Secondary Mathematics Tuition | How to Show Working in Maths: Clear Steps, Notation and Mathematical Communication

Secondary Mathematics Tuition · Clear working, notation and mathematical communication

Adrian gets the answer right.

His page does not show how.

There is one equation, then a large jump, then the final value.

If the answer is correct, the page looks efficient.

If the answer is wrong, nobody can see where the route broke.

Clara has the opposite problem.

She writes every tiny arithmetic fact, repeats the question in full, explains obvious steps in sentences and turns a six-line solution into half a page.

Her Mathematics is visible.

It is also difficult to scan.

Ben is learning a third way.

He writes enough to preserve the mathematical state, justify the important decisions, keep risky transitions inspectable, carry units and conditions, make checking possible and let another reader reconstruct the route.

He does not write everything.

He writes what the Mathematics needs.

This worldwide guide is about that skill.

Showing working in Mathematics is not the same as writing more.

It is the design of a visible mathematical route.

The route should be clear enough that the learner, tutor, teacher or examiner can answer:

What was the target?

What representation was chosen?

What relationship or method was used?

What changed from one line to the next?

Which conditions remained in force?

Where could an error be located?

How was the final result interpreted?

This page owns the world-facing Secondary Mathematics learner job of deciding what working should remain visible and how to communicate it clearly. It covers algebra, equations, geometry, graphs, probability, statistics, modelling, calculator work, exactness, units, explanation, proof-like reasoning, compression and examination readability across school systems.

The ownership boundary is deliberate. Singapore School Mathematics: Command Words, Answer Forms and Task Contracts remains the Singapore-specific command-word owner. SEC Mathematics Question Reading remains the SEC G1/G2/G3 question-reading owner. Secondary 4 Additional Mathematics Mathematical Communication remains the stage-specific A-Math owner. Secondary 1 Mathematical Communication, Working & Checking remains the Secondary 1 owner. Calculator, Formula Sheet & Essential Working in SEC Mathematics remains the Singapore SEC tool-and-working owner.

This article also does not replace How to Improve Maths Accuracy, How to Check Maths Answers or How to Get Faster at Maths. Those own error prevention, verification and fluency. Here, the question is narrower:

What should the learner put on the page so that the Mathematics remains valid, readable, checkable and recoverable without becoming unnecessarily long?

Adrian, Clara and Ben are fictional recurring learners. Their scenes are explanatory examples, not reported student records or fixed ability labels.

The core working principle is:

WRITE THE STATE → SHOW THE DECISION → PRESERVE THE INVARIANT → EXPOSE THE RISKY TRANSITION → INTERPRET THE RESULT → COMPRESS ONLY WHAT REMAINS RECONSTRUCTABLE.

This is not a compulsory format for every problem. A one-step calculation may need one line. A proof, modelling task or long algebraic derivation may need many. Good mathematical communication is proportional to the job.

1. Working is external mathematical state, not decoration around the answer

Long problems create state.

Which variable represents what?

Which equation is current?

Which values are exact?

Which condition has already been used?

Which line is still trusted?

Which unit belongs to the current quantity?

If all of that stays in the learner’s head, working memory must keep carrying it.

A written page can carry some of that state instead.

Consider:

4(2x − 3) + 5 = 29.

A compressed mental route might be:

8x − 12 + 5 = 29, so 8x = 36, so x = 4.5.

If the learner writes only x = 4.5, the result is visible but the state changes are not.

A useful written route is:

4(2x − 3) + 5 = 29

8x − 12 + 5 = 29

8x − 7 = 29

8x = 36

x = 4.5.

Why is this useful?

The expansion is visible.

The collection of constants is visible.

Equality is preserved line by line.

If x = 4.5 later fails a check, the learner can locate the first suspicious transition.

Now imagine the learner writes:

4(2x − 3) + 5 = 29 = 8x − 7 = 36 = 4.5.

This is not just ugly notation.

It says mathematically that all those expressions are equal to one another, including 29 = 8x − 7 = 36, which is not the intended chain.

Working is therefore part of the Mathematics itself.

Notation makes claims.

Line breaks make structure.

Definitions preserve meaning.

Brackets preserve grouping.

Units preserve quantity type.

Domain conditions preserve validity.

A page with good working is a temporary memory and verification system.

That is why “show your working” should not be interpreted as “write more because the teacher wants to see it”.

The deeper instruction is:

leave enough mathematical state on the page that the route can be inspected, trusted, repaired and communicated.

2. The answer and the route are different mathematical objects

An answer tells us where the Mathematics ended.

Working tells us how the result was justified.

Sometimes the answer is enough.

If the task is 17 + 28, writing 45 may be sufficient.

Sometimes the route is essential.

If the task is to solve an equation, prove a statement, construct a model, compare two methods, justify a probability or show why a result follows, the route carries mathematical content.

Take:

Solve x² − 7x + 12 = 0.

Answer-only:

x = 3 or 4.

Route-visible:

x² − 7x + 12 = 0

(x − 3)(x − 4) = 0

x = 3 or x = 4.

The route shows why the roots arise.

It also shows that factorisation, not numerical guessing, generated them.

Now take:

A bag contains 5 red and 3 blue counters. Two are drawn without replacement. Find the probability both are red.

Answer-only:

5/14.

Route-visible:

P(RR) = 5/8 × 4/7 = 20/56 = 5/14.

The second fraction shows that the state changed after the first draw.

That is not clerical detail.

It is the mathematical reason the probability is correct.

Working is especially valuable when multiple wrong routes can accidentally land on a correct answer.

A learner might make two compensating arithmetic errors.

The final number can still be right.

Without the route, neither the learner nor the teacher knows whether the underlying Mathematics is secure.

This creates an important rule:

the more a task depends on relationships, conditions or reasoning, the less trustworthy an answer becomes when isolated from the route.

Good working therefore protects learning, not only marking.

3. Essential working is the minimum visible route that preserves meaning, validity and recoverability

Students often ask:

“How much working do I need to show?”

There is no single line count.

A useful answer is based on function.

Working is essential when removing it would make one of these harder to see:

representation — what the symbols stand for;

method — what relationship is being used;

validity — why the transformation is legal;

condition — what restriction or assumption matters;

state — what values or expressions are current;

risk — where an error could propagate;

interpretation — what the final result means.

Consider reverse percentage.

An item costs 144 after a 20% discount. Find the original price.

One possible line is:

144 ÷ 0.8 = 180.

That may be enough for a learner who clearly understands the relationship.

But during learning, the stronger working is:

Final price = 80% of original

0.8P = 144

P = 180.

The first line makes the base relationship explicit.

Now consider geometry.

A right triangle has hypotenuse 10 and one leg 6. Find the other leg.

A compact route:

x² + 6² = 10²

x² = 64

x = 8.

The diagram or earlier statement must make clear that 10 is the hypotenuse and the triangle is right-angled.

If those conditions are not visible anywhere, the algebra alone hides method authority.

Essential working therefore depends on context.

A diagram can carry state that would otherwise need words.

A labelled variable can reduce later explanation.

A table can replace repeated sentences.

A correct symbolic chain can communicate more efficiently than prose.

The aim is not maximal visibility.

It is sufficient visibility.

4. Mathematical notation is not shorthand alone; every symbol makes a claim

Weak working often comes from treating notation as decoration.

But symbols have logical jobs.

The equals sign says two expressions have the same value.

An implication arrow says one statement follows from another under the stated conditions.

An approximation sign says exact equality is not being claimed.

Brackets control grouping.

Units classify the quantity.

Domain restrictions define where an expression is valid.

Consider:

1/3 = 0.333.

That is not exactly true if 0.333 means a terminating decimal.

A better statement is:

1/3 ≈ 0.333

or

1/3 = 0.333…

depending on the context.

Consider:

√49 = ±7.

In standard school notation, the principal square root √49 is 7.

The equation x² = 49 has solutions x = ±7.

Small notation differences change the mathematical claim.

Consider:

(x² − 9)/(x − 3) = x + 3.

This needs the original restriction x ≠ 3.

After factorisation and cancellation, the simplified expression may look unrestricted, but the original function or expression was not defined at x = 3.

Good working preserves that state.

Notation discipline improves accuracy because it forces the page to say what the learner means.

It improves communication because another reader does not have to guess.

It improves checking because contradictions become visible.

The rule is simple:

do not write a symbol merely because it is conventional; write it because its mathematical claim is true.

5. Line breaks should follow mathematical transitions, not arbitrary page space

A good line break answers:

What changed?

One equation became an equivalent equation.

One expression changed form.

A value was substituted.

A condition was applied.

A quantity was interpreted.

When several high-risk changes happen in one line, the working becomes hard to inspect.

When every tiny arithmetic fact gets its own line, the route becomes noisy.

Consider:

2(3x − 5) − 4 = 18.

Clear:

6x − 10 − 4 = 18

6x − 14 = 18

6x = 32

x = 16/3.

Too compressed for a learner with sign risk:

2(3x − 5) − 4 = 18 → x = 16/3.

Too expanded for most secondary learners:

2 × 3x = 6x

2 × (−5) = −10

−10 − 4 = −14

18 + 14 = 32

32 ÷ 6 = 16/3.

Expanded working can be appropriate temporarily when a component is being repaired.

It should later compress.

The page should show transitions at the resolution needed by the learner.

Ben’s sign work is stable, so he can combine some steps.

Adrian still drops negatives, so his working keeps distribution visible.

Clara is fluent in direct equations but expands her working again when algebraic fractions introduce new risk.

This produces a useful principle:

the more fragile the transition, the more visible it should remain.

6. Equation working should preserve equivalence so every line remains mathematically accountable

Solving equations is one of the clearest places to see why good working matters.

The learner is not merely chasing a value.

They are producing a sequence of equivalent statements whose solution sets match.

Consider:

5x − 7 = 18.

Useful working:

5x = 25

x = 5.

The first line records the effect of adding 7 to both sides, even if the operation is not written in prose.

Now consider:

3(x − 4) = 2x + 7.

Useful working:

3x − 12 = 2x + 7

x − 12 = 7

x = 19.

Each line remains an equation.

The equals sign connects quantities that really are equal.

A common weak style is:

3(x − 4) = 2x + 7

3x − 12 − 2x

= 7

x = 19.

The middle line no longer states clearly what the expression equals. The learner may know what they intended, but the page no longer preserves the full state.

Another weak style uses arrows for everything:

3(x − 4) = 2x + 7 → 3x − 12 = 2x + 7 → x = 19.

This can be readable, but arrows can hide what logical relationship is being asserted. Ordinary aligned equations are usually cleaner for straightforward equivalence transformations.

For inequalities, the working must preserve order conditions.

Example:

−3x > 12

x < −4.

A learner who writes only x = −4 has changed the mathematical object entirely.

The inequality sign is part of the state.

For equations with denominators, working should expose any restrictions before simplification if those restrictions matter.

Example:

(x + 2)/(x − 1) = 3, x ≠ 1.

x + 2 = 3(x − 1)

x + 2 = 3x − 3

5 = 2x

x = 5/2.

The excluded value x = 1 never becomes relevant to the final solution here, but recording the restriction establishes legitimate manipulation.

Equation working should make three things visible:

the current equation;

the transformation;

the surviving condition.

When those are visible, checking becomes cheaper and repair becomes local.

7. Algebraic manipulation should show form changes without pretending that every form serves the same purpose

Algebra is not only about obtaining a final answer.

Often the goal is to change form.

Expand.

Factorise.

Simplify.

Rearrange.

Complete the square.

Substitute.

Each form exposes different structure.

Good working makes the transformation traceable.

Consider factorising:

x² − 7x + 12.

One concise route is:

x² − 7x + 12

= (x − 3)(x − 4).

If the factorisation is fragile, an intermediate reasoning line can help:

Need two numbers with product 12 and sum −7: −3, −4.

Then:

x² − 7x + 12 = (x − 3)(x − 4).

That sentence is temporary scaffolding, not a permanent requirement.

Consider completing the square:

x² − 6x + 5

= x² − 6x + 9 − 9 + 5

= (x − 3)² − 4.

The inserted +9 and −9 are visible.

The equivalence can be checked by expansion.

Consider algebraic fractions:

(x² − 9)/(x² − x − 6).

Factor both parts:

[(x − 3)(x + 3)] / [(x − 3)(x + 2)].

Then:

= (x + 3)/(x + 2), with original restrictions x ≠ 3 and x ≠ −2.

The cancellation is visible because common factors are visible.

Without factorisation, crossing out x² or 3 by appearance would be invalid.

Algebraic working should answer:

What form did I start with?

What form am I aiming for?

What equivalence justifies the transition?

What conditions survive?

That structure prevents algebra from becoming a sequence of symbol movements without meaning.

8. Geometry working should connect the diagram, the theorem condition and the calculation

Geometry communicates partly through diagrams.

That means the working is distributed across the page.

A labelled diagram can carry more information than several sentences.

But the diagram should not be treated as self-justifying.

Suppose a right triangle has hypotenuse 13 and one leg 5.

A clear route is:

x² + 5² = 13²

x² = 144

x = 12.

The diagram should make the right angle and hypotenuse relationship clear.

If the right angle is not given or proved, the calculation is not justified merely because the triangle looks right-angled.

For trigonometry:

sin 35° = x/10

x = 10 sin 35°

x ≈ 5.74.

The diagram should identify the reference angle and the side roles.

For similarity, the working should preserve correspondence.

If triangles ABC and DEF are similar, writing:

AB/DE = AC/DF

is meaningful only if the correspondence A↔D, B↔E, C↔F has been established.

Good working may therefore include:

△ABC ~ △DEF

before the ratio.

For angle geometry, short reasons can be useful:

∠ABC = 65° (alternate angles)

or

∠AOC = 2∠ABC (angle at centre).

The exact theorem vocabulary depends on curriculum, but the communication principle is general:

write enough of the geometric reason that the calculation is not detached from the conditions that authorise it.

Do not narrate the whole diagram in prose if notation and markings already carry the information.

Use the diagram as working.

But make sure the diagram carries evidence, not assumptions.

9. Graph working should show scale, coordinates, algebraic relationships and interpretation as one connected system

Graph questions often produce pages where the learner writes almost nothing because the graph itself seems to contain the work.

Sometimes that is enough.

Sometimes important decisions disappear.

If gradient is required from two points, a clear route is:

m = (y₂ − y₁)/(x₂ − x₁)

= (−4 − 6)/(4 − (−1))

= −10/5

= −2.

The point order remains consistent.

The negative result can be compared with the graph falling from left to right.

If an equation of a line is required:

y = mx + c

m = 3

using (2, 11): 11 = 3(2) + c

c = 5

so y = 3x + 5.

This is more informative than writing y = 3x + 5 directly, even if the learner saw the answer visually.

For graphical estimates, notation should distinguish exact from approximate results.

x ≈ 2.4

is more appropriate than x = 2.4 when the value was read from a graph.

Graph working should also carry context when relevant.

If gradient represents 3 dollars per item, write the unit or interpretation once:

gradient = 3 dollars/item.

If the intercept represents an initial fee, state it.

Graphs are representations, not pictures detached from units and variables.

A clear working page connects:

axis meaning;

scale;

selected points;

calculation;

equation;

interpretation.

10. Probability and statistics working should show the state and aggregation logic before the arithmetic

Probability can produce deceptively short answers.

The challenge is often not multiplication or addition.

It is the event structure.

Consider two draws without replacement from 5 red and 3 blue counters.

P(two red) = 5/8 × 4/7 = 5/14.

The changed numerator and denominator show the state transition.

If the learner writes only 5/14, the crucial logic disappears.

For “at least one” events, a complement route should be visible:

P(at least one success)

= 1 − P(no successes).

That line explains the strategy before the arithmetic.

For statistics, the aggregation relationship should be visible.

Group A: n = 8, mean = 15 → total = 120.

Group B: n = 12, mean = 21 → total = 252.

Combined mean = (120 + 252)/(8 + 12)

= 372/20

= 18.6.

This working shows why a simple average of 15 and 21 is not being used.

For data interpretation, working may be verbal rather than algebraic.

“Median is unchanged, but the range increases from 8 to 14, so the centre is stable while spread increases.”

The communication should match the task.

The underlying principle is:

show the probability state or statistical aggregation rule that gives the arithmetic meaning.

11. Modelling working should show the bridge from the real situation to the mathematical model and back again

Modelling problems are especially dependent on visible working because the hardest step is often not the algebra.

It is deciding what the symbols mean.

Consider:

A storage company charges a fixed fee of 12 dollars plus 3 dollars per day.

A weak solution might jump straight to:

12 + 3d.

A stronger route defines the quantity:

Let d = number of days.

Total cost C = 12 + 3d.

This small definition matters.

It gives the symbols meaning.

If the question later asks when the cost reaches 42 dollars, the working becomes:

42 = 12 + 3d

30 = 3d

d = 10 days.

The final unit and interpretation return the result to the original situation.

Suppose the model is only valid for 0 ≤ d ≤ 30 because the storage contract lasts at most one month.

That domain is part of the working too.

Without it, the equation may be algebraically valid but physically misused.

Working in modelling should therefore expose:

the variable definition;

the relationship;

the assumptions;

the domain;

the solution;

and the interpretation.

Adrian tends to skip definitions because he believes the symbols are obvious.

They are obvious to him while the problem is fresh.

Two pages later, x may mean something else.

Good working protects the reader from symbol drift.

Clara over-writes models in sentences.

She learns that one variable definition and one clean equation can replace several lines of prose.

Ben learns that the model boundary is not optional decoration.

It determines where the equation still describes the world.

Model working should make it possible to ask:

What does this variable mean?

Why this equation?

Where is the model valid?

What does the numerical answer mean in the original context?

That is mathematical communication from world to model and back again.

12. Units should travel through the working whenever they are doing mathematical work

Units are not labels added only at the end.

They can reveal whether the calculation itself makes sense.

Consider speed:

distance = 540 m

time = 2 min = 120 s

speed = 540/120 = 4.5 m/s.

The conversion line matters because the denominator unit changes from minutes to seconds.

Without it, a learner may divide 540 by 2 and write 270 m/s.

The numerical arithmetic is correct for 540 ÷ 2.

The quantity calculation is wrong.

For area:

length = 12 cm

width = 8 cm

area = 12 × 8 = 96 cm².

The squared unit communicates dimension.

For volume, the unit becomes cubic.

For rates, compound units matter:

dollars/hour;

km/h;

grams/cm³;

people/km².

Working with units can also catch impossible operations.

You cannot add 5 metres to 3 square metres and obtain a meaningful physical quantity.

You can add 5 metres to 3 metres.

Dimension acts as a check.

Adrian writes units only on the final line.

That is often fine for pure algebra.

It becomes risky in multi-stage applied problems where conversions occur.

Clara writes units on every single intermediate numerical symbol, making the page heavy.

She learns a balanced rule:

carry units at transitions where quantity type or conversion matters;

preserve the final unit clearly;

do not repeat a unit mechanically where it adds no new information.

Ben uses units as a verification tool.

If the question asks for a rate and his final answer is in dollars rather than dollars per item, he knows the route is incomplete.

The working rule is:

show units where they protect meaning, conversions, dimensions or interpretation.

13. Exactness and approximation should be visible so the page does not quietly change the kind of answer being produced

Many Secondary Mathematics problems move between exact and approximate values.

The working should show when that change occurs.

Consider:

x = √55.

If the problem accepts an exact answer, stop there.

If a decimal is required:

x = √55 ≈ 7.416…

so x ≈ 7.42 to 3 significant figures.

The approximation sign makes the status clear.

Now consider a multi-stage problem.

If an intermediate value is 7.416198… and the learner rounds it to 7.4 too early, later calculations may drift.

Good working often preserves exact form or sufficient calculator precision until the final stage.

For example:

area = (1/2)(√55)(12)

= 6√55

≈ 44.5.

This is cleaner than converting √55 to 7.4, multiplying and then rounding again.

Exactness also matters with fractions.

2/3 is not the same object as 0.67 if exact equality is required.

π is not 3.14.

3.14 is an approximation to π.

Working should make the transition deliberate.

Use one of these states:

exact: fraction, surd, π, symbolic form;

calculator display: more digits than needed;

reported approximation: rounded to the requested precision.

Adrian tends to round early because decimals feel easier.

Clara keeps everything exact even when the final context requires a practical decimal.

Ben learns to preserve exactness until approximation becomes part of the answer contract.

This supports communication because the reader can see where information was lost.

It supports checking because unexpected discrepancies can often be traced to premature rounding.

The working principle is:

do not let exact mathematics become approximate by accident.

14. Calculator working should show the mathematical expression before or around the button sequence

A calculator can evaluate a correct expression perfectly and still produce the wrong mathematical answer if the expression entered is wrong.

That is why calculator work needs a paper trail at the right level.

Suppose the intended quantity is:

(12 + 8)/5.

If the learner types 12 + 8 ÷ 5, the calculator may return 13.6 rather than 4.

The calculator is not wrong.

The mathematical grouping was not entered faithfully.

Good working can show:

(12 + 8)/5 = 4.

The paper expression becomes the source of truth.

For trigonometry:

x = 10 sin 35°.

Writing the expression before pressing buttons helps the learner notice:

which function is being used;

which angle enters;

which value multiplies;

and whether degree mode is appropriate.

For a long expression:

x = [−b ± √(b² − 4ac)]/(2a),

substitute coefficients on paper first.

If a = 1, b = −9, c = 17:

x = [9 ± √(81 − 68)]/2.

Only then use the calculator if a decimal is required.

This working separates algebra from device execution.

If the output looks wrong, the learner can ask:

Was the formula substituted correctly?

Was the grouping entered correctly?

Was the mode correct?

Was the result copied accurately?

Ben often writes only calculator outputs.

When one is wrong, the tutor cannot tell what expression was entered.

After adding the intended expression to the page, recovery becomes much cheaper.

Clara writes every button press.

That is usually unnecessary unless the calculator procedure itself is being learned.

The right communication level is:

show the mathematical expression; let the calculator carry the routine arithmetic.

15. Some working should be verbal because not every mathematical reason can be compressed safely into symbols

Mathematics is not purely symbolic communication.

Some decisions need words.

Examples include:

why a model is appropriate;

why two triangles are similar;

why a probability state changes;

why one summary statistic is more appropriate;

why a root is rejected;

why a theorem condition holds;

why a result is unreasonable.

Consider:

x² = 25.

Mathematical solution:

x = ±5.

If x represents a physical length, the final working may say:

x = 5 cm, rejecting −5 because length is positive.

The rejection needs a contextual reason.

Consider a weighted mean.

If two group sizes differ, the explanation:

“A simple average of the two means is invalid because the groups have unequal sizes”

makes the decision explicit.

Consider a graph.

“The negative gradient means the quantity decreases by 2 units for every 1-unit increase in x.”

This communicates meaning that the symbol m = −2 does not fully express.

Good verbal working is concise.

It should explain the reason, not narrate every action.

Weak:

“First I looked at the equation and then I decided to subtract 7 because I wanted to get x on its own.”

Stronger:

“Subtract 7 from both sides.”

Or simply show the equivalent equation.

Weak:

“I used Pythagoras because I saw a triangle.”

Stronger:

“The triangle is right-angled, so Pythagoras applies.”

Words should carry the mathematical reason that symbols cannot safely carry alone.

The result is communication that is neither silent nor over-written.

16. Good working creates a return path when something goes wrong

One of the strongest reasons to show working has nothing to do with presentation.

It is recovery.

A long solution is a journey through mathematical states.

If the final answer fails a check, the learner needs to know where to return.

Suppose:

3x + 2y = 16

x + 2y = 8.

Subtracting gives:

2x = 8

x = 4.

Then substituting into x + 2y = 8:

4 + 2y = 8

2y = 4

y = 2.

Check:

3(4) + 2(2) = 16.

Everything is visible.

Now imagine the learner instead writes:

x = 4, y = 3.

The check fails.

Where did 3 come from?

There is no return path.

Good working creates trusted checkpoints.

After elimination, x = 4 is one trusted state.

After substitution, 2y = 4 is another.

If y is wrong, the learner can return to the last correct line rather than restart the whole problem.

This matters even more in long modelling, geometry or probability questions.

A learner may spend several minutes building a route.

Restarting every time is expensive.

Visible working lowers recovery cost.

Use a recovery rule:

when a check fails, return to the last line whose meaning and validity you still trust.

Then inspect the next transition.

Was a number copied wrongly?

Was a sign changed?

Was a theorem applied without its condition?

Was a unit converted incorrectly?

Was the calculator expression grouped wrongly?

Adrian used to erase large sections immediately.

After learning to preserve line-by-line state, he repairs locally.

Clara used to use arrows and scraps of arithmetic around the page.

She now aligns a main route so the trusted state is easy to find.

Ben adds a small side calculation when needed but brings the result back into the main chain explicitly.

Working should help the learner answer:

Where am I?

What do I still trust?

What changed next?

That is fault-tolerant mathematical communication.

17. Side working is useful when it supports the main route without fragmenting it

Not every calculation belongs in the main chain.

Some problems need side working.

Factor pairs.

Quick estimates.

Small substitutions.

Angle calculations.

Temporary arithmetic.

A diagram sketch.

The danger is page fragmentation.

If side calculations are scattered without labels, the final route becomes difficult to reconstruct.

Suppose a learner is solving:

x² − 11x + 24 = 0.

Side working:

factor pairs of 24:

1,24;

2,12;

3,8.

Need sum 11 → 3 and 8.

Main route:

(x − 3)(x − 8) = 0

x = 3 or x = 8.

The side work supports the decision but does not break the main solution.

In geometry, a side calculation may find an angle:

180° − 72° − 48° = 60°.

That value should then be marked clearly on the diagram or used explicitly in the next relationship.

In probability, a small complement calculation may sit beside a tree, but the main event relationship should remain readable.

Use three rules.

Label side working when its purpose is not obvious.

Bring its result back into the main route.

Do not let side working replace the main mathematical narrative.

Clara’s old pages contain numbers circled everywhere.

She knows what they meant during the lesson but cannot understand them two days later.

After using labelled side work, revision becomes easier because the route survives time.

This is another communication test:

could future-you understand the page without replaying the entire lesson in memory?

18. Working and checking should support each other without becoming the same route repeated twice

Working creates the route.

Checking challenges it.

If the check merely repeats the same working, the same error may survive.

Good working makes alternative checks possible.

Equation solution:

main route isolates x;

check substitutes x into the original.

Factorisation:

main route factors;

check expands.

Reverse percentage:

main route reconstructs original;

check applies the change forward.

Geometry:

main route uses trigonometry;

check inspects magnitude, angle and diagram relationships.

Model:

main route solves algebraically;

check substitutes into the contextual relationship and tests domain.

Working should preserve enough original structure to make this possible.

If the original equation is copied incorrectly or discarded, substitution may check the wrong statement.

If the diagram is not labelled, geometry checking becomes harder.

If units are stripped away, dimensional checking disappears.

If exact values are rounded too early, independent numerical comparison becomes weaker.

Ben uses a small check mark next to high-risk transitions.

Not every line.

Only where checking adds useful confidence.

Adrian learns to keep the original equation visible until the solution is verified.

Clara learns that a second full recalculation is not always the best check.

The working/checking partnership is:

write a route that exposes independent evidence.

The 240 checking owner develops verification fully.

Here, the communication lesson is that good working should leave enough structure for meaningful checking to happen.

19. Fluency should shorten working only after the shorter route remains mathematically transparent

As learners improve, their written work should often become shorter.

That is normal.

A fluent learner does not need to write every multiplication fact or explain every inverse operation.

But compression should follow stability.

Consider:

7x + 3x − 4 = 26.

A beginner may write:

10x − 4 = 26

10x = 30

x = 3.

A fluent learner may write:

10x = 30

x = 3,

mentally combining the like terms and adding 4.

That can be fine if the transition is stable.

Now consider:

−4(2 − 3x) + 5 = 29.

A learner who often loses signs should not compress the expansion.

Write:

−8 + 12x + 5 = 29

12x − 3 = 29

12x = 32

x = 8/3.

Compression has a risk profile.

Safe candidates often include:

simple arithmetic;

stable collection of like terms;

well-practised fraction simplification;

routine substitution.

High-risk candidates often include:

negative distribution;

changing inequality direction;

domain restrictions;

multi-step fraction algebra;

long calculator entries;

unit conversions;

state changes in probability.

A useful compression test is:

If this line were wrong, could I still see what decision was made and recover quickly?

If no, the compression may be too aggressive.

The 340 fluency owner develops automaticity and efficient working more deeply.

Here, the communication rule is:

shorter is better only when it remains inspectable.

20. Independent learners use working to think, not merely to display thinking after the fact

Students sometimes imagine that strong mathematicians solve problems mentally and write the solution only afterward.

Sometimes experts do.

In harder problems, writing is part of the thinking process.

A diagram reveals a relationship.

A table exposes a pattern.

An equation reduces verbal complexity.

A factorisation reveals roots.

A labelled state clarifies probability.

A rewritten expression suggests a method.

Working therefore has a generative role.

Adrian meets an unfamiliar word problem.

He does not know the whole route.

He writes:

Let x = smaller number.

Larger number = x + 7.

Product = 60.

Then:

x(x + 7) = 60.

The writing creates the problem state he can solve.

Clara meets a geometry problem and redraws the relevant triangle separately.

The redraw is not presentation after insight.

It produces the insight.

Ben meets a data problem and reconstructs group totals in a small table.

The table makes weighted aggregation visible.

This changes the meaning of “show your working”.

The learner should not wait until they know the answer and then manufacture a clean route for the page.

Use writing to discover.

Represent.

Test.

Compare.

Recover.

Then clean the final route enough that it remains readable.

The 140 unseen-problem owner develops cue-free startup in depth.

Here, the key connection is:

working is one of the tools independence uses to create the route.

21. The working should lead naturally to the answer form the problem actually requires

A mathematically correct route can still end badly if the final answer is not in the required form.

Working should therefore keep the target visible all the way to the last line.

Suppose the question asks for:

a length;

an angle;

a probability;

a percentage;

a coordinate;

an equation;

a proof;

a statement in context;

or a rounded value.

The final line should answer that object.

Consider:

A rectangle has area 60 m² and width 5 m. Find the length.

Working:

length = area/width

= 60/5

= 12.

Final answer:

12 m.

The unit completes the quantity.

Now consider:

Find the probability as a percentage.

Working may produce:

P = 3/8 = 0.375.

The answer is not finished until:

P = 37.5%.

Now consider:

Give the equation of the line.

If working produces gradient m = 2 and intercept c = −5, the final answer should be:

y = 2x − 5.

Writing only m = 2, c = −5 leaves the target incomplete.

Good working should therefore maintain a small target contract.

What type of object must the final line contain?

What units?

What precision?

What notation?

What contextual interpretation?

Adrian often solves the Mathematics and loses the answer contract at the end.

He now circles the target word or writes a tiny margin cue:

length;

probability;

equation;

nearest integer;

explain.

Clara uses the final line to convert from mathematical result to requested form.

Ben treats that conversion as part of the route, not an afterthought.

This makes mathematical communication complete:

the working should terminate in the object the question asked for.

22. Reasoning and proof need claims, reasons and logical direction—not just more symbols

Some Secondary Mathematics tasks ask for more than calculation.

They ask the learner to justify, explain, establish, show, prove or demonstrate.

In these tasks, working must communicate logical structure.

A useful pattern is:

claim → reason → consequence.

Consider proving that the sum of two consecutive odd integers is divisible by 4.

Let the consecutive odd integers be:

2n + 1 and 2n + 3.

Their sum is:

(2n + 1) + (2n + 3)

= 4n + 4

= 4(n + 1).

Since n + 1 is an integer, the sum is a multiple of 4.

The final sentence matters.

The algebraic factorisation shows the structure, but the conclusion links it to the property being proved.

Consider a geometry reasoning task.

If two angles are equal because they are alternate angles, and another pair is equal because they are vertically opposite, those reasons should be visible before similarity is asserted.

The working may look like:

∠ABC = ∠DEF (alternate angles)

∠ACB = ∠DFE (vertically opposite angles)

Therefore △ABC ~ △DEF (AA).

Exact notation conventions vary, but the logical order should remain clear.

Weak proof working often has one of four problems.

Missing reason: the learner writes a true statement without showing why it follows.

Wrong direction: the learner assumes the conclusion they are meant to prove.

Example instead of proof: the learner checks a few numerical cases and treats them as universal.

Unstated condition: the learner uses a theorem without establishing its hypotheses.

Ben is comfortable with algebraic manipulation but weak at the final logical sentence.

He learns to ask:

“What exactly has the algebra established, and why does that prove the target?”

Adrian writes long prose around proofs.

He learns to let equations carry the algebra and words carry only the logical bridge.

Clara uses counterexamples when a universal claim looks suspicious.

That is also communication: one valid counterexample can show that a universal statement is false.

Good proof working is not necessarily long.

It is logically sufficient.

23. Page layout should make the main route visually obvious without turning presentation into a design exercise

Readable Mathematics benefits from spatial organisation.

Not decoration.

Organisation.

The main route should be easy to find.

Variables should not change meaning halfway down the page.

Side calculations should not collide with the main chain.

Diagrams should be close enough to the calculations they support.

Final answers should be identifiable.

Use a simple page architecture.

Main column: the central mathematical route.

Margin or side area: short side calculations, labels or checks.

Diagram area: one clean representation rather than multiple uncontrolled sketches.

Final line: the answer in the requested form.

Alignment helps.

For equations:

3x + 7 = 25

3x = 18

x = 6.

The equals signs do not have to form a perfect printed column, but the structure should be visually traceable.

For simultaneous equations, label them if substitution or elimination will reuse them:

(1) 2x + y = 11

(2) x − y = 1.

Then working can refer to:

(1) + (2): 3x = 12.

This is clearer than rewriting both equations repeatedly when labels help.

For multi-part questions, separate parts visibly.

Do not let working from part (a) merge into part (b) without indicating which result is being handed forward.

Clara once writes tiny side numbers in every corner.

When she later reviews the page, she cannot tell which belong to which question.

She starts boxing only key handoff results and crossing out abandoned routes lightly rather than erasing all evidence.

Adrian leaves no space between questions and sometimes continues algebra into the wrong answer area.

He begins using one blank line or a clear part label.

Ben’s page is compact but readable because the layout follows the mathematical hierarchy.

The purpose is not neatness for its own sake.

It is lower reading cost.

A good page lets the eye recover the structure faster than the mind could reconstruct it from memory.

24. Examination working should preserve marks, pace and recoverability without becoming slower than the Mathematics itself

In an examination, working has an additional constraint:

time.

The learner cannot write a textbook solution for every routine question.

Nor should they reduce every route to an unsupported final answer.

The practical job is to preserve enough state to earn and protect the solution while maintaining pace.

Use risk-weighted working.

Show more when:

the question has multiple stages;

method choice matters;

a theorem condition matters;

partial progress could be meaningful;

a sign or domain error could propagate;

the calculator expression is long;

or later parts depend on the result.

Show less when:

the operation is routine;

the transition is highly stable;

the calculation is one-step;

and the answer can be independently checked cheaply.

Suppose a learner solves a simple equation.

Three lines may be sufficient.

Suppose a learner solves a modelling problem with two variables and an interpretation.

The variable definitions, equations, route and final contextual answer should stay visible.

Examination working also protects re-entry.

If a learner leaves a difficult question and returns later, a visible state prevents them from starting again from zero.

Write enough to know:

what was established;

what remained unknown;

which route was being attempted.

Adrian used to erase incomplete working before moving on.

He now leaves the trusted state and a small mark indicating where he stopped.

On return, he can continue.

Clara used to write fully polished explanations under time pressure.

She learns to reserve prose for tasks that actually require reasons.

Ben uses compact but interpretable algebra, keeping high-risk transitions visible.

Assessment rules differ across systems and papers, so this article does not claim one universal marking convention.

The transferable principle is:

under time pressure, working should protect the mathematical route without consuming more time than the route is worth.

25. Tutors should teach working as an adaptive control system, not as one fixed house style

A tutor can easily turn “show your working” into a formatting rule.

Always write this line.

Always use three steps.

Always circle the answer.

Always write the formula first.

Some routines are useful.

But a rigid house style can disconnect working from mathematical purpose.

The tutor should instead diagnose what the learner’s page is failing to preserve.

Is the state disappearing?

Are signs being compressed too early?

Are variables undefined?

Are units missing at conversions?

Are reasons absent in geometry?

Is the calculator expression invisible?

Is the learner writing too much and losing pace?

Is side working fragmented?

Is the final answer not returning to context?

Then change the working at the first weak point.

Adrian’s tutor asks him to expand one risky line, not every line.

Clara’s tutor asks her to compress stable arithmetic, not to “write less” everywhere.

Ben’s tutor asks him to preserve a domain restriction he tends to drop.

Working should also fade.

When a transition becomes stable, the learner can compress it.

When a new topic adds risk, working can expand again temporarily.

This is adaptive communication.

The tutor should be able to explain every requested line:

“Keep this because it protects the sign.”

“Define this variable because the model uses it for several lines.”

“Write the expression before the calculator because grouping is the current risk.”

“You can now combine these two arithmetic steps because they are secure.”

That explanation makes working learner-owned rather than teacher-imposed.

The goal is not to produce pages that all look identical.

The goal is to produce learners who know what their Mathematics needs to remain visible.

26. Practice should include working-quality feedback, not only answer accuracy

If working matters only in examinations, it will often collapse under pressure.

Working should be trained during ordinary practice.

Not as a handwriting exercise.

As a mathematical control system.

After a practice set, review two things separately:

Mathematics: was the route correct?

Communication: was enough of the route visible to inspect and recover?

A correct answer with invisible reasoning may need a communication repair.

A wrong answer with a clear route may be easier to fix because the first invalid transition is visible.

Use a working-quality checklist only where useful:

variables defined;

state preserved;

risky step visible;

notation valid;

units/conditions carried;

final answer complete;

check possible.

Do not apply every item to every problem.

A one-line arithmetic question does not need variable definitions or a recovery scaffold.

A modelling problem may need all of them.

Adrian’s practice includes one “working audit” problem per session.

He solves normally.

Then he asks:

Could another reader reconstruct the route?

Where did I compress?

Was that compression safe?

What line would I return to if the answer failed?

Clara practises compression deliberately.

She takes an over-expanded solution and shortens it without deleting the high-risk transitions.

Ben practises expansion deliberately.

He takes a too-short solution and inserts the one or two missing states that make it inspectable.

This makes “show working” a skill with two directions:

expand when the route is opaque;

compress when the route is secure.

Practice should gradually teach the learner to regulate that balance independently.

27. Digital and AI-supported Mathematics still needs a visible learner-owned route

Digital tools change where working happens.

Some state sits on paper.

Some sits in a calculator.

Some sits in graphing software.

Some may be generated by an AI system.

The communication problem does not disappear.

It becomes more important to know which part of the route belongs to the learner.

Suppose an AI system returns:

x = 4.

If the learner copies the answer, there is no visible route.

If the AI returns a full solution and the learner copies every line, the page looks excellent but may still contain little learner-owned Mathematics.

A stronger AI workflow is:

1. learner writes the target and first representation;

2. learner attempts a route;

3. AI is asked for one hint, critique or alternative method;

4. learner closes the output;

5. learner reconstructs the route in their own working;

6. learner solves a changed problem without the AI solution visible.

The written page should show the mathematical decisions the learner can reproduce.

The same principle applies to graphing tools.

If software displays an intersection, the learner should still know what the intersection means in the original problem.

If software produces a regression or table, the learner should still identify variables and interpret parameters.

If a calculator solves an equation, the learner should still understand the equation being solved and check whether the output belongs to the required domain.

Digital tools can carry computation.

They should not silently carry all mathematical meaning.

Working becomes the receipt of learner control.

Not every button press.

The decisions that remain learner-owned.

28. When “show more working” does not improve performance, diagnose what the page is actually missing

More working is not automatically better.

Sometimes a learner already writes many lines and still makes the same mistakes.

The problem may be that the wrong information is visible.

Failure 1 — lots of algebra, no representation

The learner manipulates symbols but the original word problem was modelled incorrectly.

Repair: define variables and relationship before algebra begins.

Failure 2 — every step visible, condition missing

The learner shows a perfect trigonometric calculation on a triangle where the method is not authorised.

Repair: make theorem or method conditions visible.

Failure 3 — units appear only at the end

The learner performs an invalid conversion early and then attaches the correct-looking unit to the final answer.

Repair: expose the conversion transition.

Failure 4 — calculator output replaces mathematical expression

The learner cannot diagnose a grouping or mode error.

Repair: write the intended expression before entry.

Failure 5 — working is too compressed

Errors cannot be localised.

Repair: expand the first historically fragile transition.

Failure 6 — working is too expanded

The learner loses time and hides the main structure in routine detail.

Repair: compress stable arithmetic and keep strategic decisions visible.

Failure 7 — working is scattered

The page contains useful calculations but no main route.

Repair: create a central chain and label side working.

Failure 8 — final answer is disconnected

The algebra ends correctly but does not answer the contextual target.

Repair: add a final interpretation line with unit/form.

Failure 9 — reasons are written after the fact

The learner writes theorem names beside a route they actually chose by visual guessing.

Repair: require condition evidence before the method line.

Failure 10 — neatness is being mistaken for communication

The page is beautifully presented but mathematically ambiguous.

Repair: prioritise valid notation and visible relationships over appearance.

The correct intervention is not “write more”.

It is:

make the missing mathematical state visible.

29. Frequently asked questions about showing working in Maths

Do I need to show every step?

No. Show enough to preserve the route, important decisions, risky transitions, conditions and final interpretation. Stable routine arithmetic can often be compressed.

Can I lose marks for not showing working?

Assessment conventions vary by system and paper. Regardless of marking rules, visible working protects method, checking and recovery. Consult the official instructions for your specific examination.

Is mental Maths bad?

No. Mental calculation is useful when the operation is stable and low risk. Write the result into the route where the next step depends on it. Do not keep too much multi-step state mentally if that increases error or recovery cost.

Should I always write the formula first?

Not always. Write it when doing so clarifies the relationship, protects substitution or helps checking. If the relationship is already obvious and the transition is routine, repeating a formula mechanically may add little.

Should I write units on every line?

No universal rule is needed. Show units where quantity type, conversion or interpretation matters, and ensure the final answer has the correct unit.

Should I circle or box my answer?

A clear final answer is useful, but boxing is a formatting convention rather than Mathematics. Follow local assessment expectations where relevant. Do not let a box replace units, precision or interpretation.

What is wrong with writing only the answer if it is correct?

For simple answer-only tasks, nothing may be wrong. For multi-step or reasoning tasks, the route contains mathematical evidence, supports partial diagnosis and makes checking possible.

How do I make algebra working less messy?

Keep one main equation per state, align equivalent transformations, label side calculations when needed and avoid combining several fragile sign/fraction operations in one invisible jump.

How do I know when I can skip a line?

Skip or combine a transition when it is stable, meaningful and recoverable. If errors recur there or you cannot explain what happened, keep it visible.

What should I show when using a calculator?

Usually show the mathematical expression being evaluated, key substitutions and the final rounded/exact result. Button-by-button logs are rarely necessary unless calculator technique itself is being taught.

Should diagrams count as working?

Yes. A correctly labelled diagram can carry mathematical state, conditions and relationships. But visual appearance alone is not proof; mark only given or derived information.

What about graphs and tables?

They are working when they organise or reveal the relationship. Label axes/quantities and connect the representation to the calculation or interpretation.

How much explanation should I write?

Use words for reasons that symbols do not communicate safely: method conditions, contextual interpretation, model assumptions, rejection of roots, theorem reasons or comparative conclusions.

Should I cross out wrong working?

If the route is abandoned, make that clear so the reader does not treat it as part of the final solution. During learning, preserving enough of the wrong route to diagnose it can be useful. Follow examination rules for your context.

Can neat working improve accuracy?

Organisation can reduce copying, sign, state and re-entry errors. But neatness alone does not guarantee correct Mathematics. The page must preserve valid relationships.

What if writing slows me down?

Compress stable transitions, not essential state. The solution is usually better working, not no working. Under time pressure, use risk-weighted visibility.

How can I practise working?

Take one too-short solution and expand the missing decisions. Take one over-expanded solution and compress safe routine steps. Compare whether both remain checkable and reconstructable.

Does good working look the same for every student?

No. Learners at different stages need different visibility. A student repairing negative signs may show more intermediate algebra than a fluent learner. The communication job stays the same even when the number of lines changes.

Does good working look the same in every country or exam?

No. Notation and assessment conventions vary. This guide focuses on transferable mathematical communication. Official syllabus and examination instructions remain the authority for local requirements.

What is the best rule to remember?

Write enough that the Mathematics can be understood, checked and repaired; compress only what is already stable.

30. The complete working system: make the route visible enough to trust and compact enough to use

Return to Adrian, Clara and Ben.

Adrian began with answers that appeared from nowhere.

His Mathematics was sometimes correct but difficult to diagnose.

He learned to preserve state at the transitions where errors could propagate.

Clara began with pages full of detail.

Her communication was safe but inefficient.

She learned to let stable arithmetic disappear while keeping representation, conditions and risky transitions visible.

Ben learned that working is not merely for other people.

It is a tool for his own thinking, checking and recovery.

The complete system is:

READ THE TARGET → CHOOSE A REPRESENTATION → DEFINE WHAT SYMBOLS MEAN → WRITE THE CURRENT STATE → SHOW THE METHOD OR RELATIONSHIP → PRESERVE EQUALITY, CONDITIONS, UNITS AND EXACTNESS → EXPOSE HIGH-RISK TRANSITIONS → USE SIDE WORK WITHOUT LOSING THE MAIN ROUTE → INTERPRET THE RESULT → CHECK INDEPENDENTLY → COMPRESS ONLY STABLE STEPS → LEAVE A RETURN PATH.

This system does not require every problem to be long.

A short problem should still be short.

A complex problem should not be forced into answer-only compression.

The reader job decides the visibility.

The neighbouring BTT owners remain separate.

How to Improve Maths Accuracy owns prospective error prevention.

How to Check Maths Answers owns verification.

How to Get Faster at Maths owns fluency and efficient execution.

How to Solve Unseen Maths Problems Independently owns cue-free startup and transfer.

How to Build Confidence in Maths owns calibrated self-trust.

Evidence and scope note: the visibility rules, recovery ideas, fictional scenes and working-quality frameworks in this guide are original instructional design tools rather than validated scoring instruments. Public research and professional guidance support related principles: the NCETM Five Big Ideas emphasise representation and structure, mathematical thinking, fluency, variation and coherence; the What Works Clearinghouse problem-solving guide gives strong-evidence recommendations within grades 4–8 for monitoring/reflection and visual representations; the WWC algebra guide includes recommendations about solved problems, algebraic structure and alternative strategies with stated evidence ratings. These sources support bounded teaching principles, not one universal number of lines or one global marking scheme.

Public references: NCETM — Five Big Ideas in Teaching for Mastery; What Works Clearinghouse — Improving Mathematical Problem Solving in Grades 4 Through 8; What Works Clearinghouse — Teaching Strategies for Improving Algebra Knowledge in Middle and High School Students.

Good mathematical working leaves the route visible enough that the answer is not a mystery, the error is not a catastrophe and the next reader does not have to guess what the Mathematics meant.

Appendix A — Working-quality checkpoint: twenty-four cases

This checkpoint is original explanatory material. It is not a universal marking scheme and has no validated cut score. Its job is to test whether a learner can decide what needs to remain visible, identify what is mathematically wrong with weak working, and improve a solution without making it unnecessarily long.

For each case, use four questions:

What mathematical state must be visible?

What is missing or over-written?

What is the smallest useful repair?

What could be safely compressed later?

Task 1 — The disappearing equation

A learner solves 5x − 7 = 18 and writes:

5x − 7 = 18

25

5.

Diagnosis: the numbers are connected in the learner’s mind but not on the page. The equation state disappears.

Repair:

5x − 7 = 18

5x = 25

x = 5.

Why: the intermediate line preserves the unknown and shows the equivalent equation. The final line identifies the solution.

Compression: a highly fluent learner might combine simple arithmetic mentally, but the equation state should not disappear entirely.

Task 2 — Incorrect equals-sign chain

A learner writes:

2x + 3 = 11 = 2x = 8 = x = 4.

Diagnosis: the equals signs claim all neighbouring expressions are equal. In particular, 11 is not equal to 2x as a free expression and 8 is not equal to x.

Repair:

2x + 3 = 11

2x = 8

x = 4.

Why: each line is a complete equivalent equation. The equals sign now carries a true claim.

Transfer: this same issue appears in arithmetic simplification, algebraic rearrangement and formula work. Notation is logical structure, not visual glue.

Task 3 — Reverse percentage with no base relationship

An item costs 144 after a 20% discount. The learner writes:

144 ÷ 0.8 = 180.

Diagnosis: the working can be sufficient for an experienced learner, but if the learner often confuses forward and reverse percentage, the critical relationship is invisible.

Repair for a developing learner:

Final price = 80% of original

0.8P = 144

P = 180.

Why: the working exposes which quantity is the 100% base.

Compression: once base selection is reliable, the first two lines can often compress to 144 ÷ 0.8 = 180 without losing meaning.

Task 4 — Pythagoras with no right-angle evidence

A learner sees a triangle that looks right-angled and writes:

x² + 6² = 10².

Diagnosis: the calculation may be valid only if a right angle is given or derived. Visual appearance alone is not enough.

Repair: mark or state the right-angle condition before the theorem is used.

For example:

∠C = 90°, so x² + 6² = 10².

Why: the method’s authority becomes visible.

Transfer: theorem conditions should be visible at least while method selection is still fragile. Later a correctly marked diagram may carry the condition without extra prose.

Task 5 — Premature approximation

A learner needs the area of a triangle with height √55 and base 12. They write:

√55 = 7.4

Area = 1/2 × 12 × 7.4 = 44.4.

Diagnosis: √55 has been rounded early, and equality is used incorrectly.

Repair:

Area = 1/2 × 12 × √55

= 6√55

≈ 44.5.

Why: exactness is preserved until approximation is actually needed.

Communication lesson: use ≈ when a value is approximate and avoid information loss before the final reporting stage.

Task 6 — Calculator output with invisible expression

A learner writes only:

13.6.

The intended calculation was (12 + 8)/5.

Diagnosis: the calculator result suggests the learner probably entered 12 + 8 ÷ 5. The mathematical grouping was never written.

Repair:

(12 + 8)/5 = 4.

Why: the paper expression becomes the source of mathematical meaning. If the device output disagrees, the learner can inspect entry rather than rebuild the whole problem.

Compression: no button sequence is needed. The expression is the essential working.

Task 7 — Unit conversion hidden inside the answer

A cyclist travels at 4.5 m/s for 2 minutes. The learner writes:

4.5 × 2 = 9 m.

Diagnosis: the time units are incompatible with the rate. The missing state is the conversion.

Repair:

2 min = 120 s

distance = 4.5 × 120

= 540 m.

Why: the conversion line prevents a dimension error.

Transfer: whenever units change mid-problem, make the conversion visible at the transition where it matters.

Task 8 — Weighted mean without the aggregation rule

Group A has 8 values with mean 15. Group B has 12 values with mean 21. A learner writes:

(15 + 21)/2 = 18.

Diagnosis: unequal group sizes make the two means unequally weighted. The current working hides the definition of mean.

Repair:

Total A = 8 × 15 = 120

Total B = 12 × 21 = 252

Combined mean = (120 + 252)/(8 + 12)

= 18.6.

Why: the working exposes the aggregation rule.

Compression: an experienced learner may write directly (8×15 + 12×21)/20 = 18.6, but the weighting must remain visible.

Task 9 — Algebraic fraction with lost restriction

A learner simplifies:

(x² − 9)/(x − 3)

= x + 3.

Diagnosis: the simplified expression is correct only on the original domain x ≠ 3.

Repair:

(x² − 9)/(x − 3)

= [(x − 3)(x + 3)]/(x − 3)

= x + 3, x ≠ 3.

Why: the factor structure and domain restriction are both visible.

Transfer: simplification should not erase the conditions inherited from the original expression.

Task 10 — Side working with no return to the main route

A learner solving x² − 11x + 24 = 0 writes factor pairs in the margin, circles 3 and 8, then writes x = 3 or 8 without showing factorisation.

Diagnosis: the side working found the right numbers, but the main algebraic route is missing.

Repair:

(x − 3)(x − 8) = 0

x = 3 or x = 8.

Why: the factor pair result is brought back into the equation, revealing the zero-product structure.

Communication lesson: side working should support the main route, not replace it.

Task 11 — Graph gradient with points but no scale reading

A learner chooses two plotted points and obtains gradient −2. The graph uses an x-axis scale of 5 units per small square, but the learner read it as 1.

Diagnosis: the arithmetic route is visible but the representation state is wrong.

Repair: record or verify coordinates from the axis scale before applying the gradient formula.

For example:

Points: (5, 8) and (15, −12)

m = (−12 − 8)/(15 − 5)

= −20/10

= −2.

Why: the working shows where the coordinates came from, so scale errors are easier to detect.

Task 12 — Correct root, incomplete contextual answer

A rectangle problem produces:

w² + 7w − 60 = 0

(w + 12)(w − 5) = 0

w = −12 or 5.

The learner stops.

Diagnosis: the algebraic solution set is complete, but the contextual answer is not.

Repair:

Width w = 5 m, rejecting −12 because a physical length must be positive.

If length = w + 7, then length = 12 m.

Why: the final line returns to the original problem and completes the answer contract.

Transfer: contextual constraints can remove mathematically valid algebraic solutions; that interpretation belongs in the working.

Task 13 — An inequality loses its inequality

A learner solves:

−3x > 12

and writes:

x = −4.

Diagnosis: the learner has collapsed a solution set into a single value and lost the order reversal.

Repair:

−3x > 12

x < −4.

Why: dividing by a negative reverses the inequality, and the final object is a range of values rather than one number.

Check: test x = −5 and x = 0 to confirm which side of −4 satisfies the original inequality.

Communication lesson: preserve the type of mathematical object from problem to answer.

Task 14 — A proof that checks examples but proves nothing general

A learner wants to show that the sum of two consecutive odd integers is divisible by 4. They write:

3 + 5 = 8

7 + 9 = 16

11 + 13 = 24

so it is always divisible by 4.

Diagnosis: the examples support a conjecture but do not establish the universal statement.

Repair:

Let the consecutive odd integers be 2n + 1 and 2n + 3.

Their sum is:

4n + 4 = 4(n + 1).

Since n + 1 is an integer, the sum is divisible by 4.

Why: the working covers all integer n, not only selected examples.

Compression: once proof habits are secure, the prose can be concise; the logical generality cannot be skipped.

Task 15 — Trigonometry formula written correctly, sides substituted incorrectly

A right triangle has reference angle 35°, opposite side x and hypotenuse 10. The learner writes:

sin 35° = 10/x.

Diagnosis: the trigonometric function is correct, but the side ratio is reversed.

Repair: label the side roles on the diagram, then write:

sin 35° = x/10

x = 10 sin 35°.

Why: the diagram carries relational working. It makes the substitution visible rather than relying on memory of a formula string.

Transfer: rotate the triangle and repeat with the same relationship to ensure the working is relational, not positional.

Task 16 — Probability tree drawn, state not updated

A learner draws a tree for two draws without replacement from 5 red and 3 blue counters. Every second-stage denominator remains 8.

Diagnosis: the representation exists, but it communicates the wrong state.

Repair: after the first draw, total counters become 7. Branch numerators also update according to the colour removed.

For red then red:

5/8 then 4/7.

For blue then red:

3/8 then 5/7.

Why: a diagram is only useful working if its labels are mathematically faithful.

Communication lesson: representation quality matters more than representation presence.

Task 17 — Formula substitution with a hidden sign error

A learner uses the quadratic formula on:

x² − 5x + 6 = 0.

They write:

x = [−5 ± √(25 − 24)]/2.

Diagnosis: b = −5, so −b = 5. The formula itself may be remembered, but coefficient identification is hidden.

Repair:

a = 1, b = −5, c = 6

x = [−(−5) ± √((−5)² − 4(1)(6))]/2

= [5 ± 1]/2

x = 3 or 2.

Why: writing coefficients makes the sign state inspectable.

Compression: a fluent learner may omit the coefficient line once sign substitution is reliable, but the risky sign should remain visible while errors recur.

Task 18 — A model gives a mathematically valid answer outside its real domain

A tank starts with 20 litres and fills at 6 litres per minute until its 80-litre capacity. A learner writes:

V = 20 + 6t.

Then later uses t = 15 to claim V = 110 litres.

Diagnosis: the algebraic expression has been extended beyond its physical validity.

Repair:

V = 20 + 6t, 0 ≤ t ≤ 10.

Because 20 + 6(10) = 80, the linear filling model stops at 10 minutes.

Why: the domain is part of the model and should remain visible if later calculations depend on it.

Communication lesson: working should preserve model boundaries, not only equations.

Task 19 — Over-expanded arithmetic hides the strategy

A learner solving 25% of 84 writes eight lines converting 25 to 25/100, simplifying by several tiny steps and performing long multiplication.

Diagnosis: the working is mathematically valid but the efficient structure 25% = 1/4 is hidden.

Repair:

25% of 84 = 1/4 × 84 = 21.

Why: good communication should reveal useful structure, not merely preserve every operation.

Transfer: for 12.5% of 80, 1/8 × 80 = 10 may be similarly efficient.

Compression lesson: shorter working can be more informative when it exposes the relationship that makes the arithmetic cheap.

Task 20 — A simultaneous-equation solution changes method but does not show where

A learner begins elimination, then suddenly substitutes a value into one equation without indicating which result was obtained.

Diagnosis: the route may be correct, but the method handoff is invisible.

Repair: label the equations and show the intermediate result.

(1) 2x + y = 11

(2) x − y = 1

(1) + (2): 3x = 12

x = 4

Substitute into (2):

4 − y = 1

y = 3.

Why: the reader can see where elimination ends and substitution begins.

Check: substitute (4,3) into both originals.

Task 21 — A graph answer reports false exactness

A learner reads an intersection from a graph at approximately x = 2.37 and writes:

x = 2.37 exactly.

Diagnosis: a graphical estimate has been reported as exact.

Repair:

x ≈ 2.37

or state an appropriate interval/precision based on the graph scale.

Why: the notation should communicate the evidential strength of the representation.

Transfer: the same principle applies to measured data and rounded values. Working should not claim more precision than the source provides.

Task 22 — Correct calculation, wrong dimension

A similar figure has length scale factor 3:5. A learner is asked for the area ratio and writes 3:5.

Diagnosis: the visible ratio is the linear scale, not the target dimension.

Repair:

Length ratio = 3:5

Area ratio = 3²:5² = 9:25.

Why: the working shows the dimension transition.

Transfer: for volume, use the cube of the linear scale factor. The target object’s dimension should become visible in the route.

Task 23 — A learner copies a full worked solution but cannot reproduce the first decision

The page is beautifully complete. The learner copied a tutor’s solution to a word problem:

let x = number of adult tickets;

let y = number of student tickets;

form two equations;

eliminate;

solve;

check.

When given a changed problem, the learner cannot decide what x and y should represent.

Diagnosis: visible working on the page is not proof that the learner owns the route.

Repair: close the model and ask the learner to reconstruct the representation before seeing later steps.

Communication lesson: working is valuable evidence only to the extent that it reflects reproducible learner control.

Next test: changed context, no example, variable definitions first.

Task 24 — Design the shortest acceptable working that remains recoverable

Problem:

Solve 4(x − 2) + 3 = 19.

One fully expanded solution is:

4(x − 2) + 3 = 19

4x − 8 + 3 = 19

4x − 5 = 19

4x = 24

x = 6.

A fluent learner proposes:

4x − 5 = 19

x = 6.

Question: is this compression acceptable?

Answer: it can be, if expansion/collection and inverse operations are stable and the learner can reconstruct them. The route still preserves the crucial equation state and final solution.

If the learner has a history of negative-bracket or collection errors, keep at least:

4x − 8 + 3 = 19

4x − 5 = 19

4x = 24

x = 6.

Final lesson: there is no universal shortest working. The shortest acceptable route is the shortest route that remains mathematically valid, inspectable and appropriate to the learner and task.

How to read the checkpoint

Do not total the tasks into one “working score”.

Classify the first communication failure instead.

State failure: the current equation, quantity or representation disappears.

Notation failure: symbols make a false or ambiguous claim.

Condition failure: method authority or domain is missing.

Precision failure: exact and approximate values are confused.

Unit failure: quantity type or conversion is hidden.

Route failure: side working or method changes are not connected to the main solution.

Interpretation failure: the final mathematical result does not return to the question.

Compression failure: too much has been hidden.

Expansion failure: routine detail obscures the structure.

The next working intervention should target the first failure, not simply add more lines to every solution.

Appendix B — Working routing matrix: symptom → missing state → smallest repair → next independence test

This matrix turns “show more working” into a diagnostic decision. It is not a universal marking scheme. Start from the page in front of you. Identify the first mathematical state that has disappeared or become ambiguous. Repair only that state, then test whether the learner can keep it visible on a changed problem without being reminded.

Symptom: answers appear with almost no route

Likely missing state: representation and intermediate mathematical state.

Smallest repair: require one line that shows the relationship used and one line that shows the main transformation. Do not immediately demand every arithmetic step.

Example: instead of writing only x = 7 for 4x − 7 = 21, write 4x = 28, then x = 7.

Next independence test: use a changed equation and see whether the learner chooses an appropriate intermediate state without a prompt.

Fade condition: once the learner consistently preserves the equation state and can recover from local errors, routine arithmetic between states may remain mental.

Symptom: many lines are written, but the main method is hard to see

Likely missing state: hierarchy. Routine detail is obscuring the strategic relationship.

Smallest repair: mark the one or two decisions that actually determine the route. Compress repetitive arithmetic that is already secure.

Example: for 25% of 84, replace a long decimal conversion with 25% = 1/4, so 84 ÷ 4 = 21.

Next independence test: ask the learner to rewrite one over-expanded solution in half the number of lines without deleting any high-risk transition.

Fade condition: when the shorter route remains understandable two days later and error rate does not rise.

Symptom: equals signs are used as separators between unrelated steps

Likely missing state: logical meaning of equality.

Smallest repair: require each line to be either a complete equation or a valid expression-equivalence chain.

Example: replace 2x + 3 = 11 = 2x = 8 = x = 4 with three complete equations.

Next independence test: give an expression simplification and an equation-solving task side by side. Ask the learner to explain when an equals sign connects expressions and when it connects equations.

Fade condition: none. Correct equality notation is not scaffolding; it is permanent mathematical meaning. What may fade is the verbal explanation.

Symptom: arrows are used for every transition

Likely missing state: distinction among equality, implication and process sequencing.

Smallest repair: use equals signs for equal values/expressions and ordinary line breaks for equivalent equations. Reserve arrows for places where implication or direction is actually intended.

Next independence test: ask the learner to rewrite one arrow-heavy solution using the correct logical relations.

Fade condition: once notation choice reflects mathematical meaning automatically.

Symptom: a formula is copied correctly, but substitution is wrong

Likely missing state: mapping between problem quantities and formula variables.

Smallest repair: write the variable/value identification before substitution.

Example: for the quadratic formula, write a = 1, b = −5, c = 6 before inserting values if sign errors are recurring.

Next independence test: use a changed equation where b is positive, then one where a is not 1. Remove the coefficient-identification line only when substitution remains stable.

Fade condition: repeated accurate substitution across varied signs and coefficients.

Symptom: a geometry method is correct only when the diagram looks familiar

Likely missing state: theorem condition and relational labels.

Smallest repair: make the condition visible on the diagram: right angle, corresponding vertices, parallel lines, reference angle or side roles.

Next independence test: rotate the diagram while keeping the mathematics simple. Ask the learner to mark the condition before calculating.

Fade condition: the learner identifies the relationship correctly across orientation changes without extra labels supplied by the tutor.

Symptom: a probability answer is numerically plausible, but the event logic is invisible

Likely missing state: event and state transition.

Smallest repair: write the event or branch sequence and show how counts change.

Example: P(RR) = 5/8 × 4/7, not merely 5/14.

Next independence test: contrast with-replacement and without-replacement problems and ask the learner to show only the state change before completing the arithmetic.

Fade condition: branch-state updates remain accurate without explicit reminders.

Symptom: a statistics calculation is correct but no reader can tell why that statistic was chosen

Likely missing state: selection reason or aggregation rule.

Smallest repair: add one relationship line or one concise reason.

Example: unequal group sizes → reconstruct totals before combining means.

Next independence test: pair equal-size and unequal-size group-mean questions so the learner must decide whether weighting is necessary.

Fade condition: the learner selects and interprets the statistic correctly in mixed work.

Symptom: units appear only in the final answer and conversion errors are common

Likely missing state: dimensional transition.

Smallest repair: write units only at the conversion or quantity-type transition where they matter.

Example: 2 min = 120 s before multiplying by a speed in m/s.

Next independence test: mix compatible and incompatible units. Ask the learner to identify the conversion before calculating.

Fade condition: conversions remain accurate and the final dimension is correct without repeated unit annotation.

Symptom: the learner rounds every intermediate decimal immediately

Likely missing state: exactness/precision status.

Smallest repair: preserve the exact expression or calculator value through intermediate stages and round only when the answer contract requires it.

Next independence test: give a multi-stage problem where early rounding creates a visibly different final result.

Fade condition: the learner independently chooses when exact form or stored calculator precision should be preserved.

Symptom: calculator errors cannot be diagnosed

Likely missing state: intended mathematical expression.

Smallest repair: require the expression or substituted formula to appear on paper before the output.

Next independence test: provide an implausible calculator answer and ask whether the problem lies in algebra, grouping, mode or transcription.

Fade condition: the expression remains visible even when button-level guidance disappears.

Symptom: variable letters appear without definitions in modelling problems

Likely missing state: semantic meaning of symbols.

Smallest repair: define each non-obvious variable once.

Example: let d = number of days; C = total cost.

Next independence test: use a changed context with two unknowns and see whether the learner defines variables before forming equations.

Fade condition: definitions may become shorter, but meaning should remain unambiguous throughout the route.

Symptom: the model equation is correct but later used outside its context

Likely missing state: domain or assumption.

Smallest repair: write the valid interval or condition next to the model when later use could violate it.

Next independence test: ask the learner to identify where a linear model ceases to describe the situation.

Fade condition: domain awareness remains active even when the interval is not explicitly prompted.

Symptom: side calculations are everywhere and cannot be traced

Likely missing state: connection between side work and main route.

Smallest repair: label the side calculation and write its result back into the main chain.

Next independence test: ask the learner to revisit the page the next day and explain the route without teacher narration.

Fade condition: side work may become more compact, but every retained side result should have an obvious destination.

Symptom: the learner restarts from the beginning after every error

Likely missing state: trusted checkpoints.

Smallest repair: keep main transformations on separate lines so the learner can identify the last valid state.

Next independence test: give a solution with one planted error and ask the learner to repair from the first invalid line rather than redo the whole question.

Fade condition: recovery becomes local and spontaneous.

Symptom: an answer is mathematically correct but not in the requested form

Likely missing state: target contract.

Smallest repair: keep a small cue for the required object—length, percentage, equation, exact value, nearest integer, explanation.

Next independence test: use a multi-part question where different parts require different answer forms.

Fade condition: the learner naturally returns from calculation to the target without a cue.

Symptom: proof working contains true statements but no logical bridge

Likely missing state: reason or implication.

Smallest repair: add the theorem, definition, algebraic factorisation or condition that makes the next claim follow.

Next independence test: give a short proof with one missing reason and ask the learner to supply only the bridge.

Fade condition: standard reasons may become concise, but logical sufficiency must remain.

Symptom: the learner writes long prose around simple algebra

Likely missing state: symbolic fluency and division of labour between symbols and words.

Smallest repair: let equations carry transformations; reserve prose for conditions, interpretations and reasons not safely carried by symbols.

Next independence test: rewrite a prose-heavy solution using half the words while preserving every mathematical reason.

Fade condition: the learner can choose when prose adds meaning rather than merely describing actions.

Symptom: neatness improves but errors do not

Likely missing state: mathematical rather than visual organisation.

Smallest repair: identify whether the first error comes from representation, condition, sign, unit, calculator entry or interpretation. Make that state visible.

Next independence test: compare two versions of the same page: one visually neat but mathematically opaque, one less polished but structurally explicit. Ask which is easier to verify and why.

Fade condition: visual organisation remains useful but is no longer treated as a substitute for valid reasoning.

Symptom: working is excellent in homework but disappears under time

Likely missing state: risk-weighted compression under load.

Smallest repair: identify the two or three high-risk transition types that must remain visible even during timed work.

Next independence test: use a short timed mixed cluster and compare which lines disappeared when errors increased.

Fade condition: the learner maintains a compact recoverable route at sustainable pace.

Symptom: a learner copies perfect working from AI or a tutor

Likely missing state: learner ownership.

Smallest repair: close the model and require reconstruction of the first representation, first decision and one check.

Next independence test: changed problem, same invariant, no model visible.

Fade condition: the learner can reproduce the route architecture without copying the surface sequence.

Symptom: the learner’s working is different from the teacher’s but mathematically valid

Likely state: possible healthy method variation.

Do not repair automatically. First test validity, completeness, efficiency and readability.

Smallest intervention: if the route is valid, ask the learner to explain why it works and compare costs with the alternative.

Next independence test: use a problem where the preferred method changes and see whether the learner can choose flexibly.

Principle: good mathematical communication does not require one canonical surface format when multiple valid routes exist.

How to use the matrix

The matrix should eventually disappear.

The learner should internalise a shorter set of questions:

What state must remain visible?

Where is the risky transition?

What can I safely keep mental?

What condition or unit could disappear?

Could I recover if this answer failed?

Could another reader reconstruct the route?

When those questions become natural, working has become self-regulated mathematical communication rather than a teacher-imposed format.

Appendix C — Eight composite working laboratories: from opaque pages to inspectable mathematical routes

These laboratories bring several working decisions together. Each case begins with a page that could plausibly appear in real Secondary Mathematics study. The aim is not to make every solution look identical. The aim is to identify what the page must preserve so that the Mathematics remains valid, readable, checkable and learner-owned.

Laboratory 1 — Simultaneous equations: the route is correct but the handoff is invisible

Problem:

A school shop sells notebooks and pens. Three notebooks and two pens cost 13 dollars. One notebook and two pens cost 7 dollars. Find the price of each item.

Adrian writes:

3n + 2p = 13

n + 2p = 7

2n = 6

n = 3

p = 2.

The answer is correct.

The page is close to sufficient, but one useful transition is missing: how p = 2 was obtained.

A stronger compact route is:

(1) 3n + 2p = 13

(2) n + 2p = 7

(1) − (2): 2n = 6

n = 3

Substitute into (2):

3 + 2p = 7

p = 2.

Therefore notebook = 3 dollars, pen = 2 dollars.

What the working preserves: variable meaning, equation labels, elimination operation, substitution handoff and contextual interpretation.

What can be compressed later: the line 3 + 2p = 7 may combine mentally to p = 2 if substitution is stable. The equation labels may be omitted if the system is short and still readable.

What should not disappear: enough of the method handoff that the reader can see where the second value came from.

Recovery value: if p is later found to be wrong, the learner can return directly to the substitution rather than redo elimination.

Laboratory 2 — Geometry: the calculation is clean but the theorem is unauthorised

Problem:

A triangle has side lengths 6 cm, x cm and 10 cm. A diagram looks right-angled at the vertex between the 6 cm and x cm sides, but no right-angle symbol or statement is given. Find x.

Clara writes:

x² + 6² = 10²

x² = 64

x = 8 cm.

The working is beautifully laid out.

It is not justified from the stated information.

The missing state is not another algebraic line.

It is the theorem condition.

If the problem actually includes a right-angle condition, the repaired route is:

Triangle is right-angled, with 10 cm as hypotenuse.

x² + 6² = 10²

x² = 64

x = 8 cm.

If no right angle can be established, then Pythagoras cannot be used solely from visual appearance.

What the working preserves: method authority.

What this laboratory teaches: more algebra cannot repair a missing condition. Good communication begins before calculation.

Compression rule: once theorem-condition recognition is secure, a correctly marked diagram may carry the condition without an extra sentence.

Laboratory 3 — Modelling: the symbols are right, but the model does not yet mean anything

Problem:

A car-rental company charges 35 dollars plus 0.40 dollars per kilometre. A customer pays 83 dollars. Find the distance travelled.

Ben writes:

35 + 0.4x = 83

0.4x = 48

x = 120.

The route is mathematically correct.

For a short one-off problem, many readers can infer that x is distance.

For robust modelling communication, especially in longer problems, one definition improves the page:

Let d = distance travelled in km.

35 + 0.40d = 83

0.40d = 48

d = 120.

Distance travelled = 120 km.

What the working preserves: semantic meaning of the variable and final unit.

What can be compressed: the sentence “distance travelled” need not be repeated on every line.

Changed version: if a daily kilometre limit or maximum rental distance exists, the valid domain should be stated where it constrains later use.

Recovery value: if the final number is implausible, the learner can inspect whether the fixed fee, rate or variable meaning was modelled incorrectly.

Laboratory 4 — Graphs: the algebra is correct but the coordinates came from a misread scale

A graph shows a straight line. The x-axis increases by 5 units per major division, while the y-axis increases by 2 units per major division.

Adrian picks two visible points and writes:

m = (6 − 2)/(4 − 2)

= 4/2

= 2.

The arithmetic is internally correct.

The x-coordinates were read as 2 and 4 even though the plotted points are actually at x = 10 and x = 20.

A repaired working page begins at the representation:

Read axis scale first.

Points used: (10, 2) and (20, 6).

m = (6 − 2)/(20 − 10)

= 4/10

= 0.4.

What the working preserves: source coordinates and their scale.

Communication lesson: a formula cannot protect a value that was read incorrectly before the formula began.

Compression: the sentence “read axis scale first” is training scaffolding. The coordinates themselves should remain if they are used in the calculation.

Laboratory 5 — Probability: the tree is present, but its branches do not represent the changing state

Problem:

A bag contains 4 red and 3 blue counters. Two counters are drawn without replacement. Find the probability that both counters are red.

Clara draws a neat tree with first branch R = 4/7 and second R branch also = 4/7.

Then she writes:

P(RR) = 4/7 × 4/7 = 16/49.

The representation looks sophisticated.

It communicates the wrong state.

Repair the tree or write the state directly:

First red: 4/7.

After one red is removed: 3 red remain out of 6 total.

P(RR) = 4/7 × 3/6

= 12/42

= 2/7.

What the working preserves: sequential state.

What the page teaches: a diagram is working only when its labels are mathematically faithful.

Compression: once no-replacement state control is secure, the explicit sentence can disappear and the changing fractions can carry the meaning.

Laboratory 6 — Statistics: a compact formula is better than a long narrative when it exposes the weighting

Problem:

Class A has 18 students with mean score 64. Class B has 12 students with mean score 76. Find the combined mean.

Ben starts writing a paragraph explaining that the groups have different sizes and should not be averaged equally.

The explanation is correct.

The Mathematics can communicate the same idea more efficiently:

Total score A = 18 × 64 = 1152.

Total score B = 12 × 76 = 912.

Combined mean

= (1152 + 912)/(18 + 12)

= 2064/30

= 68.8.

A short reason may be added if the task asks for explanation:

Group sizes differ, so totals must be weighted by group size.

What the working preserves: aggregation logic.

What can be compressed: an experienced learner could write (18×64 + 12×76)/30 = 68.8 directly.

What should remain visible: the weighting. Writing (64 + 76)/2 would hide the essential structure and produce the wrong answer.

Laboratory 7 — Calculator-supported quadratic: paper working should separate algebra from device execution

Problem:

Solve x² − 9x + 17 = 0, giving decimal answers if required.

Adrian types something into a calculator and writes:

x = 6.30, 2.70.

Those values are plausible, but there is no way to inspect coefficient signs, discriminant or calculator grouping.

A stronger route is:

a = 1, b = −9, c = 17

x = [−b ± √(b² − 4ac)]/(2a)

= [9 ± √(81 − 68)]/2

= [9 ± √13]/2

≈ 6.30 or 2.70.

What the working preserves: coefficient mapping, exact structure and the point at which approximation begins.

What the calculator carries: numerical evaluation of √13 and final division.

What can be compressed later: the coefficient-identification line may disappear once sign substitution is consistently accurate. The exact expression [9 ± √13]/2 is still valuable because it exposes the mathematical result before decimal approximation.

Laboratory 8 — Proof-like reasoning: correct algebra needs a final sentence that actually proves the target

Problem:

Show that the difference between the squares of two consecutive integers is always odd.

Clara writes:

(n + 1)² − n²

= n² + 2n + 1 − n²

= 2n + 1.

Then she stops.

The algebra is correct.

The proof target has not been explicitly closed.

Add:

Since 2n + 1 is odd for every integer n, the difference is always odd.

What the working preserves: general variable, algebraic transformation and logical conclusion.

What can be compressed: a fluent proof may combine expansion and simplification in one line.

What should not disappear: the bridge from algebraic form to the property being proved.

What the laboratories show together

Clear working is not one visual style.

In algebra, it may mean preserving equivalence.

In geometry, it may mean preserving theorem conditions.

In modelling, variable meaning and domain.

In graphs, scale and source coordinates.

In probability, changing state.

In statistics, aggregation logic.

With calculators, the intended expression.

In proof, the logical bridge to the conclusion.

The shared job is always the same:

make the mathematically decisive state visible at the point where losing it would make the route ambiguous, fragile or impossible to verify.

Appendix D — Compact working templates that should fade as judgement improves

Templates can help learners who do not yet know what to leave visible. They are scaffolds, not permanent formatting laws.

Template 1 — Equation

Original state

Write the equation accurately.

Main transformations

One equivalent equation per important state.

Solution

Identify the unknown clearly.

Check if risk justifies it

Substitute into the original.

Fade the template when the learner preserves equality and can recover without it.

Template 2 — Word problem / model

Define: variable and unit.

Relationship: equation/model.

Solve: visible main route.

Interpret: answer in context.

Boundary: domain/assumption if it matters.

Fade labels once the learner performs these jobs naturally.

Template 3 — Geometry

Evidence: mark given/derived conditions.

Relationship: theorem, ratio or similarity statement.

Calculate: compact algebra.

Interpret: length/angle/area with units and precision.

Check: magnitude or geometric consistency where useful.

Template 4 — Probability

Event: what outcome is required?

State: what changes between stages?

Representation: direct product, tree, table, complement or cases.

Calculation: show the event structure.

Check: result in [0,1], event logic sensible.

Template 5 — Statistics

Target summary: what measure/comparison is required?

Aggregation: totals, counts or frequencies if needed.

Calculation: visible weighting or formula.

Interpretation: what does the statistic say?

Limit: what does it not show?

Template 6 — Calculator-supported problem

Mathematical expression: write the intended structure.

Substitution: show values/signs if fragile.

Device evaluation: calculator carries arithmetic.

Output status: exact or approximate?

Check: sign, magnitude, unit or domain.

Template 7 — Proof / reasoning

Given/representation: define variables or establish conditions.

Claim: what needs to be shown next?

Reason: theorem, definition, algebra or earlier result.

Consequence: what follows?

Closure: explicitly connect the final statement to the proof target.

Template 8 — Recovery

Check fails.

Last trusted state: circle or identify it.

First suspect transition: sign, copy, condition, unit, calculator entry, interpretation.

Repair locally.

Re-enter main route.

This template should disappear once local recovery becomes habitual.

The final template is no template

The goal is not to make every student carry eight formats.

The goal is to teach enough contrast that they can answer one question independently:

What does this problem need to keep visible?

A fluent learner may use three lines on one problem and fifteen on another.

A developing learner may show a sign-sensitive transition that an expert keeps mental.

A modelling problem may need a variable definition while a pure-number problem does not.

A proof may need a reason sentence while a calculation does not.

That variability is not inconsistency.

It is mathematical judgement.