Secondary 3 Mathematics Tuition in Yishun | Choosing the Method When Questions Change

Teaching edition: 15 September 2026. The original archive address is retained; the learning material below has been rewritten for the scope stated here.

You can know a method and still fail to recognise the moment when it becomes useful. That is the gap this Secondary 3 Mathematics workshop is designed to make visible. Instead of giving you another tour through the syllabus, it places questions beside one another and asks a more demanding question: what makes this method appropriate here, and what would make it inappropriate?

Work with a pencil. Read one comparison, attempt the small task, and open its answer only after you have made a decision. The numbers, people, quotations, prices and school-event situations in the worked examples are invented for teaching. Leah, Harun and Wen are fictional learners, not accounts of actual pupils or promised tuition results.

Scope: the main audience is a Secondary 3 learner studying Mathematics at G3, including students who call it E-Math. Some tasks repair lower-secondary prerequisites; others connect ideas that a school may teach later. Follow your school’s actual sequence. This is not an Additional Mathematics syllabus guide, a complete national examination paper or a diagnostic assessment with validated cut-off scores.

For the relevant examination cohort, SEAB’s SEC overview explains the introduction of the Singapore-Cambridge Secondary Education Certificate from 2027. Its 2027 G3 syllabus list identifies Mathematics as K310. The K310 syllabus includes selecting relevant mathematics, moving between representations and connecting ideas. Those are the particular demands practised here. G1, G2 and G3 are not interchangeable teaching plans.

---

1. Three learners, three different problems

Leah opens a mixed worksheet and immediately looks for an equation to solve. Harun searches for a question that resembles yesterday’s example. Wen reads the question carefully but spends so long comparing possible approaches that he completes only half the set. All three might say the same thing afterwards: “I need more practice.” They do, but not necessarily the same practice.

Give Leah the expression 4(2x − 3) and ask her to simplify it. She writes 8x − 12 = 0 and obtains x = 1.5. Her algebra after the invented equality is orderly. The problem lies earlier: she changed the task. There was no equation to solve. A worksheet of further linear equations will not directly address the distinction she missed.

Harun can solve 3x + 8 = 29. He subtracts eight, divides by three and checks x = 7. When a question describes a hire fee of eight dollars plus three dollars per hour, with a total bill of twenty-nine dollars, he hesitates. The arithmetic is not harder. The relationship has arrived in prose rather than symbols. Harun needs to practise constructing the representation, not merely operating on one supplied to him.

Wen sees both structures correctly. In a reverse-percentage question, however, he tries a bar model, abandons it, writes an equation, changes to a multiplier, and then checks the same calculation three times. Every approach could work. His difficulty is deciding when he has enough reason to commit to one. He needs a stopping rule for method comparison and a purposeful check, not pressure to calculate carelessly.

What these cases do and do not establish

These invented cases illustrate a distinction. They do not establish that every student belongs to one permanent category. The same learner can select well in geometry and poorly in probability. A normally secure student can misread a particular sentence. A single blank response cannot tell you whether the cause was missing knowledge, unfamiliar language, fatigue or lack of time.

That is why this workshop asks for visible decisions. Beside a question, write a brief statement such as “The sale price is eighty per cent of the original,” or “Two sides of a right-angled triangle are known.” That sentence gives the learner and teacher something to inspect before the calculation begins. It is not an additional essay to write in every examination; it is a training device for revealing the reason behind a choice.

A useful decision sentence contains a relationship, not only a topic name. “Percentages” is too broad. “The question gives the result after a reduction, so I need the original base” explains the direction. “Trigonometry” is too broad. “The known side is adjacent to the given angle and I need the opposite side” points towards a particular ratio.

Method knowledge has a boundary

Knowing a method includes knowing when its conditions are absent. Pythagoras is powerful, but not a licence to square the sides of every triangle. Direct proportion is useful, but a fixed starting fee prevents the total cost from being directly proportional to quantity. The mean of two group means is not automatically the mean of all observations. A cancelled denominator does not erase an excluded input.

We will repeatedly place a valid method beside a tempting invalid one. The aim is not to make students suspicious of everything. It is to build a small set of decisive checks: identify the object, name the relationship, retain the conditions, and return the result to the question.

Do not attempt every chapter at once. Start with the decision check, then select two comparison families that expose a real uncertainty. A learner who already distinguishes percentage bases should move forward. A learner who cannot yet manipulate fractions should slow down there. The route is determined by work on the page, not by a requirement to complete the longest reading journey.

The central question throughout is practical: after this explanation, can you make a better choice on a fresh question without someone pointing at the formula? The following chapters are built to let you test that question rather than simply agree with it.

2. The starting decision check

Before reading the worked comparisons, try these eight short tasks. For each, write a proposed first relationship and then solve only as far as you can without notes. Do not use the chapter names in the contents as clues. The complete discussions appear in Chapter 27, away from the starting questions.

You may record “not yet taught” when a topic is genuinely unfamiliar. That is different from knowing a method but not recognising it. You may also write a precise question about missing information. Not every mathematical prompt deserves an immediate numerical answer.

Decision check 1: the reduced price

A model-making kit costs $68 after a 15% discount on its original price. Find the original price. Before calculating, complete this sentence: “Sixty-eight dollars represents ___ per cent of ___.” Do not fill the second blank with “the answer”; identify the actual quantity.

Decision check 2: the service rule

A service uses a fixed booking fee plus a constant charge per hour. A two-hour booking costs $17 and a five-hour booking costs $29. Find the cost of an eight-hour booking. Explain why multiplying $17 by four would not automatically give the correct answer.

Decision check 3: the mathematical object

First simplify 3(2x − 5) + 4. Then, as a separate task, solve 3(2x − 5) + 4 = 19. Your two responses should not be the same type of object. Identify the distinction before working.

Decision check 4: two conditions

A box contains only two-dollar and five-dollar tokens. There are fourteen tokens altogether and their total face value is $46. Find how many tokens of each type are in the box. Define each unknown so that its unit is clear.

Decision check 5: the sloping line

A straight line passes through A(−2, 9) and B(4, −3). Find its gradient. Predict its sign before calculating. Explain whether 9 divided by −2 is sufficient to find the gradient of this line.

Decision check 6: a larger similar panel

Two panels are similar. A length of 6 cm on the smaller panel corresponds to a length of 9 cm on the larger. The smaller panel has area 40 cm². Find the larger area. State whether the multiplier applies to length or area.

Decision check 7: a right-angled frame

A triangular frame is right-angled. Its two perpendicular sides measure 9 cm and 12 cm. Find the longest side. State the information that permits your chosen theorem, rather than merely naming the theorem.

Decision check 8: drawing counters

A bag contains four red counters and two blue counters, all equally likely to be selected. Two counters are drawn one after another without replacement. Find the probability that both are blue. Describe what changes between the first and second draw.

Read your work without turning it into a label

Do not total these into a predicted examination grade. Eight questions sample a few relationships; they do not cover a syllabus, represent every difficulty or establish a stable learner profile. The useful evidence is narrower. Did you identify the relevant base? Did you preserve the difference between an expression and an equation? Did you notice the changing sample space?

For a question you answered incorrectly, cover the arithmetic and inspect the first relationship. A correct equation followed by an arithmetic slip calls for a different response from an incorrect equation solved accurately. A blank first line calls for a small probe: can you explain the situation in ordinary language, or can you solve a simpler version with the relationship already supplied?

For a question you answered correctly, ask how the choice was made. “I recognised that the difference in bills removes the fixed fee” is useful evidence. “It looked like the last question I did” may indicate a cue worth testing with a variation. A correct answer is welcome, but it does not make the explanation irrelevant.

Keep the original attempt. Do not erase it into a neat corrected page. Use a different section of your notebook for the repair, and record one later attempt. The comparison between the first decision and the later decision will tell you more than a page on which every trace of uncertainty has disappeared.

A parent working alongside a learner should avoid supplying the topic name too soon. Ask what the quantities represent. Ask which condition matters. When direct teaching becomes necessary, give it plainly and record that support was supplied. Supported success is useful; it simply answers a different question from independent success.

3. Read targets, quantities and conditions

A word problem contains more than numbers waiting to be collected. It contains objects, relationships, restrictions and a requested result. A productive first read separates those components. This is not a ritual of highlighting every sentence. It is a way to prevent an attractive calculation from answering the wrong question.

Consider this invented prompt: a rectangular noticeboard has width w metres. Its length is 0.8 m greater than twice its width. The perimeter is 8.8 m. Find the area. Three quantities are already in play: width, length and perimeter. The target is a fourth quantity, area. Solving for w will be an intermediate achievement, not the final answer.

Write length = 2w + 0.8. The perimeter condition becomes 2[w + (2w + 0.8)] = 8.8. Simplifying gives 6w + 1.6 = 8.8, so w = 1.2. The length is 3.2 m. Only now calculate the requested area: 1.2 × 3.2 = 3.84 m². Check the perimeter independently: 2(1.2 + 3.2) = 8.8 m.

Target words change the stopping point

Change only the final sentence to “Find the width,” and the same early working ends sooner. Change it to “Find the amount of trim needed to go around the board once,” and the perimeter is already supplied. Change it to “Find the cost of covering the board at $14 per square metre,” and the area must be multiplied by fourteen, giving $53.76 under that stated pricing rule.

The shape has not changed. The appropriate stopping point has. Students who rely on a remembered sequence for “rectangle questions” can do unnecessary work or stop one stage too early. The target determines which intermediate values are worth finding.

Conditions tell you what you may assume

Suppose the prompt merely says a board has perimeter 8.8 m. Its area cannot be uniquely determined from that information alone. A rectangle measuring 1.2 m by 3.2 m has that perimeter, but so does one measuring 2 m by 2.4 m. Their areas are 3.84 m² and 4.8 m². The length-width relationship was not decorative wording; it removed the uncertainty.

In geometry, a condition might be a right angle, equal sides or parallel lines. In probability, it might be replacement or equal likelihood. In a cost problem, it might be a constant rate, a fixed fee or a rounding rule. In a sequence, it might be the statement that the pattern continues with a constant difference. Remove the condition and the method may lose its justification.

Units organise the unknowns

Let w stand for a width measured in metres, not for “the board”. Let n stand for the number of tickets, not for “ticket money”. Let C stand for a cost measured in dollars. A variable definition should distinguish a count from a rate and a rate from a total.

For example, if each ticket costs $6, then 6n is a total cost in dollars when n is a ticket count. If n instead represents dollars, writing 6n without reconsidering its meaning would describe a different quantity. Algebra can be syntactically tidy while the model is semantically wrong.

A small annotation that earns its space

Beside the board question, write “Target: area; condition: perimeter and length-width rule; unknown: width in metres.” That note is enough. It need not become a coloured page of restated prose. Its value lies in holding the target and conditions steady while the calculation proceeds.

Now try this related task. A rectangular mat has length 1.5 m more than its width. Its perimeter is 13 m. Find its area. Before solving, say why adding 1.5 directly to the perimeter would be meaningless. The 1.5 describes a difference between two lengths, not an extra boundary attached to the whole perimeter.

Mat task: method and answer

Let the width be w metres, so the length is w + 1.5. Then 2[w + (w + 1.5)] = 13, giving 4w + 3 = 13 and w = 2.5. The length is 4 m and the area is 10 m². Checking 2(2.5 + 4) = 13 confirms the boundary condition. The area uses the product of the two dimensions; it is not the perimeter multiplied by an arbitrary number.

Reading mathematically is therefore selective. Keep information that constrains the model. Keep the requested quantity. Keep the unit and any domain restriction. Other story details may help the situation feel concrete, but they should not be promoted into operations merely because they appear in the paragraph.

4. Expression, equation, identity or inequality?

Before choosing an algebraic method, identify the object you have been given. An expression represents a value. An equation asserts equality, usually to be satisfied by particular inputs. An identity asserts equality for every input in the permitted domain. An inequality compares values without asserting equality. Similar symbols can therefore ask for very different kinds of work.

Take 2(3x − 4) + x. Simplifying gives 7x − 8. There is no numerical value of x to find because no condition has been stated that selects one. At x = 2 the expression equals 6; at x = 5 it equals 27. Simplification produces a more convenient expression for all allowed inputs, not a single answer for the variable.

Now take 2(3x − 4) + x = 27. The same simplification gives 7x − 8 = 27. Adding eight and dividing by seven gives x = 5. The equality selects a value. Substituting five into the original left side gives 2(15 − 4) + 5 = 27, so the solution satisfies the stated condition.

An identity is not waiting for one special value

The statement 2(3x − 4) + x = 7x − 8 is true for every real x. It records equivalent expressions. Testing x = 5 is consistent with that identity but does not establish it for all values. Expanding and collecting like terms establishes the relationship generally because the operations hold throughout the domain.

This distinction matters when checking. One counterexample can disprove a proposed identity. If somebody claims 2(x + 3) = 2x + 3 for every x, using x = 0 gives 6 = 3, which is false. But a handful of matching substitutions cannot replace a general algebraic justification when the question asks for one.

An inequality describes a set of possibilities

For 7x − 8 < 27, the result is x < 5, not x = 5. A learner who changes the inequality symbol into an equality has changed the set of acceptable inputs. At x = 4 the left side is 20, which is less than 27; at x = 5 it equals 27 and fails the strict inequality.

Multiplying or dividing an inequality by a negative number reverses its direction. For −2x > 10, division by −2 gives x < −5. Check with x = −6: the original left side is 12, which is greater than ten. Check with x = 0: zero is not greater than ten. These tests help inspect the direction, while the algebra explains the full solution set.

Keep the task verb attached to the object

“Expand” asks for a distributive rewrite. “Factorise” asks for a multiplicative rewrite. “Solve” asks for values satisfying a relation. “Evaluate” asks for a value, often after substituting given inputs. “Show” or “prove” asks for a justified chain leading to the stated result. These verbs are not interchangeable decorations above the same calculation.

For example, factorising x² − 9 gives (x − 3)(x + 3). Solving x² − 9 = 0 gives x = 3 or x = −3. Evaluating x² − 9 at x = 2 gives −5. A single expression can participate in all three tasks, but the output must answer the particular instruction.

Ask Leah from our opening case to sort six prompts before solving any of them. Put simplification and evaluation in separate piles. Put equations and inequalities in separate piles. Ask her to describe the expected output: another expression, a number, a set of values or a general argument. This small classification task directly addresses her habit of inventing an equation whenever she sees x.

Try a four-way comparison

Use the expression 3(x − 2) + 2x. First simplify it. Next evaluate it at x = −1. Then solve the equation obtained by setting it equal to fourteen. Finally solve the strict inequality obtained by asking when it is less than fourteen. Explain why the final two answers should differ.

Four-way comparison: answers

The simplified expression is 5x − 6. At x = −1 its value is −11. Setting it equal to fourteen gives 5x − 6 = 14, so x = 4. Requiring it to be less than fourteen gives x < 4. Four is the boundary but is not included in the strict inequality. The identity 3(x − 2) + 2x = 5x − 6 holds for every real x; it is not the solution x = 4.

A good first algebraic decision is often this unglamorous distinction. Once the object and required output are clear, the familiar procedures have a sensible destination. Without that distinction, accurate manipulation can take you efficiently to an answer nobody asked for.

5. Percentage: first identify the base

The most important word in a percentage question is often the word “of”, even when the question leaves it implied. Twenty per cent of the original amount and twenty per cent of the reduced amount are different quantities. Before choosing multiplication or division, identify the base to which the percentage refers.

A bag is priced at $90 and discounted by 20%. The discount is 0.20 × 90 = $18, so the sale price is $72. The direct multiplier 0.80 × 90 produces the same sale price. Both methods are valid because the original amount is known and the percentage acts on that amount.

Now reverse the information. A bag costs $72 after a 20% discount. Let its original price be P dollars. The relationship is 0.80P = 72, giving P = 90. Adding twenty per cent of seventy-two gives $86.40, not ninety dollars. That tempting operation uses the sale price as its base rather than reversing the original reduction.

The two questions look almost identical

Each question mentions a bag, a price and a twenty per cent discount. Those words do not determine the direction of calculation. The decisive difference is which quantity is known. In the direct question, the original is supplied. In the reverse question, the result of the percentage operation is supplied.

A useful working sentence is “The final amount is eighty per cent of the original.” It remains true in both questions. The same relationship therefore supports both directions. You multiply when the original is known and divide when recovering it from the final amount.

Successive changes require successive bases

A fictional subscription price increases by 10% and then receives a 10% reduction. Starting from $100, the increase gives $110. The reduction is now ten per cent of one hundred and ten, or eleven dollars, leaving $99. The two percentage changes do not cancel because their bases differ.

The multiplier form records this cleanly: 100 × 1.10 × 0.90 = 99. The product of the multipliers is 0.99, so the net change is a 1% decrease. No new rule is needed. Follow the quantity as it changes.

The order of two pure percentage multipliers produces the same final value because multiplication commutes. However, a fixed-dollar change and a percentage change need not commute. Starting at $100, subtracting ten dollars and then discounting by 20% gives $72. Discounting first and then subtracting ten dollars gives $70. The distinction is between multiplication and subtraction, not merely the order in which promotional words appear.

Percentage points are not percentage change

Suppose an invented survey reports an increase in support from 40% to 50%. The increase is ten percentage points. Relative to the original forty per cent, the percentage increase is 10/40 × 100% = 25%. The subtraction describes the gap on a percentage scale; the division describes the relative change.

State which comparison the question asks for. Writing “ten per cent increase” is ambiguous and, for a relative percentage-change question, incorrect here. Clear language protects the mathematics because it keeps the reference quantity visible.

A base can be hidden in a story

A school-event committee prepares 240 tickets. It sells 75% of them and refunds 10% of the sold tickets. How many tickets remain sold after refunds? First identify the sold quantity: 0.75 × 240 = 180. The refund is ten per cent of those 180 tickets, not ten per cent of all 240. Eighteen are refunded, leaving 162 sold tickets.

You can combine the stages as 240 × 0.75 × 0.90 = 162. The expression is useful because its factors preserve the sequence of meanings: prepared to initially sold to still sold. Writing 240 × (0.75 − 0.10) would wrongly subtract a percentage of the original stock.

Choose a method that exposes the relationship

For a learner who is uncertain about the base, an equation or a labelled bar can be more informative than a memorised multiplier. For a learner who understands the base securely, a multiplier is efficient. These are not rival camps. They are representations of the same relationship, chosen for clarity and control.

Do not demand a long algebraic solution for a question that can be answered transparently in one line. Conversely, do not insist on mental shortcuts when the learner keeps reversing the base. The goal is a justified method that the learner can inspect and adapt.

Three fresh decisions

A: A $150 fee is reduced by 12%. Find the reduced fee. B: A fee is $132 after a 12% reduction. Find the original fee. C: An amount is reduced by 12%, then increased by 12%. Is it back to its original value? Identify the base at each stage before calculating.

Percentage comparison: answers and checks

A: 150 × 0.88 = $132. B: 132 ÷ 0.88 = $150. C: the combined multiplier is 0.88 × 1.12 = 0.9856, so the final amount is 98.56% of the original, a 1.44% decrease. Substituting an easy original amount such as one hundred helps explain the changing bases, but the multiplier establishes the conclusion for any positive starting amount.

When reviewing a percentage error, do not ask only whether the calculator was used correctly. Ask what the learner believed represented one hundred per cent. That answer often reveals the first decision that needs attention.

6. Proportion is not every straight-line relationship

A quantity is directly proportional to another when their ratio remains constant over the domain being modelled. A fixed starting amount changes that structure. Both relationships can produce straight lines, but only the directly proportional one passes through the origin when extended to zero within the model.

Imagine a printing service charging $0.40 per sheet with no setup fee and no minimum charge. For n sheets, C = 0.40n. Twenty sheets cost eight dollars; fifty sheets cost twenty dollars. Doubling n doubles C. Dividing any positive cost by its corresponding sheet count gives the same unit rate, $0.40 per sheet.

A second service charges a four-dollar setup fee plus $0.40 per sheet. Its rule is C = 4 + 0.40n. Twenty sheets cost twelve dollars and fifty sheets cost twenty-four dollars. Doubling the number of sheets does not double the total cost because the setup fee is not repeated with every sheet.

The difference between a rate and an average cost

For the second service, twenty sheets cost twelve dollars, so the average cost per sheet is $0.60. That does not mean each additional sheet costs sixty cents. The additional charge is still forty cents. The average includes a share of the fixed setup fee.

This distinction is easy to miss when the question supplies only one bill. From a total of twelve dollars for twenty sheets, you cannot determine whether the rule is C = 0.60n, C = 4 + 0.40n, or another rule entirely unless further assumptions are supplied. One observation does not identify a pricing mechanism.

Two bills can reveal a constant additional rate

Suppose the question explicitly states that a service charges a fixed fee plus a constant amount per hour. A two-hour booking costs seventeen dollars and a five-hour booking costs twenty-nine dollars. The difference is twelve dollars for three extra hours, so the additional rate is four dollars per hour.

The fixed fee is 17 − 2 × 4 = nine dollars. The rule is C = 9 + 4h. An eight-hour booking costs forty-one dollars. The subtraction of the two bills works because the same fixed fee appears in both and cancels. Without the constant-rate assumption, two bills would not justify extending a straight line to every other booking length.

Choose between total ratio and difference ratio

For direct proportion, C/n is constant. For a fixed-plus-variable relationship, the ratio of changes, ΔC/Δn, is constant. Those two ratios answer different questions. The first compares totals from zero. The second compares how much one total changes when the other changes.

This provides a bridge to gradient. You do not need to treat a pricing question and a graph question as unrelated chapters. A fixed fee is an intercept; a constant additional charge is a gradient; two bills are two points. The model is the common structure beneath the representations.

When “more” does not mean a constant rate

A third hypothetical service charges forty cents per sheet for the first fifty sheets and thirty cents for every sheet beyond fifty, with no setup fee. This is not one constant-rate rule over the entire domain. Seventy sheets cost 50 × 0.40 + 20 × 0.30 = twenty-six dollars. Multiplying seventy by thirty cents applies the discounted rate to sheets that do not qualify.

The key condition is the threshold. A table or a split calculation makes it visible. The graph would have different gradients in different sections. The question still concerns printing, but the mathematical relationship has changed again.

Try the decision without calculating everything

A club pays a fixed $18 hire fee and $3 for each chair. Another pays $4.50 per chair with no fixed fee. Before solving, predict which is cheaper for a very small order and which may become cheaper for a large order. Then find the order size at which the costs are equal.

Chair-hire comparison: answer

Let n be the chair count. Compare 18 + 3n with 4.5n. Equality gives 18 = 1.5n, so n = 12. At twelve chairs both cost $54. Below twelve chairs the no-fixed-fee option is cheaper; above twelve the lower per-chair rate outweighs the fixed fee. Because chairs are counted in whole numbers, practical orders use non-negative integers. This is an algebraic model using invented charges, not a quotation from a real supplier.

The transferable lesson is not “always subtract bills”. It is “identify what stays the same and what changes”. Subtracting is useful when it removes a shared fixed component. Dividing totals is useful when direct proportion is justified. Choosing a method means recognising that distinction before arithmetic begins.

7. Two unknowns and enough information

Two unknowns do not automatically mean that a problem can be solved by writing any two equations. The equations must represent the stated relationships, and they must provide enough independent information. A repeated version of the same condition does not add a new restriction.

Consider the token box from the starting check. Let x be the number of two-dollar tokens and y the number of five-dollar tokens. The count condition is x + y = 14. The value condition is 2x + 5y = 46. Notice the units: the first equation counts tokens; the second counts dollars of face value. The same unknowns participate in two different relationships.

Multiply the count equation by two to obtain 2x + 2y = 28. Subtract it from the value equation: 3y = 18. Thus y = 6 and x = 8. Check both conditions: eight plus six is fourteen tokens, and 2 × 8 + 5 × 6 = 46 dollars. Checking only the total value would leave the count condition untested.

Substitution and elimination answer the same model

You could instead write x = 14 − y and substitute into the value equation. That gives 2(14 − y) + 5y = 46, then 28 + 3y = 46, so y = 6. Neither route is inherently more intelligent. Elimination is short here because matching the coefficients of x is easy. Substitution is transparent because one equation already isolates a simple total.

Choose a route that preserves meaning and reduces unnecessary manipulation. A student who consistently loses signs when subtracting equations may initially find substitution easier to inspect. That does not remove the need to learn elimination; it simply recognises that method choice also depends on the structure of the particular equations and the learner’s present control.

One condition can leave many answers

Suppose the prompt tells you only that the box contains fourteen tokens of the two types. Then x + y = 14 has many non-negative integer solutions: x = 0, y = 14; x = 1, y = 13; and so on. Inventing a total value in order to obtain one answer would be a modelling error.

Suppose the prompt supplies x + y = 14 and 2x + 2y = 28. These are two written equations but only one independent condition. The second merely doubles the first. It does not tell you how the fourteen tokens are divided between the types.

Now suppose it supplies x + y = 14 and 2x + 2y = 30. The conditions conflict: doubling the first gives twenty-eight, not thirty. There is no solution satisfying both. In a real data task, that would be a reason to inspect the data or assumptions, not to force an answer through an inappropriate procedure.

The domain can reject an algebraic result

A count of tokens must be a non-negative integer. An algebraic solution such as y = 5.5 would therefore require attention. It might reveal an error in the student’s equations, an arithmetic mistake or an impossible set of supplied conditions. The correct response is not to round 5.5 tokens to six and continue as though the original conditions still hold.

For example, keep fourteen tokens but change the total face value to forty-five dollars. The equations give 3y = 17 and y = 17/3. No whole-number mixture of two-dollar and five-dollar tokens satisfies those exact conditions. This is a meaningful conclusion, not an unfinished calculation.

A second context, the same structure

A school exhibition sells student tickets for four dollars and adult tickets for seven dollars. It sells thirty tickets and collects $156. Let s and a be the respective ticket counts. Before solving, name the two conditions and predict whether your final values must be whole numbers.

Exhibition tickets: two methods and a check

The equations are s + a = 30 and 4s + 7a = 156. Subtract four times the first equation from the second: 3a = 36, so a = 12 and s = 18. Alternatively substitute s = 30 − a into the money equation. The counts total thirty and the money is 18 × 4 + 12 × 7 = $156. Both values are non-negative integers, as required.

Harun might recognise this as a “ticket question”. Ask him to go one level deeper: it is a count-and-value system. The token box and exhibition use different objects, but they share the structure. A later problem about two types of packages could use the same reasoning. Conversely, a ticket question with a fixed booking charge may require a different model. The noun does not choose the method.

8. Expand, factorise or solve?

An algebraic form is useful because of what it makes visible. Expanded form can make coefficients and like terms easy to compare. Factorised form can expose zeros and common factors. An equation may require either form on the way to finding its solutions. The method should follow the target, not a habit of expanding every bracket immediately.

Take (x − 4)(x + 2). Expanding gives x² − 2x − 8. Both forms represent the same value for every real x. If the question asks for the coefficient of x, expanded form displays −2 directly. If it asks when the expression equals zero, factorised form already provides a useful starting point.

For (x − 4)(x + 2) = 0, at least one factor must be zero. Hence x = 4 or x = −2. This is the zero-product property. It does not say that the factors of every product must separately equal the value on the other side of an equation.

A small change invalidates a tempting shortcut

Change the equation to (x − 4)(x + 2) = 7. You cannot set x − 4 = 7 or x + 2 = 7 and claim those values solve the product equation. A product can equal seven through many pairs of non-zero factors. The zero-product shortcut was special to a product equal to zero.

Here expand and rearrange: x² − 2x − 8 = 7, so x² − 2x − 15 = 0. Factorising gives (x − 5)(x + 3) = 0, hence x = 5 or x = −3. Check: at five the original factors are one and seven; at negative three they are negative seven and negative one. Both products equal seven.

The decisive step was not the act of factorising. It was rewriting the equation so that a product equalled zero. Teach that condition explicitly. Otherwise a learner may remember the surface instruction “split the brackets” and apply it where it is invalid.

Do not divide away a possible solution

Consider x² = 5x. Dividing both sides by x gives x = 5, but that division assumes x is not zero. Zero actually satisfies the original equation. A safer complete route is x² − 5x = 0, then x(x − 5) = 0, giving x = 0 or x = 5.

The missing solution is not a calculator problem. It is a decision about a transformation. Before dividing by an expression, consider whether that expression could be zero. If it can, either treat that case separately or choose a route that retains it.

Factoring is not guessing without a check

To factorise x² + x − 12, seek two numbers with product −12 and sum 1. Four and negative three fit, so the factorisation is (x + 4)(x − 3). Expanding the proposed factors verifies the identity. It is possible to find the correct product but the wrong sum, especially when signs are involved.

For 2x² + 7x + 3, one factorisation is (2x + 1)(x + 3). Expanding gives 2x² + 6x + x + 3 = 2x² + 7x + 3. The middle coefficient comes from two cross-products. A learner who checks only the first and last terms can accept a false factorisation.

Not every quadratic has convenient integer factors

Do not spend unlimited time searching for neat factors that do not exist. For x² − 2x − 1 = 0, the quadratic formula gives x = 1 ± √2. A numerical approximation may be appropriate if the question requests it, but the exact form records the solutions without rounding.

Method selection here includes noticing what the task permits and what form is requested. Factorisation is efficient when factors are readily available. A general formula provides another route when they are not. An examination instruction such as “by factorisation” or a particular curriculum task can constrain the allowed approach; do not disregard it simply because another method is familiar.

Try a pair designed to expose the condition

A: Solve x(x − 6) = 0. B: Solve x(x − 6) = 16. Before calculating, explain why the first equation can use the zero-product property immediately but the second cannot.

Product equations: answers

A gives x = 0 or x = 6. For B, rearrange to x² − 6x − 16 = 0, then factorise as (x − 8)(x + 2) = 0. Thus x = 8 or x = −2. Substitution into the original product gives sixteen in both cases. Setting either original factor equal to sixteen would not be a valid use of the zero-product property.

When you review this family, record the trigger as “make one side zero before using a product of factors”. That is a transferable condition. “Use the brackets trick” is not.

9. Fractions, cancellation and restrictions

Cancellation is division by a common non-zero factor. It is not the removal of similar-looking symbols from anywhere in a fraction. A student who can state that distinction has a much better chance of spotting when a familiar-looking shortcut changes the expression.

For 6x/9, divide numerator and denominator by three to obtain 2x/3. For 6x/(9x), cancellation of x gives 2/3 only when x is not zero. The original fraction is undefined at zero, even though the simplified constant 2/3 is perfectly defined there. Simplification must carry the restriction forward.

Compare (x + 4)/x. You cannot cancel the x in the numerator against the denominator because the numerator is a sum, not a product with x as a factor. For x ≠ 0, you can instead split the fraction as x/x + 4/x = 1 + 4/x. Substituting x = 2 gives three in both forms. The invalid cancellation would miss the remaining division.

Factor before deciding whether cancellation is legal

Consider (x² − 16)/(x − 4). The numerator factorises as (x − 4)(x + 4). For x ≠ 4, the common factor x − 4 can be cancelled, leaving x + 4. At x = 4, however, the original expression is undefined because its denominator is zero. The simplified rule does not retroactively make the original expression defined there.

This matters when an equation includes the fraction. For (x² − 16)/(x − 4) = 8, with x ≠ 4, simplification gives x + 4 = 8 and the candidate x = 4. That candidate is excluded. The original equation has no solution. Writing x = 4 without checking the restriction would manufacture an answer at a point where the original equation has no meaning.

Clearing denominators must affect every term

Solve x/3 + 1/2 = 5/6. Multiplying the entire equation by six gives 2x + 3 = 5, so x = 1. A common error is to multiply only one term by six. The operation applies to both sides and every term on each side.

Brackets help keep that intention visible: 6(x/3 + 1/2) = 6(5/6). After distribution, each denominator is cleared. The working is slightly longer than a compressed shortcut, but it displays the equivalence clearly while the skill is still developing.

Now solve 3/(x − 1) = 2. State x ≠ 1 first. Multiplying by x − 1 gives 3 = 2(x − 1), so x = 2.5. Substitution gives 3/1.5 = 2. The restriction survives the transformation and the candidate satisfies it.

Cross-multiplication has an underlying operation

For a/b = c/d, with b and d non-zero, multiplying both sides by bd gives ad = bc. The cross-products are not a magic pattern. They are the result of clearing denominators. Understanding that derivation helps prevent students from cross-multiplying in expressions that are not equations or from ignoring denominator restrictions.

For example, 2/(x + 1) + 3 is not an equality to which cross-multiplication can be applied. It is an expression. To combine its terms, write 3 as 3(x + 1)/(x + 1), preserving x ≠ −1, and obtain (3x + 5)/(x + 1). The desired operation depends on the task: combine, simplify, evaluate or solve.

A numerical check can expose, but not fully prove, a simplification

If somebody proposes (x + 4)/x = 1 + 4 for every non-zero x, substitute x = 2. The left side is three and the right side is five. One counterexample disproves the claimed identity. If a proposed simplification matches at x = 1, that single match does not prove the identity. Choose tests that are informative, then use algebra for a general justification.

Also avoid testing at a forbidden input. A calculation that divides by zero is not evidence for or against a numerical identity in the usual sense; it is outside the expression’s domain. Keep “false” and “undefined” distinct.

Try two related tasks

A: Simplify (x² − 9)/(x − 3), stating the restriction. B: Solve (x² − 9)/(x − 3) = 10. Explain why a restriction belongs in both tasks even though only one asks for a numerical solution.

Fraction tasks: answers

A: factorise the numerator to obtain (x − 3)(x + 3), then simplify to x + 3 for x ≠ 3. B: the same restricted simplification gives x + 3 = 10, so x = 7. Seven is permitted and substitution into the original equation gives (49 − 9)/(7 − 3) = 40/4 = 10. The restriction records where the original denominator is non-zero.

The decision habit is compact: identify whether the numerator contains a common factor, check that the factor is non-zero, and retain the original restrictions. It protects both the manipulation and the meaning of the result.

For a deeper explanation of why some operations lose possible originals, continue to eduKate Sengkang’s How Mathematical Operations Can Preserve or Lose Information. Use it as a conceptual extension after these cancellation tasks, then return here for the mixed practice.

10. Gradient is a comparison of changes

The gradient of a straight line compares a change in the vertical quantity with the corresponding change in the horizontal quantity. It is not generally obtained by dividing one point’s y-coordinate by its x-coordinate. That latter ratio gives the slope from the origin to the point, which is a different line unless the given line passes through the origin.

Take A(−2, 9) and B(4, −3). From A to B, x increases by six and y decreases by twelve. The gradient is −12/6 = −2. The negative sign agrees with a line that falls as x increases. This prediction does not replace the calculation, but it provides an independent check on its direction.

Dividing 9 by −2 gives −4.5. Dividing −3 by 4 gives −0.75. Neither is the gradient of the line through A and B. Those ratios compare each point with the origin rather than with the other point.

Keep the point order consistent

You may calculate from A to B or from B to A. From B to A, the vertical change is 9 − (−3) = 12 and the horizontal change is −2 − 4 = −6, again giving −2. Reversing both changes preserves the ratio. Reversing only one changes the sign incorrectly.

A reliable written form is (y of B − y of A)/(x of B − x of A). The labels make the pairing explicit. Once the student is secure, the notation can be abbreviated. During repair, however, a small amount of explicit structure is often more useful than a rapid formula written without attention to the points.

Read the rate from a table

An invented filling process is modelled by this table:

Volume under a stated constant filling rule

Time in minutes	Volume in litres
0	12
3	24
7	40

The volume increases by twelve litres in three minutes, then by sixteen litres in four minutes. Both intervals give four litres per minute. The starting volume is twelve litres, so the linear model is V = 12 + 4t. Dividing the seven-minute volume by seven would mix the initial volume into the rate and give the wrong filling rate.

The phrase “under a stated constant filling rule” matters. A few measurements compatible with a line do not prove that a real process is exactly linear outside the observed interval. In an examination model, use the assumptions supplied. In an investigation, distinguish a fitted model from a guaranteed mechanism.

A scale changes what one grid square means

On a graph, suppose one horizontal square represents two minutes and one vertical square represents five litres. A line rises six vertical squares while moving four horizontal squares. The vertical change is thirty litres and the horizontal change is eight minutes, giving 3.75 litres per minute. Using 6/4 = 1.5 would calculate squares per square, not the requested physical rate.

Read units and intervals before counting squares. A graph can look steep because of the chosen scales. Visual steepness on the page is not by itself a numerical rate. The quantity represented by each axis must be included.

Horizontal and vertical are not the same special case

A horizontal line has zero change in y while x changes, so its gradient is zero. A vertical line has zero change in x; the usual gradient ratio would require division by zero and is undefined. Saying that a vertical line has gradient zero confuses “no horizontal change” with “no vertical change”.

For a vertical line x = 3, y may take many values while x stays fixed. It cannot be written as y = mx + c with a finite constant m describing all its points. Recognising this exception prevents a student from applying a familiar straight-line form outside its conditions.

Try the graph decision in another form

A line passes through (2, 11) and (8, 29). Find the gradient and the y-intercept. Then explain the meaning of the gradient if the horizontal quantity is hours and the vertical quantity is kilometres travelled after an initial recorded distance.

Line task: answer and interpretation

The gradient is (29 − 11)/(8 − 2) = 18/6 = 3. In y = 3x + c, substituting (2, 11) gives c = 5. The equation is y = 3x + 5. In the stated distance-time interpretation, the rate of increase is three kilometres per hour and the initial recorded distance is five kilometres. This rate interpretation depends on what the axes represent; a different pair of units would give a different meaning to the same numerical gradient.

Gradient questions and fixed-fee questions belong to the same family of decisions: compare changes to identify the rate, then account separately for the starting value.

11. Choosing a useful quadratic form

A quadratic relationship can be written in different equivalent forms, and each form reveals something different. Choosing a form is a mathematical decision. It is not merely a preference for whichever expression looks shorter.

Consider y = x² − 6x + 5. Its expanded form shows the coefficients. The factorised form y = (x − 1)(x − 5) reveals the x-intercepts, because y is zero when either factor is zero. The form y = (x − 3)² − 4 reveals the minimum value over all real x, because a square is non-negative.

These forms are equivalent. Expanding the brackets verifies the equality. You do not need to treat them as three separate formulas to memorise. They are three ways of seeing the same relationship.

The target determines the useful representation

If the question asks where the graph crosses the x-axis, factorised form makes x = 1 and x = 5 visible. If it asks for the y-intercept, set x = 0 in any form to obtain y = 5. If it asks for the smallest possible value for real x, the squared form makes y ≥ −4 clear, with equality at x = 3.

A learner can obtain correct results using other valid methods, but the point of comparing forms is to reduce unnecessary work. Expanding a factorised expression before finding its zeros may hide information that was already conveniently displayed. Conversely, multiplying out can help when comparing coefficients or combining the expression with another polynomial.

Keep the domain in view

The minimum −4 is reached at x = 3 when all real inputs are allowed. Suppose instead that the problem restricts x to 0 ≤ x ≤ 2. The point x = 3 is outside the permitted interval. Over that interval, the smallest value occurs at x = 2 and is −3. The unrestricted turning point does not automatically answer a constrained question.

You can see this without calculus. For 0 ≤ x ≤ 2, the quantity x − 3 lies between −3 and −1. Its square is smallest at −1, corresponding to x = 2. Subtracting four gives −3. The domain condition changes which part of the graph matters.

A geometric model supplies another quadratic

A rectangular display has perimeter twenty metres. Let one side be x metres, so the other is 10 − x. The area is A = x(10 − x) = 10x − x². For a genuine rectangle, 0 < x < 10. The area can also be written A = 25 − (x − 5)².

The largest area is twenty-five square metres, achieved when x = 5 and both sides are five metres. The conclusion follows because the squared term cannot be negative. This is a useful extension in choosing forms; the route a school requires may depend on what has already been taught. The important idea is that an equivalent rewrite can expose a bound.

Notice how different this is from a problem asking for a missing side from a known area. With A = 21, the same model gives x(10 − x) = 21, or x² − 10x + 21 = 0. Factorisation gives x = 3 or x = 7. These describe the same three-by-seven rectangle with the side labels exchanged. The target determines whether you seek an optimum or particular solutions.

Do not extend linear habits into a curved relationship

For a linear rule, equal increases in x produce equal increases in y. For y = x², the values at x = 1, 2, 3 and 4 are 1, 4, 9 and 16. The successive increases are 3, 5 and 7, not constant. Doubling x from two to four multiplies y by four, not by two.

A quadratic table can therefore look orderly without being directly proportional or linear. Before using a constant rate from two values, check whether the question actually states or supports a linear relationship.

Try selecting the form

The same quadratic is supplied as y = x² + 2x − 8, y = (x + 4)(x − 2), and y = (x + 1)² − 9. Which form makes the zeros easiest to identify? Which makes the unrestricted minimum easiest? Which value is obtained when x = 0?

Quadratic forms: answers

The factorised form gives zeros at x = −4 and x = 2. The squared form gives a minimum of −9 at x = −1 over the real numbers. Setting x = 0 gives y = −8, the y-intercept. Every form gives the same results if used correctly, but each reveals a different feature with less work.

The transferable skill is asking what a representation makes easy to see. That question also belongs in statistics, geometry and word problems, not only quadratics.

12. Perimeter, area and purchased quantities

Two questions can show the same rectangle and require different mathematics because they ask about different quantities. Perimeter measures a boundary. Area measures a surface. A cost may depend on either, and a purchase may introduce whole packs, minimum quantities or unused material. Do not let the diagram choose the formula before the wording identifies the target.

An invented exhibition panel measures 2.4 m by 1.5 m. A strip is attached around its entire edge. The required strip length is 2(2.4 + 1.5) = 7.8 m. A sheet covers the front surface. The area is 2.4 × 1.5 = 3.6 m². These are different physical jobs, so they have different units.

If the strip costs $2 per metre, the strip cost is $15.60 under a continuous-length pricing model. If covering material costs $9 per square metre, the surface cost is $32.40 under an area pricing model. Multiplying the perimeter by the area rate would combine incompatible quantities.

A gap changes the boundary task, not necessarily the surface task

Suppose a 0.6 m section of the edge needs no strip because it rests against a support. The strip length becomes 7.8 − 0.6 = 7.2 m. The front area remains 3.6 m² unless the question also changes the surface being covered. A number describing a boundary gap should not be subtracted directly from an area.

That last mistake is not a minor unit omission at the end. It is a wrong operation on different kinds of quantity. Writing units beside intermediate values can expose it before the answer is finalised.

Squared units need squared conversion factors

One metre equals one hundred centimetres. Therefore one square metre is a square measuring one hundred centimetres by one hundred centimetres, or ten thousand square centimetres. Converting area is not the same as converting length.

The panel’s 3.6 m² equals 36,000 cm². You can check using centimetre dimensions: 240 × 150 = 36,000. Converting both dimensions first and multiplying agrees with multiplying in metres and converting the area afterwards. Agreement between these routes is a useful check on the power of the conversion factor.

Calculated need and purchased amount can differ

Suppose strip material is sold only in sealed two-metre lengths. A total need of 7.2 m requires four lengths, providing eight metres. The unused 0.8 m is a consequence of the packaging rule. Dividing 7.2 by two gives 3.6 packs, but you cannot buy a fractional sealed pack under the stated rule.

Do not generalise “always round up” to every context. Round according to the actual task. If you are reporting a measured length to a specified precision, ordinary rounding rules may apply. If you need enough whole packages to cover a requirement, a ceiling decision may apply. If you are finding how many complete pieces can be cut from a fixed length, a floor decision may apply instead.

For example, a ten-metre roll can provide five complete pieces of length 1.8 m, leaving one metre, if cutting loss is explicitly ignored. Six pieces would require 10.8 m. The target is now the number of complete pieces obtainable, not the number of rolls needed.

Composite shapes need a visible decomposition

A rectangular board measuring 3 m by 2 m has a rectangular opening measuring 0.8 m by 0.5 m. The area to cover is 6 − 0.4 = 5.6 m² if the opening is left uncovered. But trim around the outer edge alone is ten metres, while trim around both the outer edge and the opening is 10 + 2(0.8 + 0.5) = 12.6 m.

The words “outer edge” and “both edges” choose different boundaries. A learner who memorises “subtract the small rectangle” for every cut-out problem may subtract where addition is needed. Draw the specific surface or boundary being measured, and label it before choosing the arithmetic.

Try a coverage decision

A wall is 4.2 m wide and 2.5 m high, with a window 1.2 m wide and 1 m high. The wall excluding the window needs one coat of paint. Each sealed tin covers 3 m². Ignore wastage and assume the coverage figure applies exactly. How many tins are needed?

Paint task: answer

The wall area is 4.2 × 2.5 = 10.5 m². Subtract the window area of 1.2 m² to obtain 9.3 m². Dividing by three gives 3.1 tins’ worth of coverage, so four whole tins are needed. The answer follows the sealed-tin requirement. It is not a real paint recommendation; actual product coverage and surface conditions would need checking in a practical project.

A complete solution identifies the physical quantity, models the amount required, then applies the purchase or reporting rule. Skipping that final distinction can produce a mathematically accurate number that is not a usable answer.

13. Similarity and the power of the scale

Similarity is a relationship between shapes, not merely a statement that one object looks bigger. Corresponding angles are equal and corresponding lengths share one scale factor. Once that factor is established, length, area and volume respond differently because they involve different numbers of dimensions.

Suppose a length of six centimetres on a smaller panel corresponds to nine centimetres on a larger similar panel. The linear scale factor from smaller to larger is 9/6 = 1.5. Every corresponding length is multiplied by 1.5. The area is multiplied by 1.5² = 2.25, so a smaller area of forty square centimetres becomes ninety square centimetres.

The square is not an arbitrary examination rule. Imagine a rectangle whose width and height both grow by 1.5. Their product grows by 1.5 × 1.5. The same area scaling applies to similar plane figures generally, but the rectangle makes the dimensional reasoning easy to see.

Read the direction of the ratio

A ratio written as small:large = 2:3 is not the same as a multiplier of 2/3 from small to large. Moving from small to large uses 3/2. Moving back uses 2/3. Label the direction before inserting a fraction into the calculator.

If the larger similar figure has area ninety square centimetres, the smaller area is 90 × (2/3)² = forty square centimetres. Dividing by the linear factor only once would give sixty, which answers neither the correct area relationship nor a stated length question.

Area information requires a square-root decision

Suppose two similar logos have areas in the ratio 25:49. Their corresponding lengths are in the ratio 5:7, not 25:49. A length of fifteen centimetres on the smaller logo corresponds to 15 × 7/5 = twenty-one centimetres on the larger.

The question supplied an area ratio, so the first decision is to recover the linear scale. This is the reverse of the previous task. The surface words “similar logos” do not determine whether to square or square-root; the type of given quantity and the type of required quantity do.

Volume introduces a third dimension

For similar solids with linear scale factor two, volume scales by eight. A model container with volume 150 cm³ would correspond to a similar container of volume 1,200 cm³ at twice every linear dimension. This assumes the entire solid scales uniformly, including the relevant internal dimensions if the quantity is capacity.

A practical container with thicker walls or a different interior is not automatically a uniform enlargement. The mathematical model needs the similarity condition. This is another example of a useful formula whose applicability depends on the wording, not merely on the object being a box.

A single matching ratio does not prove similarity

Compare a rectangle measuring four centimetres by six centimetres with one measuring six centimetres by eight centimetres. One pair of sides has ratio 4:6 = 2:3, but the other pair has ratio 6:8 = 3:4. There is no single common scale factor, so the rectangles are not similar.

If you used 1.5² to scale the first area of twenty-four square centimetres, you would predict fifty-four. The actual second area is forty-eight. The arithmetic could be flawless while the similarity assumption was wrong.

Orientation can also distract. Rotating or reflecting a shape does not by itself change its side lengths or angles. When establishing correspondence, follow labels and geometric relationships rather than matching whichever sides happen to be horizontal on the page.

Try two directions of the same idea

A: Similar figures have corresponding lengths in the ratio 3:5. The smaller area is seventy-two square centimetres. Find the larger area. B: Similar figures have areas in the ratio 9:25. A corresponding larger length is twenty centimetres. Find the smaller length.

Similarity tasks: answers

A: the area ratio is 9:25, so the larger area is 72 × 25/9 = 200 cm². B: the length ratio is 3:5, so the smaller length is 20 × 3/5 = 12 cm. Both tasks involve the same scale relationship, but the given and requested dimensions determine which power is relevant.

In your decision sentence, name both dimensions: “Given a length ratio, required an area,” or “Given an area ratio, required a length.” This brief note is more protective than remembering a vague instruction to square whenever the word similar appears.

14. Pythagoras or trigonometry?

A right-angled triangle offers several possible relationships. The useful one depends on what is known and what is required. Recognising the right angle is necessary, but it does not by itself choose between Pythagoras and a trigonometric ratio.

If two sides are known and the third side is required, Pythagoras often provides the direct route. For perpendicular sides of lengths five and twelve centimetres, the hypotenuse h satisfies h² = 5² + 12² = 169, so h = 13 cm. The positive square root is used because h is a physical length.

If the hypotenuse is thirteen and one perpendicular side is twelve, the other side b satisfies b² + 12² = 13². Thus b² = 25 and b = 5 cm. Adding the two given squares would be wrong here because the required side is not the hypotenuse. First identify the role of each side.

An angle and a side suggest a ratio

Suppose a right-angled triangle has an acute angle of thirty degrees and hypotenuse ten centimetres. The opposite side is found from sin 30° = opposite/10. Since sin 30° = 1/2, the opposite side is five centimetres.

Pythagoras alone is not the most direct starting tool because only one side is known. The angle supplies a relationship between sides through a trigonometric ratio. The decision is based on the information available, not a preference for trigonometry over geometry.

Opposite and adjacent depend on the chosen angle

The hypotenuse is always opposite the right angle. The other side labels depend on the acute reference angle. A side opposite one acute angle is adjacent to the other. Rotating the drawing does not change those relationships, but moving your attention to the other angle does.

To choose a ratio, identify the reference angle first. Then name the known and required sides relative to it. Sine connects opposite and hypotenuse. Cosine connects adjacent and hypotenuse. Tangent connects opposite and adjacent. Choose the ratio that contains the quantities you have and the quantity you need.

For an angle θ with opposite side seven centimetres and adjacent side eleven centimetres, tan θ = 7/11. The angle is arctan(7/11), approximately 32.5°. The inverse tangent finds an angle from a ratio; it is not the reciprocal of tan θ. Check that the calculator is using degree mode for this degree-based question.

Unknown placement changes the algebra

Suppose sin 35° = 8/h and the unknown h is the hypotenuse. Multiplying by h gives h sin 35° = 8, so h = 8/sin 35°, approximately 13.9 cm. Multiplying eight by the sine would give a shorter length and contradict the role of the hypotenuse.

This is a valuable place to separate method selection from execution. A student may choose sine correctly but rearrange the equation incorrectly. Giving a lecture about choosing sine versus cosine would miss the actual weak step. Ask the student to show the equation before using the calculator, so the two decisions remain distinguishable.

Two steps can be simpler than one guessed formula

A rectangular frame is twelve centimetres high and five centimetres wide. A diagonal brace joins opposite corners. First find the brace length, then find its angle with the horizontal base. The length is thirteen centimetres by Pythagoras. For the angle θ, tan θ = 12/5, so θ is approximately 67.4°.

You could find the angle using another appropriate ratio after finding the diagonal. The direct tangent route does not require the diagonal, so it is efficient for the angle alone. If the question asks for both quantities, the two calculations answer different parts. Do not assume that a value found in part one must be used in part two unless the wording or structure makes it useful.

The right-angle condition cannot be invented

For a triangle with two sides five and twelve centimetres but no right-angle condition, the third side is not necessarily thirteen. The included angle changes the possible third length. A diagram that merely looks perpendicular is not a substitute for a given right angle or a justified geometric argument.

This is a condition check, not a memory check. The student may remember Pythagoras perfectly and still use it incorrectly because the triangle was misclassified.

Try a contrasting pair

A: A right-angled triangle has perpendicular sides eight and fifteen centimetres. Find the hypotenuse. B: A right-angled triangle has hypotenuse sixteen centimetres and a thirty-degree acute angle. Find the side adjacent to that angle. Name the different information patterns.

Triangle pair: answers

A uses two known perpendicular sides: h = √(8² + 15²) = 17 cm. B uses an angle and the hypotenuse: adjacent = 16 cos 30° = 8√3 cm, approximately 13.9 cm. In B, sine would find the opposite side instead. The numerical answer should be accompanied by the correct side role.

A strong triangle solution therefore has three distinct decisions: establish the geometry, choose the relationship containing the relevant quantities, and solve that relationship without changing it.

15. Geometry needs evidence, not appearance

Geometric diagrams help organise information, but their appearance does not automatically supply facts. A line can look horizontal without being parallel to another line. A triangle can look isosceles without equal-side information. A marked right angle is evidence; a corner that seems square on the page may not be.

Consider triangle ABC with AB = AC and angle BAC = 40°. The equal sides imply equal base angles at B and C. The remaining angle sum is 180° − 40° = 140°, so each base angle is seventy degrees. The equal-side condition is what allows the remaining total to be divided equally.

Remove AB = AC while keeping the forty-degree angle. Then angles B and C still sum to 140°, but they need not both equal seventy. They could be sixty and eighty, or another permitted pair. The numerical total survives; the equality does not.

Name the property at the step where it is used

A useful working sequence is “AB = AC, so angle ABC = angle BCA; angles in a triangle sum to 180°; therefore each is 70°.” The reasoning is short, but each step has a source. Simply writing “70° because triangle” is not enough to distinguish the equal-angle step from the angle-sum step.

The same principle applies to parallel lines. Corresponding angles are equal when a transversal crosses parallel lines. If parallelism is not given, an equality of the relevant corresponding angles may instead be used to establish parallelism through the converse property. Do not reverse a statement merely because the reversed sentence sounds plausible; identify the actual converse theorem.

A local angle and a whole shape

Suppose a straight line through B extends side AB of triangle ABC, and the exterior angle at B is 125°. The adjacent interior angle at B is fifty-five degrees because the two form a straight angle. If angle A is forty-five degrees, angle C is eighty degrees from the triangle sum.

You can also use the exterior-angle property: the exterior angle equals the sum of the two opposite interior angles, so angle C = 125° − 45° = 80°. Comparing these routes helps you see that a compact theorem can summarise a longer chain. It does not mean that any two angles near an exterior line can be added without tracking their positions.

Similarity and congruence answer different questions

Similar triangles have equal corresponding angles and proportional corresponding sides. Congruent triangles have the same shape and size. A pair of equal angles can support a similarity argument for triangles, but not necessarily congruence. A small and large equilateral triangle have equal corresponding angles while their side lengths differ.

Before using a side equality inferred from a diagram, ask whether you have established congruence or only similarity. In a similarity problem, a scale factor may still be needed. The method changes because the conclusion supported by the evidence is different.

Counterexamples sharpen a rule

Take the claim “Every quadrilateral with one right angle is a rectangle.” It is false. A quadrilateral can have one right angle and three other angles that are not right angles. You do not need to inspect every quadrilateral to reject the universal claim; one valid counterexample is sufficient.

Now take “A quadrilateral with three right angles has a fourth right angle.” In ordinary plane geometry, its interior angles sum to 360°, so the remaining angle is ninety degrees. The stronger condition changes the conclusion. This is a useful comparison because it trains attention to how much information is actually supplied.

Try a proof-and-calculation task

Triangle PQR has PQ = PR. Its angle at P is 52°. Find the angle at Q. Then explain what can still be concluded if the equal-side condition is removed. Finally, suppose instead that the angles at Q and R are known to be equal. What relationship between sides follows?

Triangle evidence task: answer

With PQ = PR, the base angles at Q and R are equal and sum to 128°, so each is 64°. Without the equal-side condition, you know only that their sum is 128°. If the base angles are given equal, the converse isosceles-triangle property gives PQ = PR. The equal sides lie opposite the equal angles; correspondence matters.

Geometry becomes more manageable when every new number is tied to a property and every property is tied to a condition. The diagram helps you hold the relationships together, but the argument—not the attractiveness of the picture—decides what you may conclude.

16. Probability begins with the experiment

Before choosing a probability calculation, describe the experiment. What can happen? Which outcomes are equally likely? Is something replaced between selections? Does an earlier outcome change the next stage? The words defining the experiment are part of the mathematics.

Take a bag containing four red counters and two blue counters, all equally likely to be drawn. For one draw, the probability of blue is 2/6 = 1/3. This calculation uses a count of equally likely individual counters. It would need reconsideration if the physical selection mechanism did not make the counters equally likely.

Replacement keeps the second-stage composition unchanged

Draw a counter, record its colour, replace it and mix before drawing again. Under the stated random model, the probability of blue on each draw is one third, and the draws are independent. The probability of two blue counters is (1/3)(1/3) = 1/9.

The multiplication follows the sequence of events. It is not because every probability question containing the word “and” permits multiplication of unconditional probabilities. If the first event changes the second, the second factor must reflect the changed condition.

Without replacement, the second probability changes

Draw two counters without replacement. For two blues, the first factor is still 2/6. After one blue is removed, one blue remains among five counters, so the second factor is 1/5. The probability is 2/30 = 1/15.

Using one third for both factors would answer the replacement experiment, not the experiment stated. The same bag and the same colours appear in both questions. The decisive difference is the word replacement.

Exactly one blue has two possible orders

Without replacement, exactly one blue can occur as blue then red or red then blue. The first route has probability (2/6)(4/5) = 8/30. The second has probability (4/6)(2/5) = 8/30. These routes are mutually exclusive, so their probabilities add to 16/30 = 8/15.

Two reds have probability (4/6)(3/5) = 12/30 = 2/5. Check the complete collection: two reds, exactly one blue and two blues have probabilities 6/15, 8/15 and 1/15, which sum to one. This does not prove every detail was correct, but it is a useful completeness check.

“At least one” suggests considering the complement

In the same two-draw experiment without replacement, the probability of at least one blue is one minus the probability of no blues. No blues means two reds, so the required probability is 1 − 2/5 = 3/5. You can check by adding exactly one blue and two blues: 8/15 + 1/15 = 9/15 = 3/5.

The complement route is efficient because “none” is simple to calculate. It is not a rule that complements are always shorter. Choose the event description that reduces the number of disjoint cases you must organise reliably.

Independent is not the same as mutually exclusive

Mutually exclusive events cannot both occur on the same trial. Independence concerns whether knowing one event occurred changes the probability of the other. For one roll of a fair six-sided die, “an even result” and “an odd result” are mutually exclusive, not independent: knowing the result is even makes the probability of odd zero.

For two independent rolls, “the first result is even” and “the second result is even” can both occur. The probability of each remains one half regardless of the other, so the joint probability is one quarter. Confusing these terms can lead to choosing addition where multiplication is required or vice versa.

Do not assume equally likely sums

Roll two fair six-sided dice. The ordered pairs are equally likely under the usual independent model, but the possible sums are not. A sum of two has one ordered pair, (1,1). A sum of seven has six: (1,6), (2,5), (3,4), (4,3), (5,2), (6,1). Counting eleven possible sums and assigning each probability 1/11 ignores their different numbers of underlying outcomes.

Try a comparison with a different bag

A bag contains three green and five yellow counters. Two are drawn without replacement. Find the probability that both are green and the probability that at least one is green. Then say which probability would change if replacement were introduced.

Counter comparison: answers

Both green without replacement: (3/8)(2/7) = 3/28. At least one green: 1 − (5/8)(4/7) = 1 − 5/14 = 9/14. With replacement, both green would be 9/64 and at least one green would be 1 − 25/64 = 39/64. Both requested probabilities change because the experiment changes.

A probability solution begins with a faithful model of what is happening. The arithmetic comes after the outcomes and conditions have been organised.

17. Statistics: identify what is being averaged

The word average can hide a choice. A question may ask for a mean, a median, a combined mean, a weighted mean or a comparison of distributions. Before selecting a calculation, identify the observations and how much weight each observation carries.

Suppose one group of eight students has a mean score of twelve and another group of twelve students has a mean score of seventeen. The combined total is 8 × 12 + 12 × 17 = 96 + 204 = 300. There are twenty students, so the combined mean is fifteen.

The simple mean of the two group means is (12 + 17)/2 = 14.5. That calculation gives equal weight to the groups, not equal weight to the students. Because the groups contain different numbers of students, it does not produce the combined student mean.

Recover the total before recombining

A mean is total divided by count. Therefore total is mean multiplied by count. This relationship is useful in missing-value and combined-group questions. It is more dependable than memorising a separate formula for each story type.

Five readings have mean eighteen. Four readings are twelve, sixteen, twenty and twenty-one. The total must be ninety, while the known total is sixty-nine. The missing reading is twenty-one. The reasoning begins by recovering the total; it does not begin by averaging the four known readings and guessing the fifth.

A frequency table weights repeated values

Suppose a fictional exercise records the number of books read by ten students. Two students read one book, five read two books and three read four books. The total is 2 × 1 + 5 × 2 + 3 × 4 = 24 books, so the mean is 2.4 books per student.

The mean of the three listed category values, (1 + 2 + 4)/3, treats each category as one observation. That is not the given data. The frequencies tell you how many observations each category represents. Read the table headings before deciding what belongs in the numerator and denominator.

The median answers a different question

For the ordered values 2, 3, 3, 4, 18, the median is three, the middle observation. The mean is thirty divided by five, or six. Both are correct summaries, but they describe the data differently. The large final value raises the mean while the middle position remains three.

Do not turn this example into the universal rule that the median is always better. A total-resource question may need the mean because it connects directly to the total. A question about a central position in a skewed set may favour the median. Explain the choice in relation to the stated purpose and distribution.

Equal means can hide different spreads

The sets 8, 10, 12 and 2, 10, 18 both have mean ten. Their ranges are four and sixteen respectively. A claim that the two sets are “the same because the averages match” ignores their different variation. If consistency matters, a measure of spread must be considered alongside the centre.

Range itself has limits: it uses only the largest and smallest values. It does not describe every detail of the distribution. The appropriate level of analysis depends on what the question supplies and asks; do not invent missing data or an unsupported conclusion about every member of a group.

A comparison needs comparable conditions

Suppose an exhibition reports that an average visitor stayed longer on Saturday than Friday. Before attributing the difference to a new display, ask whether the groups, opening hours, sampling process or other conditions differed. A numerical difference can be real without establishing the proposed cause.

In a school question, answer within the supplied evidence. You can state that the observed mean was higher. You cannot automatically state that the display caused every visitor to stay longer. This distinction between description and causal interpretation is part of reading a statistical claim carefully.

Try choosing the denominator

A: Six pupils have a mean of fourteen points and four pupils have a mean of nineteen points. Find the combined mean. B: A set contains the values 4, 6, 7, 8 and 30. Find both the mean and median. Which summary would you use to describe the middle position, and why?

Statistics tasks: answers

A: the combined total is 6 × 14 + 4 × 19 = 160, and the count is ten, so the combined mean is sixteen. B: the total is fifty-five and the mean is eleven; the median is seven. The median describes the middle ordered observation. A statement about total points would still depend on the sum and count, not only the median.

The first decision in a statistics problem is often “What are the observations?” Once that is clear, the appropriate weighting and denominator become easier to defend.

18. Rates and the wrong average

Average speed is total distance divided by total time. It is not automatically the arithmetic mean of two speeds. Whether a simple average works depends on what is equal across the journey: times, distances or neither.

Imagine a person travelling twelve kilometres at six kilometres per hour and then twelve kilometres at four kilometres per hour. The first stage takes two hours; the second takes three hours. The total distance is twenty-four kilometres and the total time is five hours. The average speed is 4.8 km/h.

The arithmetic mean of six and four is five, but the traveller spends more time at the slower speed. Giving the two speeds equal weight therefore does not reflect the time spent travelling at each. Recovering distance and time avoids the mistaken shortcut.

Equal times produce a different result

Now travel for two hours at six kilometres per hour and two hours at four kilometres per hour. The distances are twelve and eight kilometres. The total is twenty kilometres in four hours, so the average speed is five kilometres per hour. The simple average works here because the durations are equal.

The two stories contain exactly the same speeds. The decisive difference is what the question keeps equal. This is an excellent contrast pair because it forces a learner to inspect the relationship instead of copying a remembered calculation from the first version.

Unit conversion is part of the model

A machine transports a tray eighteen metres in forty-five seconds. Its speed is 18/45 = 0.4 metres per second. To express that speed in metres per minute, multiply by sixty to obtain twenty-four metres per minute. Dividing by sixty would answer a different conversion direction.

Writing the units through the conversion helps: 0.4 metres per second multiplied by sixty seconds per minute gives twenty-four metres per minute. The seconds cancel in the unit relationship. This is a reason for the operation, not an extra unit attached after the arithmetic.

A graph’s vertical quantity matters

Suppose a graph shows position along a straight track against time. Position rises from zero to twelve metres in three seconds, remains at twelve for two seconds, then falls to four metres over the next four seconds. The first-stage velocity is four metres per second, the stationary stage has zero velocity, and the final-stage velocity is negative two metres per second.

The negative gradient indicates movement in the opposite coordinate direction. It does not mean a negative physical speed. The final-stage speed is two metres per second. The total distance travelled is twelve plus eight, or twenty metres, while the overall displacement is four metres. Over nine seconds, average speed is 20/9 m/s and average velocity is 4/9 m/s.

These distinctions depend on the graph showing position, not cumulative distance travelled. A graph of total distance already travelled would not fall while the traveller returns. Read what the vertical axis represents before interpreting a downward line.

Two rates can oppose one another

An invented tank model has water entering at seven litres per minute and leaving at two litres per minute, both constant while the tank is neither empty nor overflowing. The net increase is five litres per minute. Starting from eighteen litres, the volume after six minutes is forty-eight litres.

Adding the two rates would describe their combined flow magnitudes, not the change in stored water. The direction of each flow determines the sign in the storage model. If either rate changes with time, or the tank reaches capacity, the simple linear model must be adjusted.

A rate model can have a boundary

Suppose the tank’s capacity is sixty litres. At a net increase of five litres per minute from eighteen litres, it reaches capacity after (60 − 18)/5 = 8.4 minutes. The formula V = 18 + 5t applies only until that point under the stated no-overflow storage model. Using it at twenty minutes and reporting 118 litres inside a sixty-litre tank ignores the physical constraint.

A correct equation can therefore become an incorrect answer when used outside its intended interval. The domain is not only an algebraic restriction on denominators; it can also arise from the modelled situation.

Try a journey pair

A: A vehicle travels thirty kilometres at thirty kilometres per hour and thirty kilometres at sixty kilometres per hour. Find its average speed. B: It travels for one hour at each of those speeds. Find the average speed in the second scenario.

Journey pair: answers

A: the times are one hour and half an hour, so average speed is 60/1.5 = 40 km/h. B: the total distance is ninety kilometres over two hours, giving 45 km/h. The same speeds produce different averages because the weighting conditions differ. Both results follow from total distance divided by total time.

When a question combines rates, ask what is conserved, what accumulates and what is being averaged. Those questions choose the relationship more reliably than the presence of familiar speed numbers.

19. The courtyard exhibition case

The following case joins several familiar relationships without announcing a chapter above each question. All prices, dimensions and observations are invented. The setting is a fictional courtyard exhibition, not a real school’s event, a quotation from suppliers or a plan that has been safety-checked for practical use.

Leah, Harun and Wen are helping a committee compare options. Their task is not to make every possible calculation. It is to produce a decision sheet that another person can inspect: what is required, which relationship applies, what the numbers say, and which assumptions limit the conclusion.

The information sheet

Invented exhibition data

Item	Stated information
Chair supplier A	$18 fixed delivery fee plus $3 per chair.
Chair supplier B	$4.50 per chair, no fixed fee.
Display mat	A rectangle 9 m by 6 m. A 2 m opening in its boundary needs no edging strip.
Edging strip	Sold only in sealed 5 m rolls. Ignore joining losses.
Tickets	Student tickets cost $4 and adult tickets $7. Thirty-six tickets produce $180.
Tank demonstration	Starts with 20 litres. Inflow is 6 litres per minute and outflow 2 litres per minute, both constant until capacity of 80 litres.
Visitor observations	Ten visitors in one session have mean stay 6 minutes; twenty in another have mean stay 9 minutes.
Counter activity	A bag has two gold and three silver counters. Two are selected uniformly without replacement.

Decision 1: which chair quotation applies?

The committee needs twenty-four chairs. Find the total under each quotation and recommend the cheaper one under the stated rules. Then identify the order size where the two costs are equal. Do not describe either supplier as always cheaper.

Before calculating, predict the direction. A fixed delivery charge is disadvantageous for a small order, while a lower per-chair price becomes more important as the order grows. That prediction makes a later algebraic threshold meaningful rather than merely another value of n.

Chair decision: worked recommendation

Supplier A costs 18 + 3 × 24 = $90. Supplier B costs 4.5 × 24 = $108. A is cheaper by eighteen dollars for this order. Equality occurs when 18 + 3n = 4.5n, giving n = 12. At twelve chairs both cost fifty-four dollars. Below that threshold B is cheaper; above it A is cheaper. This conclusion assumes the stated charges cover the relevant order and that no additional conditions apply.

Decision 2: how much edging must be bought?

Find the length of boundary that needs edging, then the number of whole rolls needed. Explain why the area of the mat is not the starting quantity. Also calculate the unused strip after the required edging is cut, assuming every needed piece can be obtained without additional cutting loss.

Edging decision: quantity and packaging

The full perimeter is 2(9 + 6) = 30 m. Subtract the two-metre opening to obtain 28 m. Dividing by five gives 5.6 rolls’ worth, so six whole rolls are required. They supply thirty metres, leaving two metres unused. The area, 54 m², would be relevant to covering the mat’s surface but does not measure its boundary. The whole-roll decision comes after the required length is calculated.

Decision 3: how many tickets of each type?

The total number and total receipts are known. Define two ticket-count variables, write the two conditions and solve. Then check both the count and the money. Do not treat the ratio of the two ticket prices as the ratio of the numbers sold.

Ticket decision: independent conditions

Let s be student tickets and a adult tickets. Then s + a = 36 and 4s + 7a = 180. Subtract four times the count equation from the money equation to obtain 3a = 36, so a = 12 and s = 24. The checks are 24 + 12 = 36 and 24 × 4 + 12 × 7 = 180. Prices describe values per ticket, not how many customers choose each type.

Decision 4: what does the tank model allow?

Find the volume after twelve minutes and the time when the tank first reaches capacity. Explain why the same storage equation should not be used to claim a volume above eighty litres inside this tank.

Tank decision: rate, starting value and boundary

The net inflow is 6 − 2 = 4 litres per minute, so before capacity V = 20 + 4t. At twelve minutes V = 68 litres. Capacity is reached when 20 + 4t = 80, giving t = 15 minutes. After that, the original storage model no longer applies without further information about overflow or changes to the flows. The model’s valid interval is part of the answer.

Decision 5: what can the visitor data say?

Find the combined mean stay for the thirty observed visitors. A committee member says the exhibition redesign therefore made every visitor stay longer. Explain why the calculated mean does not establish that statement.

Visitor decision: weighted mean and a restrained claim

The combined total stay is 10 × 6 + 20 × 9 = 240 minutes. Dividing by thirty gives a combined mean of eight minutes. The data do not show that every individual stayed longer, and the comparison alone does not establish that a redesign caused the difference. The two groups and conditions may differ. The defensible numerical statement is about the observed means and the combined mean.

Decision 6: what is the chance of two gold counters?

Use the stated experiment, including the replacement condition. Then describe how the calculation would change if the first counter were replaced and the bag mixed before the second draw.

Counter decision: the changed second stage

Without replacement, the probability is (2/5)(1/4) = 1/10. After one gold counter is removed, one gold remains among four counters. With replacement and independent draws under the random model, the probability would be (2/5)(2/5) = 4/25. The change in procedure changes the sample space for the second stage.

What a useful decision sheet contains

The finished sheet should distinguish calculations from recommendations. “Supplier A costs ninety dollars for twenty-four chairs” is a calculation under a pricing model. “Choose A for this order” is a recommendation based on that model. “A is always cheaper” is an overgeneralisation. The boundary between those statements matters.

Likewise, “six rolls are needed” depends on the packaging and cutting assumptions. “The tank reaches capacity at fifteen minutes” depends on the stated constant rates. “The combined visitor mean is eight minutes” describes the supplied observations but does not prove a causal story about the exhibition.

Ask each fictional learner to explain one decision to another. Leah must preserve the task rather than invent an equation where none is needed. Harun must identify the shared structures beneath tickets, chairs and water. Wen must select a valid route, complete it and choose one meaningful check instead of reopening every possible method.

The case is complete when the reasoning is inspectable. A page full of correct numbers without labels can still be hard to use. A clear decision sheet connects each result to its question, unit, assumption and consequence.

20. Compare solutions without collecting tricks

Comparing methods is useful when the comparison has a purpose. The goal is not to collect ten impressive ways to solve every question. It is to understand what each route preserves, what it makes visible, and when it reduces unnecessary work.

For the equation 4(x + 3) = 28, one route expands to 4x + 12 = 28, subtracts twelve and divides by four. Another divides both sides by four first, giving x + 3 = 7, then subtracts three. Both produce x = 4. The second route is shorter here because the entire left side is a product by four and the right side divides cleanly.

Now change the equation to 4x + 3 = 28. Dividing the whole equation by four is still legal, but it gives x + 3/4 = 7. Subtracting three first gives 4x = 25 and x = 25/4. A rule such as “always divide first when you see a four” misses the difference in structure.

Compare the point of decision

Place the two equations side by side and ask what the brackets change. In the first, the multiplier four acts on x + 3 as a whole. In the second, it acts only on x. This is a more useful comparison than timing two students and announcing that one has a better trick.

The same approach works with percentages. Comparing 0.8P = 72 with P − 0.2P = 72 reveals that the multiplier route compresses the subtraction of twenty per cent of the same base. It does not work by ignoring the discount. The compact form is justified by the expanded reasoning.

What research supports, and what it does not

In an experiment with seventy seventh-grade students, Bethany Rittle-Johnson and Jon R. Star found greater gains in procedural knowledge and flexibility when learners compared equation-solving methods rather than studying the methods separately; conceptual gains were comparable. Their 2007 study supports taking comparison seriously as a teaching approach. It does not establish that every learner should compare every method, or that this particular workshop guarantees a mark increase.

Use three comparison questions

First ask what stays valid in both solutions. They should preserve the original relationship and domain. Next ask where the methods differ. One may expose factors, another may isolate a variable sooner, and another may make a geometric condition visible. Finally ask which feature of the question makes one route convenient here.

For x² − 9 = 0, factorisation produces the roots immediately. For x² = 9, taking square roots is also direct, provided both x = 3 and x = −3 are retained. For (x − 2)² = 9, the square-root route gives x − 2 = ±3 and therefore x = 5 or x = −1. Expanding would work, but it would hide a convenient square that the question already supplied.

The comparison should end with a condition, not a slogan. “A square is already isolated” is a reason to consider square roots. “Quadratics use factorisation” is an incomplete description of a broad family.

A shorter solution is not automatically better for this learner

Wen may understand a compressed method but use it unreliably under pressure. In that case, an extra line preserving brackets can be worthwhile. Leah may benefit from writing “simplify only” above an expression until she stops inventing equations. Harun may need a variable definition before recognising a linear model.

The comparison should gradually improve their control, not reward the smallest number of written characters. Efficient working is sufficient working with a clear purpose. A line that prevents a recurrent error is not wasted merely because a more experienced mathematician could do it mentally.

Try choosing rather than performing every method

For each equation, choose a convenient first operation and explain the structural reason: 5(x − 2) = 35; 5x − 2 = 35; (x + 1)² = 16; x² + 2x − 15 = 0. You do not have to solve each in two ways. The purpose is to select thoughtfully.

Comparison task: suggested routes

For 5(x − 2) = 35, divide by five first, then add two, giving x = 9. For 5x − 2 = 35, add two first, then divide, giving x = 37/5. For (x + 1)² = 16, take both square-root possibilities, giving x = 3 or x = −5. For x² + 2x − 15 = 0, factorise as (x + 5)(x − 3), giving x = −5 or x = 3. Other valid routes remain acceptable unless the task specifies a method.

Comparison has done its job when a learner can name the feature that changed the choice. It has not done its job merely because the learner has copied two finished solutions into a notebook.

21. Starting an unfamiliar question

An unfamiliar question does not need to be familiarised completely before you can begin. You need a first justified move. That move might be defining a quantity, drawing a relationship, converting units, testing a simple case or recording a condition. It need not be the final formula.

Imagine a question about a machine that places dots along the border of rectangular cards. The context may be new, but the first task is to identify what is being counted. Are corner dots counted once or twice? Is spacing measured along the full perimeter? Is each card independent? The familiar word rectangle does not settle these questions.

Translate the target into an object

Ask whether the answer should be a length, area, count, rate, probability, equation or argument. This immediately rules out some unhelpful calculations. If the target is a count of complete packages, a decimal may be an intermediate result. If the target is an identity, testing one input is not enough. If the target is a probability, a result greater than one signals a problem.

Then identify one relationship you trust. You may know that a rectangle’s opposite sides are equal, that a mean connects total and count, or that a fixed fee cancels when two bills are subtracted. Write that relationship before worrying about the entire route.

Use a smaller case honestly

Suppose a pattern is difficult to read at the twentieth stage. Construct the first three stages carefully. A small case can expose how new pieces are added or where overlap occurs. It can suggest a rule. It does not, by itself, prove that the rule must continue unless the construction or stated pattern supports that generalisation.

This is a useful distinction for both cautious and impulsive learners. Wen can use a small case to escape endless method comparison. Leah must still justify the general rule rather than turning three matching numbers into an unsupported identity.

Work backwards from a target when it reveals a dependency

If the question asks for paint cost, you may first need painted area; to find that area, you may need dimensions; to find a dimension, you may need an equation from a perimeter condition. Writing that short dependency chain can organise the problem without performing any calculation yet.

Working backwards does not mean applying inverse operations indiscriminately. It means identifying which intermediate information would make the requested result available. Once those dependencies are clear, the forward calculation becomes less mysterious.

Recognise when information is insufficient

A prompt states that a rectangle has area twenty-four square centimetres and asks for its perimeter. Without another condition, there is no unique answer. A four-by-six rectangle has perimeter twenty; a three-by-eight rectangle has perimeter twenty-two. A correct response identifies the missing restriction rather than assuming the rectangle is a square.

Being able to say “not determined by these facts” is part of mathematical control. It should be supported by an explanation or counterexample, not used as an escape from a solvable problem. Distinguish genuinely missing information from a relationship that has not yet been noticed.

Use help in a way that preserves the evidence

When studying, ask for the smallest useful hint after a genuine attempt. A prompt such as “Which quantity represents the original one hundred per cent?” preserves more decision-making for the learner than “Divide by 0.85.” If that hint is insufficient, more explicit teaching is appropriate. Record which support was needed so that a later independent attempt can test the repaired decision.

During a timed school task, use the instructions and permitted resources for that task. A practical move-on rule can prevent one blocked item from consuming the whole paper: leave your useful representation and a brief re-entry note, then return after completing accessible questions. The exact timing should be rehearsed with the teacher and paper format, not invented as a universal number of seconds.

The desired change is modest and important. Instead of “I cannot do unfamiliar questions,” aim for “I can identify the target, preserve the conditions and make one justified move.” That gives the next piece of reasoning somewhere to begin.

22. Mixed set A

This set deliberately removes topic labels from the questions. Write a brief first relationship, complete the working, and select one check that uses meaningful evidence. Do not open an answer panel until you have recorded an attempt. These are original teaching questions, not copied examination items or an official mark scheme.

Use the set flexibly. A learner who has not yet studied a particular topic can leave that item for later and record the reason. A learner who needs fraction repair can attempt the relevant chapter before returning. The purpose is to observe decisions under mixed conditions, not to manufacture a score from material that has not been taught.

A1

An item costs $98.40 after an 18% reduction from its original price. Find the original price. State what represents one hundred per cent.

A1: answer and decision

The original price is the one-hundred-per-cent base. The reduced price is eighty-two per cent of it, so 0.82P = 98.40 and P = $120. Check by applying the stated reduction to 120: 120 × 0.82 = 98.40. Adding eighteen per cent of 98.40 would use the reduced amount as the base and would not reverse the original reduction.

A2

A hire service charges a fixed fee plus a constant hourly rate. Three hours cost $20 and seven hours cost $36. Find the cost of ten hours.

A2: answer and decision

The difference between the bills is sixteen dollars for four extra hours, giving a rate of four dollars per hour. The fixed fee is 20 − 3 × 4 = eight dollars. Thus C = 8 + 4h, and ten hours cost $48. Check the other supplied bill: 8 + 4 × 7 = 36. Dividing twenty by three would produce the average cost per hour for the first booking, not the constant additional rate.

A3

Simplify 2(3x − 1) − x. Then solve the separate equation 2(3x − 1) − x = 23. Explain why the outputs differ.

A3: answer and decision

The expression simplifies to 5x − 2. The equation becomes 5x − 2 = 23, so x = 5. Simplifying asks for an equivalent expression; solving asks for an input that satisfies the stated equality. Substituting five into the original left side gives 2(15 − 1) − 5 = 23. There is no reason to set the first expression equal to zero.

A4

Solve x² = 9x, giving every real solution. Explain why dividing immediately by x needs care.

A4: answer and decision

Rearrange to x² − 9x = 0 and factorise: x(x − 9) = 0. The solutions are x = 0 and x = 9. Dividing by x would assume x ≠ 0 and could lose the zero solution. Checking both candidates in the original equation confirms them. The factorised zero-product form preserves both possibilities without requiring a separate division case.

A5

Solve (x² − 25)/(x − 5) = 12, retaining any restriction from the original expression.

A5: answer and decision

The denominator requires x ≠ 5. Factorising the numerator gives (x − 5)(x + 5), so the equation simplifies to x + 5 = 12 for permitted x. Hence x = 7. Seven is allowed and gives (49 − 25)/(7 − 5) = 24/2 = 12. The cancellation is of a common factor, and the original excluded input remains excluded.

A6

A straight line passes through (1, 8) and (5, 0). Find its gradient and equation in the form y = mx + c.

A6: answer and decision

The gradient is (0 − 8)/(5 − 1) = −8/4 = −2. Substituting (1, 8) into y = −2x + c gives c = 10. The equation is y = −2x + 10. Check the second point: x = 5 gives y = 0. The negative gradient matches a falling line. Using the coordinate ratio 8/1 would describe the line from the origin to the first point, not the given line.

A7

Two similar designs have areas in the ratio smaller:larger = 16:81. A length on the smaller is 12 cm. Find the corresponding larger length.

A7: answer and decision

The linear ratio is the square root of the area ratio, so it is 4:9. The larger length is 12 × 9/4 = 27 cm. The key decision is that the given ratio concerns area while the requested quantity is length. Multiplying twelve by 81/16 would use an area multiplier on a length.

A8

A right-angled triangle has hypotenuse 25 cm and one other side 7 cm. Find the remaining side.

A8: answer and decision

Let the missing perpendicular side be b. Then b² + 7² = 25², so b² = 625 − 49 = 576 and b = 24 cm. The hypotenuse is already known, so subtract its partner’s square rather than adding the two given squares. The result is positive and shorter than twenty-five centimetres, as the geometry requires.

A9

Four observations have mean twelve and six observations have mean seventeen. Find the mean of all ten observations.

A9: answer and decision

Recover the totals: 4 × 12 = 48 and 6 × 17 = 102. The combined total is 150 and the count is ten, giving mean fifteen. Taking the unweighted mean of twelve and seventeen would assign equal weight to two groups of different sizes. The relevant observations, not merely the listed summary values, determine the denominator.

A10

A bag contains two black and three white counters. Two counters are selected uniformly without replacement. Find the probability of exactly one black counter.

A10: answer and decision

There are two disjoint orders: black then white, with probability (2/5)(3/4), and white then black, with probability (3/5)(2/4). Adding gives 12/20 = 3/5. The second-stage denominator is four because the first counter is not replaced. “Exactly one” excludes both-black and both-white outcomes, and the two possible orders must both be included.

A11

A traveller covers 10 km at 5 km/h and another 10 km at 10 km/h. Find the average speed over the whole journey.

A11: answer and decision

The stage times are two hours and one hour. Total distance is twenty kilometres over three hours, so average speed is 20/3 km/h, approximately 6.67 km/h. The arithmetic mean of the two speeds, 7.5 km/h, gives equal weight to unequal durations. The equal quantities in the question are distances, not times.

A12

A rectangular surface measures 8 m by 5 m and contains an opening 2 m by 1 m. The remaining surface needs covering. Material is sold only in packs covering 4 m² each. Ignore wastage. How many packs are required?

A12: answer and decision

The area to cover is 8 × 5 − 2 × 1 = 38 m². Dividing by four gives 9.5 packs’ worth of material, so ten whole packs are required. The solution has two decisions: subtract the opening’s area from the whole surface, then apply the whole-pack requirement. Reporting 9.5 as the purchase count would not satisfy the stated packaging rule.

What to record after the set

Choose two items for review, not automatically every item you disliked. For each, identify the earliest point of uncertainty. Was the relationship wrong, or was it correct and then manipulated incorrectly? Did you miss a condition such as replacement, similarity or whole packs? Did you stop at an intermediate quantity?

Then choose a fresh comparison from the relevant chapter. Redoing A1 immediately after reading its solution mainly shows that the solution is still available. A later variation with a different base or direction gives more useful evidence about whether the decision has changed.

23. Mixed set B: a later return

Use this set after time has passed and after you have worked on selected weaknesses, rather than immediately after copying set A’s answers. The surface details change, and several items deliberately change a condition. A good response should notice that change instead of carrying over the previous calculation automatically.

No universal delay is prescribed here. A teacher can choose a later lesson or another study session that fits the learner’s timetable. Record whether the attempt was made with notes, hints or fully independently. The aim is an honest comparison of conditions, not a ceremonial declaration of mastery.

B1

A price is $124.20 after an 8% reduction. Find its original value and show a forward check.

B1: answer and decision

The final price is ninety-two per cent of the original, so P = 124.20/0.92 = $135. The forward check is 135 × 0.92 = 124.20. The percentage differs from set A, but the underlying reverse relationship is the same. The original remains the one-hundred-per-cent base.

B2

A booking rule consists of one fixed fee and a constant hourly charge. Four hours cost $31 and seven hours cost $49. Find the price for ten hours.

B2: answer and decision

Three extra hours add eighteen dollars, so the rate is six dollars per hour. The fixed fee is 31 − 4 × 6 = seven dollars. The rule is C = 7 + 6h, giving $67 for ten hours. Checking the seven-hour booking gives 7 + 42 = 49. The calculation depends on the fixed-plus-constant-rate rule being supplied, not merely on seeing two prices.

B3

Simplify 5(2t + 1) − 3t, evaluate it at t = −2, and solve the equation obtained when it equals 26.

B3: answer and decision

The expression simplifies to 7t + 5. At t = −2 its value is −9. Setting it equal to twenty-six gives 7t + 5 = 26 and t = 3. The outputs are an expression, a numerical value and a solution respectively. Keeping those objects separate prevents a familiar variable from triggering an invented equation.

B4

Solve x² + 2x = 0. Explain why a method that gives only x = −2 is incomplete.

B4: answer and decision

Factorise to x(x + 2) = 0, so x = 0 or x = −2. A route that divides every term by x excludes the possible case x = 0 unless that case is considered separately. Both candidates satisfy the original equation. The important retained condition is not to divide away a factor that may be zero.

B5

Determine whether (z² − 4)/(z − 2) = 4 has a solution. State the original restriction.

B5: answer and decision

The original denominator requires z ≠ 2. For permitted inputs, the left side simplifies to z + 2. The resulting candidate is z = 2, which is excluded. Therefore the original equation has no solution. This differs from A5: cancellation is still useful, but the candidate fails the original restriction. A successful procedure must include that final domain check.

B6

A straight line passes through (2, 3) and (8, 15). Find its gradient and equation. Explain why the line need not pass through the origin.

B6: answer and decision

The gradient is (15 − 3)/(8 − 2) = 12/6 = 2. Substituting into y = 2x + c gives c = −1, so y = 2x − 1. The intercept is not zero; at x = 0 the line gives y = −1. A constant gradient establishes a linear relationship, not necessarily direct proportion.

B7

Two similar figures have corresponding lengths smaller:larger = 2:5. The larger area is 100 cm². Find the smaller area.

B7: answer and decision

The area ratio is 4:25. Moving from larger to smaller gives 100 × 4/25 = 16 cm². Unlike A7, the given ratio is already a length ratio and the required quantity is area, so square the linear ratio. The direction also matters: use the smaller-to-larger area fraction on the known larger area.

B8

A right-angled triangle has hypotenuse 14 cm and an acute angle of 60°. Find the side adjacent to that angle.

B8: answer and decision

Cosine connects adjacent side and hypotenuse: adjacent = 14 cos 60° = 7 cm. The known information is an angle and one side, unlike A8’s two-side pattern. Pythagoras would become useful after another side was found, but cosine provides the direct relationship. The adjacent side must be shorter than the hypotenuse.

B9

Seven readings have mean ten and three readings have mean twenty. Find the combined mean of the ten readings.

B9: answer and decision

The total is 7 × 10 + 3 × 20 = 130. Divide by ten to obtain thirteen. The group sizes determine the weights. The simple mean of ten and twenty would be fifteen, which overweights the smaller high-mean group relative to its number of observations.

B10

A bag contains four red counters and one white counter. Two draws are made with replacement and mixing between draws. Under the independent random-draw model, find the probability that both are red.

B10: answer and decision

The red probability is four fifths on each draw because the first counter is replaced. The joint probability is (4/5)(4/5) = 16/25. Do not reduce the second denominator to four: that would model drawing without replacement. The stated independence and replacement procedure justify using the same second-stage probability.

B11

A traveller spends half an hour at 8 km/h and half an hour at 12 km/h. Find the average speed for the full hour.

B11: answer and decision

The distances are four kilometres and six kilometres. Total distance is ten kilometres in one hour, giving average speed 10 km/h. The arithmetic mean of the speeds also works here because the time intervals are equal. The condition has changed from A11’s equal distances, so reusing the previous weighting would be inappropriate.

B12

A rectangular surface 3.2 m by 2.5 m needs one layer of covering. Each sealed pack covers 1.5 m². Ignore wastage. How many whole packs are needed, and how much coverage is left unused?

B12: answer and decision

The area is eight square metres. Eight divided by 1.5 is 5⅓ packs’ worth, so six packs are needed. They provide nine square metres of coverage, leaving one square metre unused under the stated assumptions. The unused quantity is an area, not a length or a fractional pack count presented without units.

Look for transfer, not resemblance alone

Compare the first relationships in the two sets. B5 is especially informative because a learner who remembers “cancel, solve, finish” may report the excluded value. B11 tests whether the learner re-reads what is equal rather than applying a remembered warning that averaging speeds is always wrong. B8 tests whether an angle changes the useful triangle relationship.

A correct later response is evidence for the sampled task under those conditions. It is not proof that the entire topic will never cause trouble again. Keep the conclusion proportional: “The learner now retained the denominator restriction in this fresh equation” is more defensible and more useful than “All algebra is mastered.”

24. Separate selection, execution and checking

A final answer compresses several decisions into one outcome. To improve the next attempt, separate those decisions again. A wrong answer does not automatically mean the method was poorly selected, and a correct answer does not automatically show that every part of the reasoning was secure.

Imagine a learner who writes the correct gradient relationship, (29 − 11)/(8 − 2), then records 18/6 = 4. The method selection is appropriate. The arithmetic is wrong. By contrast, a learner who writes 29/8 and calculates it accurately has executed a calculation correctly but selected the wrong comparison for this line.

Those two responses should not receive identical feedback. “Remember the gradient formula” may be useful for the second learner but redundant for the first. “Check your division” may repair the first response while leaving the second learner’s misunderstanding untouched.

Use an observation record, not a personality label

A compact record for one attempted task

Stage	What to record
Selection	The first relationship written and the condition cited.
Execution	The first line where a valid relationship was changed incorrectly.
Interpretation	Whether the result answered the requested quantity and retained its unit or restriction.
Checking	The check attempted and whether it could detect the observed error.
Support	Any hint, model answer, note or teacher prompt used.
Later evidence	A fresh related task and the conditions of that attempt.

This is a teaching record, not a validated diagnostic scale. It should not be converted into a numerical claim about intelligence or an automatic prediction of examination grades. Its purpose is to help a teacher decide what to ask or teach next.

Choose a probe that distinguishes plausible explanations

Suppose a learner fails the fixed-fee task. There are several possibilities. The learner may not understand “fixed fee”. They may understand the wording but not recognise that subtracting two bills removes the fee. They may form the right equations but make an algebraic error.

A useful follow-up changes one part at a time. Ask the learner to explain a fixed fee in ordinary language. Then give the equations C = a + 3b and C = a + 7b with the appropriate bill values already substituted. Finally ask for a fresh contextual model. The pattern of responses narrows the teaching question without pretending that one wrong answer reveals a permanent cause.

A check should be capable of catching the error

For a solved equation, substitution into the original can test whether the candidate works. For a factorisation, expansion can test the rewrite. For a percentage reversal, applying the percentage forward can confirm the recovered base. For a probability collection, checking the total can reveal missing or overlapping cases.

Not every check detects every error. A probability answer can lie between zero and one and still be wrong. A dimensionally correct area can still use the wrong dimensions. A factorisation can match at one chosen value and fail elsewhere. Treat plausibility checks as useful filters, not certificates of complete correctness.

Review a correct answer with an invalid argument

Suppose someone solves x² = 9x by dividing by x and writes x = 9, then later remembers to append zero without explanation. The final solution set is correct, but the written reasoning does not explain why zero was retained. Ask for a complete route using x(x − 9) = 0 or a separate consideration of x = 0 before division.

Similarly, a student might obtain the correct combined mean by chance when two groups happen to be the same size. A fresh version with unequal group sizes tests whether the weighting principle is understood. Correctness on one convenient example can conceal an overgeneralised method.

Write the next action narrowly

Replace “revise graphs” with “practise distinguishing a coordinate ratio from a ratio of changes on lines with non-zero intercepts”. Replace “be careful with algebra” with “retain x = 0 when an equation has a common factor x”. Replace “do more probability” with “write the second-stage composition after a draw without replacement”.

These narrower actions make later evidence easier to interpret. You can see whether the learner changed the particular decision. A broad chapter label may describe where an error appeared, but it does not explain what should happen next.

Keep the record proportionate and private. A few representative tasks usually provide a more usable teaching conversation than an elaborate spreadsheet of speculative labels. Preserve enough evidence to guide learning, not a permanent catalogue of every mistake a child has ever made.

One result, three different repairs

Use the same invented question with three learners: a price is seventy-two dollars after a twenty per cent discount. The first learner writes 72 × 1.2. The second writes 0.8P = 72 and then P = 57.6. The third writes P = 90 but reports the discount as ninety dollars. All three final responses are wrong, but the first weak decisions differ.

The first learner has selected an increase from the reduced base rather than a reversal of the original decrease. Ask them to identify the amount that represents one hundred per cent. The second learner has formed the correct equation but multiplied where division was needed. Ask which operation undoes multiplication by 0.8, then verify by substitution. The third has recovered the original price but confused the requested quantity or its label. Ask them to distinguish original price, sale price and discount amount.

Giving all three a worksheet headed “Percentages” does not guarantee that any of those distinctions will be addressed. A better next task preserves the relevant contrast. The first learner receives a direct and reverse pair. The second receives simple multiplier equations such as 0.6P = 48 and a request to check the candidate. The third receives one situation with three separate questions about its original, final and changed amounts.

Now imagine each learner succeeds on the follow-up. Record exactly what happened. The first identified the original base with a labelled bar. The second inverted a multiplier without a prompt. The third labelled all three quantities correctly. These observations are encouraging, but they are not identical evidence and should not be compressed into an unsupported claim that all percentage difficulties have disappeared.

At a later session, give an unfamiliar version. A population estimate falls by a stated proportion; a length is scaled; a score is expressed relative to a maximum. Keep the mathematical scope appropriate to what has been taught. Ask whether the repaired distinction survives the changed setting. This later check tests the relationship rather than the memory of a particular bag or price.

The teacher also needs to review their own intervention. If a hint supplied the exact operation, a successful answer cannot show that the learner selected that operation independently. If a question’s heading announced the method, it did not test classification in the same way as an unlabelled item. If the later task used identical numbers, the learner may have remembered the answer. These are ordinary limits of the evidence, not reasons to disregard progress.

A useful teaching record can therefore include a sentence about the task as well as the learner: “This prompt tested reversal with the equation already supplied,” or “This version required the learner to construct the equation from prose.” That small distinction keeps the next decision honest.

25. A repeatable practice rhythm

The practice rhythm below is a proposed way to use this workshop. It is not a scientifically validated schedule with guaranteed results. Adjust it to the learner’s current topics, school workload, available time and response to the tasks. The organising principle is to separate the work of learning a relationship from the work of selecting it among alternatives.

Appointment one: learn one contrast

Choose a pair that addresses an observed uncertainty. A learner who adds a percentage of the reduced price back to recover the original should start with direct versus reverse percentage. A learner who divides by x and loses zero should start with equivalent transformations and domain conditions. Do not select a topic merely because it sounds advanced.

Read the worked pair. Ask the learner to name what changed between the two questions. Then solve one fresh near variation. The teacher may model or prompt at this stage. The point is to build the relationship clearly, not to pretend that independent performance already exists.

Appointment two: practise the decision separately

Present several short prompts and ask only for the first relationship. A student can identify whether a task involves a total ratio, a difference ratio or a changing rate without completing every numerical calculation. This isolates selection from arithmetic and makes the reasoning easier to discuss.

Keep some neighbouring alternatives in the set. A percentage reversal should sit beside a direct increase. A right-triangle question with two known sides should sit beside one with an angle and side. A fraction that permits cancellation should sit beside a sum that does not. The contrast is the point.

Appointment three: mix with other secure skills

Once the learner can explain and execute the relationship in focused work, place it inside a short mixed set. Include some secure topics so that the student has to choose among genuine alternatives. Avoid making every question a version of the newly repaired skill while calling the worksheet mixed.

Doug Rohrer, Robert Dedrick and Sandra Stershic studied interleaved and blocked mathematics practice with 126 seventh-grade students. Their original research paper reports better later test performance for interleaved practice in that study. It supports practising method selection when question types are mixed. It does not establish a universal mixing ratio, a Singapore-specific effect size or a requirement to remove focused practice before a new method is understood.

Appointment four: return after a delay

Use a fresh question later, without placing the earlier solution beside it. Record whether the learner recognises the relationship and retains its conditions. A correct response after reading the answer immediately is different evidence from a correct response on a later independent task.

Andrew Butler’s 2010 study of repeated testing and transfer found advantages over repeated study in experiments using prose passages and later inference questions. That is relevant background for taking retrieval seriously, but it is not a direct trial of this Mathematics programme. The practical proposal here is modest: test a fresh decision later rather than assuming an immediate correction proves durable learning.

Appointment five: review the evidence and choose again

If the later task succeeds, increase variation or place the relationship into a multi-step problem. If it fails, inspect where. Do not automatically repeat the entire explanation. A student may now choose correctly but mishandle a negative sign. That is a different next job from failing to recognise the method at all.

For example, a learner first chooses the wrong base in a reverse-percentage problem. After teaching, the learner writes 0.92P = 124.20 correctly but divides inaccurately. The relationship has improved even though the final answer is wrong. Recognise that progress while repairing the remaining execution error.

Make practice small enough to review properly

A large set is useful only if its evidence can be acted upon. Six carefully reviewed questions can reveal more about a learner’s decisions than thirty completed questions marked only with ticks and crosses. This is an instructional judgement about how to use limited study time, not a claim that six is an optimal research-established number.

Keep one page with the current contrast, one corrected example, one sentence stating the decisive condition and one later attempt. Avoid copying a whole chapter into the error log. The record should be easy to revisit when the same structure appears in another topic.

Do not turn a practice rhythm into a punishment

When a learner struggles, reduce unnecessary complexity before increasing volume. Read unfamiliar vocabulary together. Supply a clean diagram. Separate the representation step from the algebra. A shorter supported task can be the appropriate next move when the original task demanded too many uncertain decisions at once.

Support should then be reduced deliberately. The teacher can stop naming the method, remove the completed first line, change the context or delay the revisit. The exact adjustment depends on the observed work. A fixed programme that ignores those observations would undermine the purpose of the diagnostic approach.

The rhythm is therefore a loop with a decision at every stage: choose a contrast, learn the relationship, practise selection, mix it, revisit it, inspect the result. Its value lies in what the learner can do differently afterwards, not in completing an impressive calendar of worksheets.

26. The parent and tutor conversation

For a family comparing Secondary 3 Mathematics tuition in Yishun, this workshop offers a concrete way to discuss teaching. Instead of asking only how many worksheets a lesson includes, ask how the tutor distinguishes a selection problem from an execution problem and how the learner will eventually perform without prompts.

Bring representative marked work and the student’s actual subject level and school topic sequence. A broad label such as Secondary 3 is not enough to establish what has been taught, which examination cohort applies or whether the student takes Additional Mathematics separately. The learning plan should begin with those facts rather than assumptions based on an old course title.

Ask for a worked explanation of the diagnosis

A useful conversation might sound like this: “The learner formed the correct equation on two direct tasks but could not form it from the booking description. We are testing whether the difficulty is the fixed-fee language or the translation into algebra.” That statement identifies evidence, uncertainty and the next probe.

A weaker statement is “The student has poor problem-solving ability.” It is too broad to guide a lesson and can easily become a discouraging label. Ask which observable decision needs to change. The tutor should be able to point to a task and explain why the proposed teaching addresses it.

Ask what will be different on the next attempt

If the intervention is simply “more questions”, ask what those questions will vary. Will the unknown change? Will a denominator restriction be introduced? Will the price contain a fixed fee? Will a direct-percentage question sit beside a reverse one? Useful variation tests a relationship, not merely endurance.

Then ask what support will be removed. A learner who can solve only after the teacher says “use simultaneous equations” has shown execution with a prompt, not independent method selection. That can be a legitimate stage of learning, but it should not be reported as the final outcome.

Keep feedback specific and respectful

A student who repeats an error needs accurate feedback, not humiliation. “You used the reduced price as the base” identifies a mathematical decision. “You never think” makes an unsupported judgement about the person. The first statement creates an action; the second can close the conversation.

Likewise, do not praise an answer as proof of complete mastery after one successful attempt. A more useful response is “You retained the zero case this time; now let us test it when the common factor looks different.” Recognition of progress can be precise without becoming an inflated promise.

Preserve the learner’s own explanation

Before a parent or tutor supplies a correction, ask the student what they intended. A crossed-out equation may contain a useful idea. A wrong final number may follow a correct model. A blank might mean the topic has not been taught. Listening first prevents a correction from targeting the wrong problem.

This does not mean leaving a learner indefinitely stuck. When teaching is needed, teach. Explain the relationship, model the step and provide a manageable attempt. The important distinction is to make the support visible and later test what the learner can do without it.

Separate educational guidance from service claims

This article describes learning decisions and provides original practice. It does not confirm a current class opening, location, fee, timetable, tutor assignment or guaranteed result. Those details should be checked through the centre’s current enquiry route. Historical wording elsewhere in an archive should not be treated as a current service promise.

For the learning conversation, the useful closing question is simple: “What specific decision will the student be better able to make, and what fresh task will show that?” A tutor who can answer that clearly is making the proposed work understandable to both student and parent.

27. Answers to the starting decision check

Return to your original attempt before reading these discussions. Compare the first relationship, not only the final answer. Several valid methods may reach the same result. The important requirements are to preserve the stated conditions, use a justified transformation and answer the requested quantity.

Decision check 1: recover the original base

Sixty-eight dollars represents eighty-five per cent of the original price. Let the original be P dollars. Then 0.85P = 68, so P = 80. Applying the discount forward gives 80 − 0.15 × 80 = 68, confirming the result. The original amount is the base of the discount.

A learner who adds fifteen per cent of sixty-eight has not reversed the given operation. The next teaching step should focus on what represents one hundred per cent. A labelled bar, equation or multiplier can all make that relationship visible; the choice of representation is less important than preserving the base.

Decision check 2: remove the fixed component

The five-hour bill is twelve dollars higher than the two-hour bill. Those three additional hours cost four dollars each. The fixed fee is seventeen minus eight, or nine dollars. Hence C = 9 + 4h, and an eight-hour booking costs forty-one dollars.

Multiplying the two-hour bill by four would repeat the fixed fee four times. It would model four separate two-hour bookings under that pricing rule, not necessarily one eight-hour booking. That distinction explains why the attractive proportion calculation is inappropriate.

Decision check 3: keep the outputs separate

The expression 3(2x − 5) + 4 simplifies to 6x − 11. It remains an expression because the task supplies no equality selecting one value of x. The separate equation becomes 6x − 11 = 19, so 6x = 30 and x = 5.

Substitution checks the equation: 3(10 − 5) + 4 = 19. It does not turn the simplification task into a numerical question. A learner who writes x = 5 for both tasks needs to revisit the difference between rewriting an expression and solving a condition.

Decision check 4: use count and value

Let x be the count of two-dollar tokens and y the count of five-dollar tokens. Then x + y = 14 and 2x + 5y = 46. Subtracting twice the count equation from the value equation gives 3y = 18, so y = 6 and x = 8.

Both checks matter: eight plus six is fourteen, and sixteen dollars plus thirty dollars is forty-six. The units distinguish the two equations. A count cannot be added directly to a money total without a price-per-token relationship converting between them.

Decision check 5: compare changes, not coordinates from zero

From A to B, the vertical change is −3 − 9 = −12 and the horizontal change is 4 − (−2) = 6. The gradient is −2. The sign is consistent with the line falling as x increases. Reversing both point orders gives the same result.

The ratio 9/(−2) describes the gradient of the line joining the origin to A. The question does not state that the given line passes through the origin. In fact its equation is y = −2x + 5, so its y-intercept is five. The difference between a total ratio and a ratio of changes is the central decision.

Decision check 6: convert the dimensional scale

The linear multiplier from smaller to larger is 9/6 = 1.5. Area therefore uses the squared multiplier, 2.25. The larger area is 40 × 2.25 = 90 cm². Moving back gives 90 × (2/3)² = 40 cm².

If the task had asked for a corresponding length, the multiplier would be 1.5 rather than 2.25. If it had supplied an area ratio and asked for a length ratio, a square-root decision would be needed. Name the dimension of both the given and required quantities.

Decision check 7: use the right-angle condition

The two given sides are perpendicular, so they are the shorter sides in a right-angled triangle. The longest side h satisfies h² = 9² + 12² = 81 + 144 = 225. Thus h = 15 cm.

The positive root is appropriate for a length. The result is longer than either perpendicular side, which is consistent with the geometry. The right-angle condition, not merely the presence of two lengths, is what permits Pythagoras in this form.

Decision check 8: update the experiment after the first draw

The first blue probability is 2/6. After a blue is removed, one blue remains among five counters, so the second probability is 1/5. The required probability is (2/6)(1/5) = 1/15.

If the first counter were replaced and the draws were independent under the model, the result would be (2/6)² = 1/9. The changed answer does not mean one formula is unreliable. It means the two experiments have different conditions. The method must describe the experiment that was actually specified.

Choose the next chapter from a precise observation

If the first relationship was correct but the arithmetic was not, repair the execution. If the relationship changed when the wording changed, return to the relevant contrast family. If a candidate failed an original restriction, practise the final interpretation check. If the topic was not yet taught, arrange the necessary teaching before treating the item as a performance test.

The aim is not to memorise these eight answers. It is to recognise the distinctions they represent when the numbers, objects and presentation change. Keep one sentence from each repaired task stating the condition that makes the method appropriate.

28. Further reading and evidence notes

Continue with a purpose. Use the Secondary Mathematics directory to find a level or topic, or return to the Mathematics Tuition & Learning Hub for the wider subject routes. These are the existing library entrances; this workshop is a practice companion, not a replacement hub.

For the separate Additional Mathematics pathway, use the Additional Mathematics Hub. For a broader problem-first discussion, use the Student Diagnostics Hub. To browse beyond Mathematics, return to The Library. Directory labels may retain historical wording, so read an article’s current scope rather than inferring it solely from its dated URL.

Official curriculum references

SEAB: Secondary Education Certificate; SEAB: 2027 G3 syllabuses for school candidates; and the 2027 K310 Mathematics syllabus. Use the official document for your actual examination year and subject level. This article’s practice sequence is an editorial teaching sequence, not an official order of school chapters.

Learning-research references

Rittle-Johnson, B., and Star, J. R. (2007), Does Comparing Solution Methods Facilitate Conceptual and Procedural Knowledge? An Experimental Study on Learning to Solve Equations. The study is discussed with its scope in Chapter 20.

Rohrer, D., Dedrick, R. F., and Stershic, S. (2015), Interleaved Practice Improves Mathematics Learning. The linked author manuscript and the journal publication concern the same study; they are not independent evidence sources. Its teaching relevance and limits are discussed in Chapter 25.

Butler, A. C. (2010), Repeated Testing Produces Superior Transfer of Learning Relative to Repeated Studying. This research used prose-learning tasks, so its findings are not presented as a direct evaluation of this Mathematics workshop.

All mathematical situations, comparison tasks and answer discussions here are original teaching constructions. The observation record and practice rhythm are proposed instructional tools, not validated assessments or promised outcomes. Current fees, class availability and programme arrangements are outside this article’s evidence scope.

The next useful move: choose one question you could not classify, state the condition you missed, practise a contrasting pair, and return later to a fresh version. A method becomes useful when you can recognise its reason for being there.

Return to the reading routes