Final-answer fixation
Only the last number is examined.
Repair the failure by isolating one clean instance, requiring an independent attempt, comparing with a reliable standard and then testing the same underlying decision on a fresh item.
8. Same-question correction only
The original problem is redone repeatedly.
Repair the failure by isolating one clean instance, requiring an independent attempt, comparing with a reliable standard and then testing the same underlying decision on a fresh item.
How the Strategy Changes Across Subjects
Mathematics
Practise structure recognition, formula conditions, algebraic control, mixed sets and clear checks.
The subject standard remains decisive. General exam strategy can protect marks, but it cannot replace knowing what counts as valid reasoning, relevant evidence or sufficient working in this discipline.
Science
Include calculations, graphs, variables, experiments and physical reasonableness.
The subject standard remains decisive. General exam strategy can protect marks, but it cannot replace knowing what counts as valid reasoning, relevant evidence or sufficient working in this discipline.
Economics
Combine models, diagrams, calculations and interpretation rather than treating numbers as isolated answers.
The subject standard remains decisive. General exam strategy can protect marks, but it cannot replace knowing what counts as valid reasoning, relevant evidence or sufficient working in this discipline.
Technical subjects
Use representative problem families and realistic notation, units and time constraints.
The subject standard remains decisive. General exam strategy can protect marks, but it cannot replace knowing what counts as valid reasoning, relevant evidence or sufficient working in this discipline.
Primary School, Secondary School and Examination Years
Primary school
Use short mixed sets and simple prompts: What do you know? What are you finding? Which method might help? Does the answer make sense?
Secondary school
Use mixed problem classification, timed sections and error logs separating method choice, algebra, arithmetic, reading and checking.
Before major examinations
Use new past-paper-level problems under realistic timing. Practise moving on from stalled questions and returning later.
A 60-Minute Practice Session
- 10 minutes — decode or classify short examples related to problem-solving exam preparation.
- 10 minutes — inspect one strong model and identify the decisions that control quality.
- 15 minutes — complete a guided attempt with only targeted prompts.
- 15 minutes — complete a fresh item independently under a defined time limit.
- 5 minutes — compare against a mark scheme, model or reliable source.
- 5 minutes — record the first meaningful error and schedule a delayed retest.
The exact timing can change, but every session should contain independent evidence. Watching a tutor perform the method is not enough.
Diagnostic Checklist
- Can the learner state the unknown?
- Can they identify relevant information?
- Can they choose a method without a heading?
- Can they remain productive when stuck?
- Is working readable?
- Are units checked?
- Is the final answer plausible?
- Can they identify the first wrong step?
If several checks fail, begin with the earliest decision in the chain. A misread question should be repaired before polishing the final answer; a wrong method choice should be repaired before drilling arithmetic speed.
Practice Laboratory
Practice 1: Method-only classification
Classify ten mixed questions without solving them.
After the attempt, require the learner to explain the controlling decision in one sentence. Then alter one surface feature so success cannot depend only on remembering the previous item.
Practice 2: No-solution interval
Work for five minutes before viewing any solution.
After the attempt, require the learner to explain the controlling decision in one sentence. Then alter one surface feature so success cannot depend only on remembering the previous item.
Practice 3: Estimate first
Predict sign and rough magnitude before calculating.
After the attempt, require the learner to explain the controlling decision in one sentence. Then alter one surface feature so success cannot depend only on remembering the previous item.
Practice 4: Representation swap
Turn words into a diagram or equation.
After the attempt, require the learner to explain the controlling decision in one sentence. Then alter one surface feature so success cannot depend only on remembering the previous item.
Practice 5: Wrong-method contrast
Compare a plausible wrong method with the correct one.
After the attempt, require the learner to explain the controlling decision in one sentence. Then alter one surface feature so success cannot depend only on remembering the previous item.
Practice 6: Clear-working audit
Rewrite a messy solution in readable steps.
After the attempt, require the learner to explain the controlling decision in one sentence. Then alter one surface feature so success cannot depend only on remembering the previous item.
Practice 7: Stuck-question rehearsal
Leave a stalled item with a re-entry cue and return later.
After the attempt, require the learner to explain the controlling decision in one sentence. Then alter one surface feature so success cannot depend only on remembering the previous item.
Practice 8: Fresh-variant retest
After correction, solve a changed problem using the same principle.
After the attempt, require the learner to explain the controlling decision in one sentence. Then alter one surface feature so success cannot depend only on remembering the previous item.
Three Student Cases
Case 1: Maren memorizes examples
Maren reproduces familiar solutions but hesitates when wording changes.
Her tutor mixes structurally similar questions and requires method classification before calculation.
Performance becomes more flexible as selection is practised.
The teaching logic is diagnosis → targeted change → independent retest → delayed verification. The fictional details matter less than the decision that changed.
Case 2: Iona checks answers immediately
Iona opens the solution whenever she feels uncertain.
A no-solution interval forces her to write knowns, unknowns and a route first.
Wrong starts become visible and teachable.
The teaching logic is diagnosis → targeted change → independent retest → delayed verification. The fictional details matter less than the decision that changed.
Case 3: Leonie loses marks through working
Leonie knows the method but compresses steps and loses signs or units.
Clearer working and estimation become required.
Avoidable errors become easier to catch.
The teaching logic is diagnosis → targeted change → independent retest → delayed verification. The fictional details matter less than the decision that changed.
How Parents Can Help Without Taking Over
Parents can ask: “What exactly is this question asking?”, “What clue tells you the method?”, “How many marks or parts are involved?”, “What would make you move on?”, and “What will you check before leaving it?” The purpose is to expose reasoning, not provide the answer.
Once the learner can perform a decision independently, remove the prompt. Support should transfer control rather than become permanent supervision.
How Tutors Can Use a Three-Student Small Group
A three-student tutorial makes problem-solving exam preparation visible because learners can compare how they interpret the same prompt, choose different routes and check one another’s assumptions. One learner attempts, another questions, and the third checks task coverage or method validity.
After discussion, every learner completes a fresh item independently. Shared insight is useful only when it transfers into individual performance.
How to Measure Progress
- Method-selection errors decline.
- New problems feel less unfamiliar.
- Solution peeking decreases.
- Working becomes clearer.
- Reasonableness checks become automatic.
- Timed completion improves.
- Fresh variants succeed more often.
- Repeated error families shrink.
Marks are the final indicator, but process improvements usually appear first: fewer missed subparts, faster method selection, better pacing, fewer repeated errors and more useful final checks.
A Four-Week Implementation Plan
Week 1 — Make the decisions visible
Use short examples and identify where problem-solving exam preparation breaks first. Keep time pressure moderate so diagnosis is clear.
Keep only the routines that measurably improve independent performance. Exam strategy should become simpler as it becomes internalised.
Week 2 — Reduce prompts
Use fresh items with fewer hints. Require the learner to justify the strategy before checking.
Keep only the routines that measurably improve independent performance. Exam strategy should become simpler as it becomes internalised.
Week 3 — Add realistic time and variation
Change wording, order, representation and mark allocation. Practise move-on and checking routines.
Keep only the routines that measurably improve independent performance. Exam strategy should become simpler as it becomes internalised.
Week 4 — Simulate
Use full or representative exam sections under authentic conditions. Review both knowledge errors and execution errors.
Keep only the routines that measurably improve independent performance. Exam strategy should become simpler as it becomes internalised.
Evidence and Responsible Use
Cornell’s current exam guidance recommends practising new problems, identifying knowns and unknowns, resisting premature answer checking, showing work, estimating where useful and checking reasonableness. Princeton similarly recommends working through new problems, identifying common techniques and practising under test-like conditions before consulting solutions.
No test-taking technique can manufacture subject knowledge that was never learned. Strategy protects, retrieves and expresses knowledge; content teaching and genuine practice remain essential.
- Cornell: How to Tackle Exam Questions
- Princeton: Exam Success—Taking Exams
- Princeton: Preparing for Exams
Frequently Asked Questions
Should I repeat the same problems?
Redo once to verify correction, then use fresh variants to test transfer.
When should I look at the solution?
After a genuine attempt and useful stuck work.
Should all practice be timed?
No. Build understanding first, then increase timed work.
Why estimate first?
It provides a plausibility check for signs, units and calculator output.
What if I choose the wrong method repeatedly?
Use contrast sets where similar-looking questions require different methods.
How much working should I show?
Enough to make reasoning clear and meet assessment expectations.
Can worked examples help?
Yes for initial learning, provided support is later faded.
How do I improve speed?
Stabilise method recognition and prerequisite fluency before forcing faster execution.
Helpful Reading on eduKateSingapore
- How to Use Practice Exams
- How to Check Your Work
- How to Learn From Mistakes
- Why Students Stay on a Stuck Exam Question Too Long
Problem-Solving Exams Reward Method Selection as Much as Procedure
A student cannot prepare only by watching correct solutions and expect flexible performance on unfamiliar questions.
Classify. Represent. Attempt. Estimate. Work clearly. Check. Diagnose. Then solve a fresh variant.
Properly taught kids shine a bright light into the future.
Extended Diagnostic Workshops
Workshop 1: Understand the problem × Solution peeking
Create one fresh exam-style task where the learner practises understand the problem while watching specifically for solution peeking. Begin under a defined time limit. Identify what is being asked, what information is given and which parts must be answered. Require the student to state the task, commit to a strategy and mark the first uncertainty before checking.
The answer appears before the learner has generated a route. After correction, change one meaningful feature—wording, values, representation, evidence, mark allocation, order or time pressure. Then use this related exercise: Predict sign and rough magnitude before calculating. Schedule a delayed retest so the next success cannot be explained only by immediate familiarity.
Workshop 2: Classify the structure × Messy working
Create one fresh exam-style task where the learner practises classify the structure while watching specifically for messy working. Begin under a defined time limit. Identify the concept or method family and distinguish it from nearby alternatives. Require the student to state the task, commit to a strategy and mark the first uncertainty before checking.
Errors become difficult to locate. After correction, change one meaningful feature—wording, values, representation, evidence, mark allocation, order or time pressure. Then use this related exercise: Compare a plausible wrong method with the correct one. Schedule a delayed retest so the next success cannot be explained only by immediate familiarity.
Workshop 3: Represent the information × Same-question correction only
Create one fresh exam-style task where the learner practises represent the information while watching specifically for same-question correction only. Begin under a defined time limit. Use a diagram, equation, table or labelled quantities when useful. Require the student to state the task, commit to a strategy and mark the first uncertainty before checking.
The original problem is redone repeatedly. After correction, change one meaningful feature—wording, values, representation, evidence, mark allocation, order or time pressure. Then use this related exercise: Leave a stalled item with a re-entry cue and return later. Schedule a delayed retest so the next success cannot be explained only by immediate familiarity.
Workshop 4: Attempt before checking × Blocked practice only
Create one fresh exam-style task where the learner practises attempt before checking while watching specifically for blocked practice only. Begin under a defined time limit. Delay answer-key access until a real attempt exists. Require the student to state the task, commit to a strategy and mark the first uncertainty before checking.
Every item uses one method. After correction, change one meaningful feature—wording, values, representation, evidence, mark allocation, order or time pressure. Then use this related exercise: Classify ten mixed questions without solving them. Schedule a delayed retest so the next success cannot be explained only by immediate familiarity.
Workshop 5: Estimate or predict × Stuck equals failure
Create one fresh exam-style task where the learner practises estimate or predict while watching specifically for stuck equals failure. Begin under a defined time limit. Predict sign, range or rough magnitude where possible. Require the student to state the task, commit to a strategy and mark the first uncertainty before checking.
The learner quits too early. After correction, change one meaningful feature—wording, values, representation, evidence, mark allocation, order or time pressure. Then use this related exercise: Predict sign and rough magnitude before calculating. Schedule a delayed retest so the next success cannot be explained only by immediate familiarity.
Workshop 6: Execute clearly × Memorized pattern matching
Create one fresh exam-style task where the learner practises execute clearly while watching specifically for memorized pattern matching. Begin under a defined time limit. Use consistent notation, logical steps and units. Require the student to state the task, commit to a strategy and mark the first uncertainty before checking.
The student recognizes layout rather than structure. After correction, change one meaningful feature—wording, values, representation, evidence, mark allocation, order or time pressure. Then use this related exercise: Compare a plausible wrong method with the correct one. Schedule a delayed retest so the next success cannot be explained only by immediate familiarity.
Workshop 7: Check reasonableness × No estimation
Create one fresh exam-style task where the learner practises check reasonableness while watching specifically for no estimation. Begin under a defined time limit. Ask whether the result fits the units, context and expected scale. Require the student to state the task, commit to a strategy and mark the first uncertainty before checking.
Calculator output is accepted automatically. After correction, change one meaningful feature—wording, values, representation, evidence, mark allocation, order or time pressure. Then use this related exercise: Leave a stalled item with a re-entry cue and return later. Schedule a delayed retest so the next success cannot be explained only by immediate familiarity.
Workshop 8: Analyse the first wrong step × Final-answer fixation
Create one fresh exam-style task where the learner practises analyse the first wrong step while watching specifically for final-answer fixation. Begin under a defined time limit. Trace errors to the earliest unsupported decision. Require the student to state the task, commit to a strategy and mark the first uncertainty before checking.
Only the last number is examined. After correction, change one meaningful feature—wording, values, representation, evidence, mark allocation, order or time pressure. Then use this related exercise: Classify ten mixed questions without solving them. Schedule a delayed retest so the next success cannot be explained only by immediate familiarity.
Workshop 9: Understand the problem × Solution peeking
Create one fresh exam-style task where the learner practises understand the problem while watching specifically for solution peeking. Begin under a defined time limit. Identify what is being asked, what information is given and which parts must be answered. Require the student to state the task, commit to a strategy and mark the first uncertainty before checking.
The answer appears before the learner has generated a route. After correction, change one meaningful feature—wording, values, representation, evidence, mark allocation, order or time pressure. Then use this related exercise: Predict sign and rough magnitude before calculating. Schedule a delayed retest so the next success cannot be explained only by immediate familiarity.
Workshop 10: Classify the structure × Messy working
Create one fresh exam-style task where the learner practises classify the structure while watching specifically for messy working. Begin under a defined time limit. Identify the concept or method family and distinguish it from nearby alternatives. Require the student to state the task, commit to a strategy and mark the first uncertainty before checking.
Errors become difficult to locate. After correction, change one meaningful feature—wording, values, representation, evidence, mark allocation, order or time pressure. Then use this related exercise: Compare a plausible wrong method with the correct one. Schedule a delayed retest so the next success cannot be explained only by immediate familiarity.
Workshop 11: Represent the information × Same-question correction only
Create one fresh exam-style task where the learner practises represent the information while watching specifically for same-question correction only. Begin under a defined time limit. Use a diagram, equation, table or labelled quantities when useful. Require the student to state the task, commit to a strategy and mark the first uncertainty before checking.
The original problem is redone repeatedly. After correction, change one meaningful feature—wording, values, representation, evidence, mark allocation, order or time pressure. Then use this related exercise: Leave a stalled item with a re-entry cue and return later. Schedule a delayed retest so the next success cannot be explained only by immediate familiarity.
Workshop 12: Attempt before checking × Blocked practice only
Create one fresh exam-style task where the learner practises attempt before checking while watching specifically for blocked practice only. Begin under a defined time limit. Delay answer-key access until a real attempt exists. Require the student to state the task, commit to a strategy and mark the first uncertainty before checking.
Every item uses one method. After correction, change one meaningful feature—wording, values, representation, evidence, mark allocation, order or time pressure. Then use this related exercise: Classify ten mixed questions without solving them. Schedule a delayed retest so the next success cannot be explained only by immediate familiarity.
Workshop 13: Estimate or predict × Stuck equals failure
Create one fresh exam-style task where the learner practises estimate or predict while watching specifically for stuck equals failure. Begin under a defined time limit. Predict sign, range or rough magnitude where possible. Require the student to state the task, commit to a strategy and mark the first uncertainty before checking.
The learner quits too early. After correction, change one meaningful feature—wording, values, representation, evidence, mark allocation, order or time pressure. Then use this related exercise: Predict sign and rough magnitude before calculating. Schedule a delayed retest so the next success cannot be explained only by immediate familiarity.
Workshop 14: Execute clearly × Memorized pattern matching
Create one fresh exam-style task where the learner practises execute clearly while watching specifically for memorized pattern matching. Begin under a defined time limit. Use consistent notation, logical steps and units. Require the student to state the task, commit to a strategy and mark the first uncertainty before checking.
The student recognizes layout rather than structure. After correction, change one meaningful feature—wording, values, representation, evidence, mark allocation, order or time pressure. Then use this related exercise: Compare a plausible wrong method with the correct one. Schedule a delayed retest so the next success cannot be explained only by immediate familiarity.
Workshop 15: Check reasonableness × No estimation
Create one fresh exam-style task where the learner practises check reasonableness while watching specifically for no estimation. Begin under a defined time limit. Ask whether the result fits the units, context and expected scale. Require the student to state the task, commit to a strategy and mark the first uncertainty before checking.
Calculator output is accepted automatically. After correction, change one meaningful feature—wording, values, representation, evidence, mark allocation, order or time pressure. Then use this related exercise: Leave a stalled item with a re-entry cue and return later. Schedule a delayed retest so the next success cannot be explained only by immediate familiarity.
Workshop 16: Analyse the first wrong step × Final-answer fixation
Create one fresh exam-style task where the learner practises analyse the first wrong step while watching specifically for final-answer fixation. Begin under a defined time limit. Trace errors to the earliest unsupported decision. Require the student to state the task, commit to a strategy and mark the first uncertainty before checking.
Only the last number is examined. After correction, change one meaningful feature—wording, values, representation, evidence, mark allocation, order or time pressure. Then use this related exercise: Classify ten mixed questions without solving them. Schedule a delayed retest so the next success cannot be explained only by immediate familiarity.
Workshop 17: Understand the problem × Solution peeking
Create one fresh exam-style task where the learner practises understand the problem while watching specifically for solution peeking. Begin under a defined time limit. Identify what is being asked, what information is given and which parts must be answered. Require the student to state the task, commit to a strategy and mark the first uncertainty before checking.
The answer appears before the learner has generated a route. After correction, change one meaningful feature—wording, values, representation, evidence, mark allocation, order or time pressure. Then use this related exercise: Predict sign and rough magnitude before calculating. Schedule a delayed retest so the next success cannot be explained only by immediate familiarity.
Workshop 18: Classify the structure × Messy working
Create one fresh exam-style task where the learner practises classify the structure while watching specifically for messy working. Begin under a defined time limit. Identify the concept or method family and distinguish it from nearby alternatives. Require the student to state the task, commit to a strategy and mark the first uncertainty before checking.
Errors become difficult to locate. After correction, change one meaningful feature—wording, values, representation, evidence, mark allocation, order or time pressure. Then use this related exercise: Compare a plausible wrong method with the correct one. Schedule a delayed retest so the next success cannot be explained only by immediate familiarity.
Workshop 19: Represent the information × Same-question correction only
Create one fresh exam-style task where the learner practises represent the information while watching specifically for same-question correction only. Begin under a defined time limit. Use a diagram, equation, table or labelled quantities when useful. Require the student to state the task, commit to a strategy and mark the first uncertainty before checking.
The original problem is redone repeatedly. After correction, change one meaningful feature—wording, values, representation, evidence, mark allocation, order or time pressure. Then use this related exercise: Leave a stalled item with a re-entry cue and return later. Schedule a delayed retest so the next success cannot be explained only by immediate familiarity.
Workshop 20: Attempt before checking × Blocked practice only
Create one fresh exam-style task where the learner practises attempt before checking while watching specifically for blocked practice only. Begin under a defined time limit. Delay answer-key access until a real attempt exists. Require the student to state the task, commit to a strategy and mark the first uncertainty before checking.
Every item uses one method. After correction, change one meaningful feature—wording, values, representation, evidence, mark allocation, order or time pressure. Then use this related exercise: Classify ten mixed questions without solving them. Schedule a delayed retest so the next success cannot be explained only by immediate familiarity.
Workshop 21: Estimate or predict × Stuck equals failure
Create one fresh exam-style task where the learner practises estimate or predict while watching specifically for stuck equals failure. Begin under a defined time limit. Predict sign, range or rough magnitude where possible. Require the student to state the task, commit to a strategy and mark the first uncertainty before checking.
The learner quits too early. After correction, change one meaningful feature—wording, values, representation, evidence, mark allocation, order or time pressure. Then use this related exercise: Predict sign and rough magnitude before calculating. Schedule a delayed retest so the next success cannot be explained only by immediate familiarity.