Secondary Mathematics Tuition | How to Use a Maths Textbook: Learn Definitions, Worked Examples and Exercises Properly

Secondary Mathematics Tuition · Textbook learning, worked examples and independent exercises

A Mathematics textbook can look deceptively simple.

There is a chapter title.

A few definitions.

Some worked examples.

Several pages of exercises.

Answers at the back.

Jo opens the book and begins at Question 1.

She solves until she gets stuck, checks the answer, copies part of the method, and continues.

Two hours later she has covered eight pages and is not sure what the chapter was trying to teach.

Aisha does the opposite.

She reads every word carefully, highlights definitions, copies each example into a notebook and postpones the exercises because she wants to “understand everything first”.

Her notes look excellent.

When the book is closed, the method is difficult to reconstruct.

Ben uses the same textbook differently.

He first maps the chapter: what is new, what earlier Mathematics it depends on, what definitions matter, what examples represent distinct problem types, and what the exercises are progressively testing.

He attempts before reading full solutions when possible.

He explains why a worked example takes each step.

He selects exercises for a purpose rather than simply doing every question in order.

He uses answers to test, not to replace, his own Mathematics.

He returns after delay.

Then he leaves the textbook and checks whether the learning survives on changed problems.

This worldwide guide is about that workflow.

How to use a Maths textbook is not a question about how many pages to read.

It is about turning a Mathematics textbook into a sequence of active learning states: orientation, definition, representation, worked-example analysis, independent attempt, feedback, changed practice, retrieval and transfer.

A good Secondary Mathematics textbook contains much more than information. It contains an instructional architecture. Definitions set the legal meaning of objects. Notation compresses relationships. Worked examples expose routes. Exercises vary difficulty and representation. Review sections test whether earlier material remains available. Answers and solutions provide feedback. Chapter order encodes dependencies.

But the learner still has to use those components correctly.

This page owns the worldwide learner-facing textbook workflow across Secondary Mathematics. It does not replace How to Study Maths Effectively, which owns the broader study system; How to Practise Maths Effectively, which owns deliberate practice design; How to Use Worked Examples in Maths, which owns example fading; How to Remember Maths, which owns retrieval and spacing; or How to Catch Up in Maths, which owns accumulated-gap recovery.

This article owns a narrower job:

how to extract the intended learning system from a Mathematics textbook and convert it into independent mathematical control.

Jo, Aisha and Ben are fictional recurring learners. Their scenes are explanatory examples, not reported student cases or fixed ability labels.

The central textbook system is:

MAP THE CHAPTER → IDENTIFY PREREQUISITES → READ DEFINITIONS FOR MEANING → READ NOTATION AS A LANGUAGE → PREDICT BEFORE THE WORKED EXAMPLE CONTINUES → EXPLAIN WHY EACH STEP IS LEGAL → ATTEMPT A PAIRED PROBLEM → SELECT EXERCISES BY PURPOSE → MARK FROM EVIDENCE → REPAIR THE FIRST FAILURE → CLOSE THE BOOK → RETRIEVE → CHANGE THE SURFACE → MIX WITH NEIGHBOURS → RETURN AFTER DELAY → LEAVE THE TEXTBOOK WHEN THE MATHEMATICS CAN TRAVEL WITHOUT IT.

This is not a universal instruction to follow every textbook in the same order. Books differ by curriculum, publisher, level, notation, exercise design and answer support. Some are excellent explanatory texts. Some are concise reference books. Some assume substantial teacher mediation. Some include rich problem sets. Some do not. The workflow in this guide helps the learner decide what job the book can actually do.

Evidence and scope note: the chapter maps, textbook dashboards, selection matrices and fictional scenes in this article are original instructional tools, not validated scales. Public evidence used carefully here includes the What Works Clearinghouse guide Organizing Instruction and Study to Improve Student Learning, which gives different evidence ratings to practices such as spacing, worked examples, quizzing/re-exposure and explanatory questioning; the WWC Algebra guide on solved problems, algebraic structure and strategy choice; and NCETM’s Five Big Ideas around coherence, representation and structure, mathematical thinking, fluency and variation. These sources support bounded instructional principles rather than one universal way to read every Mathematics textbook.

1. A Maths textbook is a learning system, not a container of questions

Many students approach a Mathematics textbook as if its main purpose is to store exercises.

Open chapter.

Find questions.

Start calculating.

That misses the architecture.

A well-designed textbook usually has several layers.

Orientation.

Chapter title, objectives, opening examples, prerequisite reminders or a motivating problem establish what kind of Mathematics is about to be built.

Definitions.

These specify what mathematical words and objects mean.

Notation.

Symbols compress relationships and must be read precisely.

Explanations.

These connect new ideas to earlier structures and explain why procedures are valid.

Worked examples.

These expose routes through representative problems.

Exercises.

These test whether the learner can carry increasingly more of the route.

Variation.

Question sequences may change numbers, wording, representation, method choice or difficulty to reveal what is invariant.

Review.

Mixed and cumulative sections may test whether the learning survives without immediate chapter cues.

Feedback.

Answers, hints or full solutions provide evidence about performance.

Index/reference.

Later study may use the book as a lookup system rather than a chapter-by-chapter course.

Jo misses the definitions because she begins at the exercise set.

Then a question uses the phrase “directly proportional”.

She remembers that the graph is a straight line but forgets that direct proportion in the school context requires the relationship y = kx and therefore passes through the origin.

The definition was not decoration.

It controlled method legality.

Aisha reads every example but does not attempt between them.

She experiences the chapter as continuous explanation.

The book feels clear because the author is making all the decisions.

Ben notices that Example 1 introduces the relationship, Example 2 changes the unknown direction, Example 3 removes a cue, and Exercise 1.3 mixes near neighbours.

He sees the instructional sequence.

A textbook should therefore be read with two questions:

What Mathematics is this page communicating?

What learner responsibility is this page trying to transfer to me?

A definition transfers precise meaning.

A worked example transfers a route.

An exercise transfers execution.

A mixed review transfers selection.

A cumulative test transfers retrieval.

If the learner understands the job of each component, the textbook stops being a pile of pages.

It becomes a designed learning environment.

2. Map the chapter before you study it so you know what is new, what is prerequisite and what is merely a later application

Opening at page one and reading linearly can work.

It can also hide the structure of the chapter.

Before detailed study, spend a few minutes building a chapter map.

Look at:

section headings;

definitions;

worked-example titles;

exercise groups;

summary boxes;

review problems;

and any prerequisite notes.

Do not solve yet.

Ask:

What seems to be the central relationship?

What earlier Mathematics is assumed?

Which sections introduce new ideas?

Which sections appear to apply or combine them?

Which examples look like distinct problem families?

Where does the book begin mixing?

Suppose the chapter is “Quadratic Functions”.

A chapter map may reveal:

1. expanding quadratic expressions;

2. factorisation;

3. solving quadratic equations;

4. graphs and roots;

5. completing the square;

6. maximum/minimum or vertex interpretation;

7. modelling problems;

8. mixed review.

Now the learner can see dependencies.

If factorisation is weak, solving by factor form may become difficult.

If graph coordinates are weak, the graph section may fail for a different reason.

If algebraic manipulation is secure, the learner may move faster through early examples and spend more time on representation and modelling.

Chapter mapping is especially useful in catch-up.

Suppose Ben has already learned direct percentage but opens a chapter containing:

percentage change;

reverse percentage;

compound percentage;

growth and decay;

profit/discount applications.

He does not need to pretend every section is equally new.

He identifies what is secure and where the first new dependency begins.

A chapter map should remain small.

One page or less.

A useful template is:

Chapter: name.

Big idea: one sentence.

Prerequisites: three to five important earlier structures.

New definitions: list only the ones that change meaning or method.

New relationships: formulas, theorems, models.

Example families: direct, reversed, applied, mixed.

End test: what should I be able to do without the chapter open?

Aisha initially copies the table of contents into her notes.

That is not yet a map.

A map records relationships.

“Factorisation → roots → x-intercepts” is more useful than simply listing three section headings.

“Ratio → trigonometric ratio → unknown side/angle” is more useful than copying page numbers.

The map tells the learner where to return if later work fails.

It also prevents chapter study from becoming page-count completion.

3. Read definitions as rules that decide what a mathematical object is and what conclusions are legal

Students often skim definitions because they seem less important than worked examples.

In Mathematics, a definition can determine the entire legal structure of a problem.

“Prime number.”

“Function.”

“Direct proportion.”

“Similar figures.”

“Independent events.”

“Gradient.”

“Median.”

“Irrational number.”

These are not vocabulary decorations.

They define objects or relationships.

Take direct proportion.

If the textbook defines y directly proportional to x as y = kx for constant k, then:

y = 3x

is directly proportional.

y = 3x + 5

is not, even though it is linear and increasing.

The definition lets the learner reject a near neighbour.

Take a prime number.

A prime positive integer has exactly two positive divisors in the standard school definition: 1 and itself.

That definition excludes 1.

Memorising a list of primes without understanding the definition can produce confusion later.

Take similar figures.

Similarity is not “looks the same shape”.

It involves equality of corresponding angles and proportional corresponding lengths under the curriculum’s definition and theorems.

The exact formalism can vary by level, but the definition controls what follows.

Use a definition routine.

Read.

What exactly is being defined?

Paraphrase.

Can you say it in simpler language without changing its meaning?

Example.

Give one clear case that satisfies it.

Non-example.

Give one tempting case that fails.

Condition.

Which word in the definition prevents the non-example from qualifying?

Use.

What later theorem, formula or method depends on this definition?

Jo learns “direct proportion” by writing:

definition:

y = kx;

example:

y = 4x;

non-example:

y = 4x + 2;

why non-example fails:

nonzero intercept, ratio y/x not constant for all relevant x.

Ben uses a geometric definition:

radius = line segment from centre to circumference;

diameter = chord through centre, length twice radius.

Then he can reject a random chord as a diameter.

Aisha initially copies definitions word for word.

Copying may help attention.

It is not enough.

The test is whether she can use the definition to classify unfamiliar cases.

A definition has been learned when it changes what the learner can distinguish and justify.

4. Read textbook notation as a language whose symbols carry conditions, grouping and logical meaning

Mathematics textbooks compress enormous meaning into notation.

A learner who reads only the surrounding prose can miss the most important information on the page.

Consider:

x ∈ ℤ.

This may restrict x to integers.

0 ≤ x ≤ 10.

This gives a domain or feasible interval.

f(x).

This names a function value at input x; it is not automatically multiplication of f and x.

a:b.

This denotes a ratio in many school contexts, not division notation alone.

≈.

This communicates approximation rather than exact equality.

∥.

This can mark parallel lines.

⊥.

This can mark perpendicularity.

Notation should be read actively.

When a textbook introduces a symbol, ask:

What does it mean?

What does it not mean?

What condition is being compressed?

What common misread would change the Mathematics?

Grouping matters too.

3(x + 4)

is not the same as

3x + 4.

√(x + 4)

is not the same as

√x + 4.

sin²θ

has a conventional meaning different from sin(2θ).

Exact conventions vary with course and notation, so learners should follow the book’s definitions and teacher guidance.

Textbooks also use layout as notation.

A fraction bar groups numerator and denominator.

A superscript changes power.

A table header changes what every cell means.

A diagram mark can encode equality or angle conditions.

Aisha sometimes reads a formula as a string to memorise.

Ben reads its structure.

For the quadratic formula, he identifies:

whole numerator;

plus/minus branch;

discriminant inside the square root;

whole denominator 2a.

This helps him enter it correctly into a calculator and reason about when real roots exist.

Jo uses a notation translation habit:

symbol → sentence.

For example:

x ≥ 4

becomes:

x is at least 4.

Then she reverses it:

sentence → symbol.

This two-way translation exposes whether notation has become meaningful rather than decorative.

The worldwide owner Read Mathematical Notation Like a Language develops notation more broadly.

In textbook study, the practical rule is:

do not move past notation you cannot translate accurately into mathematical meaning.

5. Read explanations for the relationship they are building, not for sentences to highlight

Mathematics textbook prose is different from narrative prose.

A paragraph may exist to:

define a relationship;

explain why a rule works;

connect two representations;

state a condition;

justify a transformation;

or warn about a common failure.

Highlighting the entire paragraph preserves none of those distinctions.

Use a three-part read.

Before reading:

What question is this paragraph trying to answer?

During reading:

What relationship changes from the beginning to the end?

After reading:

Can I state the key mathematical claim without the book?

Suppose the paragraph explains why dividing by a negative reverses an inequality.

The learner should not retain only:

“flip the sign”.

The deeper relationship is order reversal under negative scaling.

If 4 > 1, multiplying both by −1 gives:

−4 < −1.

That relationship explains the rule.

Now a paragraph explains why combined means need group sizes.

The important claim is:

mean = total/count;

therefore total = mean×count;

therefore unequal groups contribute unequal weight.

Now a paragraph explains a graph:

y = 3x + 5.

The relationship may be:

5 is the state when x=0;

3 is the change in y per one unit of x.

The learner should be able to reconstruct that interpretation in a different context.

Use a margin question rather than a highlight.

Examples:

“Why does the sign reverse?”

“What stays constant here?”

“What is the 100% base?”

“Why is this only an estimate?”

“Which condition makes the theorem legal?”

These questions turn prose into a prompt for retrieval later.

Ben often writes one sentence after an explanation:

“This paragraph is really saying that…”

If he cannot finish the sentence accurately, he has not yet extracted the relationship.

Aisha writes too many notes.

Her repair is to reduce each explanation to:

one core relationship;

one condition;

one example or counterexample.

Jo barely reads explanations at all.

Her repair is the opposite:

when a rule keeps failing, go back and read the reason rather than doing another ten mechanical questions.

Textbook explanations matter most when they change the learner’s mental model.

The test is not whether the paragraph was read.

It is whether the relationship can now be explained and used.

6. On the first pass through a worked example, predict before you read the next line

A worked example can teach efficiently because it makes a valid route visible.

It can also create an illusion of understanding because every difficult decision has already been made by the author.

The learner reads:

“Yes, that makes sense.”

Then the book closes and the route disappears.

Prediction changes the reading state.

Cover the next line.

Ask:

What should happen next?

Why?

What information supports that move?

What alternative would be tempting but wrong?

Suppose a worked example solves:

3(x − 4) = 2x + 5.

The book begins:

3x − 12 = 2x + 5.

Stop.

Before reading further, predict:

One useful next state is:

x − 12 = 5

after subtracting 2x from both sides.

Then:

x = 17.

Now compare with the book.

If the book chose:

3x − 2x = 5 + 12,

the surface route differs but the mathematical goal is the same: preserve equivalence while isolating x.

The prediction has done two things.

It forces route generation.

It also prepares the learner to compare valid alternatives.

Now a geometry example.

The diagram establishes a right triangle.

Angle = 32°.

Adjacent side = 7.

Hypotenuse is unknown.

The book begins:

cos 32° = 7/h.

Stop.

Predict:

h = 7/cos32°.

Then read on.

Now probability.

A bag has 5 red and 3 blue counters.

Two are drawn without replacement.

The example writes:

P(R first) = 5/8.

Stop.

Predict the second red probability:

4/7.

If the learner predicts 5/8 again, the worked example has exposed the precise state-change gap before the final answer is read.

Prediction works best at high-value transitions.

Do not cover every arithmetic simplification if the learner already owns it.

Predict:

representation;

method choice;

equation formation;

theorem application;

branch-state change;

major algebraic transformation;

or final interpretation.

Jo initially reads the examples too quickly.

Her new rule is:

pause before the first decision I would have had to make independently.

Aisha pauses at every line and makes textbook reading painfully slow.

Her new rule is:

predict only the decisions that carry mathematical structure.

Ben uses a pen or sheet of paper to hide the lower half of a worked example.

That small physical change turns passive reading into partial problem solving.

The book still provides support.

The learner begins taking back the route.

7. Self-explain worked examples by asking why each important step is valid and what would make it invalid

A learner can reproduce a sequence without understanding why it works.

Self-explanation forces the route to connect to relationships and conditions.

Use questions such as:

Why is this operation legal?

What stayed invariant?

Which condition makes this theorem applicable?

Why is this representation useful?

What would change if one condition changed?

What common wrong route does this line avoid?

Consider:

x² − 7x + 12 = 0

(x − 3)(x − 4) = 0

x = 3 or x = 4.

Weak explanation:

“Factorise, then solve.”

Stronger explanation:

Factorisation rewrites the quadratic into a product of two factors.

A product equals zero only if at least one factor equals zero.

Therefore x − 3 = 0 or x − 4 = 0.

This connects the algebra to the zero-product property.

Now similarity.

The book writes:

AB/DE = AC/DF.

Ask:

Why are those sides paired?

Because the established vertex correspondence is A↔D, B↔E, C↔F.

If the triangles merely look similar but similarity has not been established, the ratio may be unjustified.

Now mean.

The book combines two groups using:

(n₁m₁ + n₂m₂)/(n₁+n₂).

Ask:

Why multiply mean by count?

Because mean = total/count, so total = mean×count.

The formula is not arbitrary.

It reconstructs group totals before combining them.

Self-explanation can also expose hidden assumptions.

A worked example might use a straight-line model.

Ask:

Over what domain is the model intended to apply?

Is the relationship assumed constant?

Does the context allow negative input?

These questions matter because textbooks sometimes simplify context to teach one relationship.

The learner should know where the simplification sits.

Aisha writes long paragraph explanations for every example.

That becomes inefficient.

Use concise prompts:

Why this line?

What condition?

What invariant?

What alternative?

What would break it?

One or two sentences can be enough.

Ben self-explains only the strategic transitions.

Routine arithmetic does not need an essay.

Jo uses self-explanation when an example “looks easy” but she is not sure she could reconstruct it.

That is often the right moment.

If you can explain why the example works, the route is becoming yours.

If you can only say what the next printed line is, the book is still carrying too much.

8. Attempt a paired problem immediately after the example while the relationship is clear but the solution is no longer visible

The transition from example to problem is where textbook learning becomes independent performance.

Do not read six worked examples in a row if the book offers a chance to attempt.

After one representative example, try a nearby problem.

The problem should preserve the core relationship while changing enough surface detail to require reconstruction.

Example:

Worked:

solve 3(x − 2) = 15.

Paired:

solve 4(x + 3) = 28.

The structure is similar.

The learner must still:

interpret the bracket;

choose an equivalent transformation;

and solve independently.

Next pair:

Worked reverse percentage:

after 20% discount, final 64; original = 80.

Paired problem:

after 15% increase, final 230; find original.

The surface direction changes.

The invariant:

final = multiplier × original.

Next pair:

Worked gradient from points (2,5) and (6,13).

Paired:

points (−1,7) and (4,−3).

The learner must preserve point order and sign control.

The paired problem should be attempted with the worked example closed or covered.

If the learner gets stuck, use a support ladder.

Level 1: reread the target and givens.

Level 2: recall the key relationship from memory.

Level 3: look only at the first decision in the worked example.

Level 4: compare the example and problem to identify what changed.

Level 5: reopen the full example.

Then close it again and complete the problem from the beginning.

The learner should not leave the worked solution visible beside the paired problem indefinitely.

That teaches matching, not reconstruction.

Jo initially flips back after every line.

Her support never fades.

Now she marks the exact point where she got stuck.

If it was:

method selection;

equation formation;

sign control;

or final interpretation,

that becomes the next repair target.

Aisha reads another example whenever she is uncertain.

Her new rule is:

attempt one changed problem before consuming another full solution.

Ben uses the textbook’s exercise sequence to locate natural pairs.

If no paired problem exists, he changes the numbers or context himself.

The paired problem is the first proof that the example taught something transferable.

9. Fade textbook examples deliberately so the book stops supplying the decisions the learner must eventually own

Textbooks vary in how much fading they provide.

Some move from fully worked examples to guided exercises to independent questions.

Some jump directly from example to large exercise sets.

The learner can create their own fading.

Use five stages.

Stage 1 — full worked example.

Read for structure and self-explain important transitions.

Stage 2 — cover-and-predict.

Hide later lines and generate the next step.

Stage 3 — skeleton solution.

Rewrite only the key states and leave gaps to complete.

Stage 4 — problem with one cue.

Keep only the relationship or first representation.

Stage 5 — independent changed problem.

No example visible.

Then add a sixth stage:

delayed return.

The same relationship comes back later without chapter context.

Consider completing the square.

Full example:

x² − 6x + 5

= x² − 6x + 9 − 9 + 5

= (x − 3)² − 4.

Stage 2:

x² − 8x + 3.

Cover the book and predict the inserted square:

+16 −16.

Stage 3:

x² + 10x − 1

= x² + 10x + ___ − ___ −1

= (x + ___)² − ___.

Stage 4:

only cue:

half the coefficient of x.

Stage 5:

x² + 14x + 20.

No cue.

Stage 6:

one week later, complete the square inside a graph or optimisation problem.

The same fading principle applies to geometry proofs, percentage models, probability trees and statistics.

Fading should target responsibility.

Which part is the book still carrying?

Target interpretation?

Representation?

Method selection?

Execution?

Checking?

Remove one responsibility at a time.

The worldwide owner How to Use Worked Examples in Maths develops this system fully.

In textbook study, the practical test is:

if the book closed right now, which decision would disappear?

That decision is what the next fading step should return to the learner.

10. Select textbook exercises by learning purpose instead of doing every question blindly or choosing only the easy ones

Exercise sets are not all the same.

One question may test basic execution.

Another changes representation.

Another combines two ideas.

Another removes a cue.

Another asks for explanation.

Another is simply additional volume.

The learner should understand what the next question is for.

Use an exercise taxonomy.

Foundation question.

Direct use of the new relationship.

Variation question.

Same relationship, changed numbers/signs/orientation.

Reverse-direction question.

Known and unknown quantities exchange roles.

Representation question.

Move between words, equations, tables, graphs or diagrams.

Near-neighbour question.

Looks similar but requires a different method or condition.

Multi-step question.

Combines the new idea with a stable prerequisite.

Mixed question.

Topic label removed; method selection is required.

Transfer question.

New context or surface structure.

Explanation/proof question.

Requires reasoning or justification rather than calculation alone.

Fluency question.

Builds speed and accuracy on a stable routine.

Suppose the book gives 30 linear-equation exercises.

Doing all 30 may be sensible if execution is very weak.

It may be wasteful if the learner is already accurate and the real problem is recognising equations in context.

Ben samples:

two direct;

two sign-sensitive;

one with variables on both sides;

one fractional;

one word-to-equation;

then a mixed question.

If errors appear, he increases volume in that exact sub-type.

Aisha chooses only questions she expects to get right.

Her textbook use confirms familiarity without testing transfer.

She now includes one “boundary question”:

a problem slightly beyond current comfort but still diagnosable.

Jo chooses only the hardest questions because she thinks easy questions are pointless.

She repeatedly fails and cannot tell whether the cause is the new concept or several unstable prerequisites.

Her new sequence is:

direct → variation → changed representation → mixed.

The best exercise selection depends on the learner state.

If knowledge is new, start with enough direct work to stabilise the relationship.

If execution is secure, reduce routine volume and increase variation/selection.

If transfer is weak, change surface and context.

If speed is weak, use short accurate fluency clusters.

If a test is near, include representative assessment conditions after foundations are stable.

The book may provide hundreds of questions.

The learner’s job is not necessarily to finish all of them.

It is to use enough of the right kinds of questions to produce reliable evidence.

11. Use the textbook’s difficulty gradient deliberately: direct work first, then variation, mixing and transfer

Many textbook exercise sets are ordered for a reason.

Early questions stabilise the relationship.

Middle questions vary the surface.

Later questions may combine topics, remove cues or require explanation.

Not every book follows this pattern perfectly, but learners should look for it.

A common study mistake is to skip the direct questions because they look easy, then fail the hardest problem and conclude the whole chapter is weak.

Another mistake is to do only direct questions and conclude mastery because the method works when the page announces it.

Use a difficulty ladder.

Level 1 — direct relationship.

The method is obvious and conditions are clean.

Level 2 — changed numbers or signs.

Same structure, slightly more execution risk.

Level 3 — changed unknown.

The relationship must be rearranged or used backward.

Level 4 — changed representation.

Words become algebra, a table becomes a graph, a diagram must be interpreted.

Level 5 — near-neighbour discrimination.

Similar-looking questions require different routes.

Level 6 — multi-step integration.

Stable prerequisites are combined with the new idea.

Level 7 — mixed or unlabeled.

The learner must choose the method.

Level 8 — transfer or unfamiliar context.

The relationship must survive a changed surface.

A textbook may not label these levels.

The learner can infer them.

Consider percentage.

Level 1:

find 15% of 240.

Level 2:

find 17.5% of 360.

Level 3:

after a 20% discount, final is 96; find original.

Level 4:

use a price table or graph.

Level 5:

percentage increase versus percentage points.

Level 6:

markup then discount.

Level 7:

percentage mixed with ratio and direct proportion.

Level 8:

unfamiliar population or concentration context.

Aisha used to stop when she could do Level 2.

Ben used to jump to Level 8 and fail noisily.

Jo now climbs the ladder only until evidence becomes stable, then samples higher levels.

If Level 3 fails, there is no need to spend an hour on Level 8 yet.

If Level 1–4 are easy, do not repeat them endlessly.

Move upward.

The difficulty ladder helps the textbook become adaptive even when the printed exercise set is fixed.

12. Use answers as feedback after an attempt, not as a route generator before the learner has committed to Mathematics

Answers at the back of a textbook are useful.

They can also destroy the diagnostic value of an exercise.

If the learner checks the final answer after every line, the page becomes a guided puzzle.

The learner no longer knows whether they could have continued independently.

Use an attempt rule.

Before checking an answer:

state the target;

choose a representation;

make a meaningful attempt;

and mark the point of uncertainty.

Then check.

If the final answer matches, do not automatically move on.

Ask:

Was the route valid?

Could two errors have cancelled?

Did I use an assumption the problem did not give?

Did I answer in the required form?

Can I explain why the method works?

Matching the answer is evidence.

It is not proof of method quality.

If the answer does not match, do not immediately read the full solution.

Use a disagreement routine.

1. Re-read the target.

Did you answer the right question?

2. Check transcription.

Did you copy values, signs or units correctly?

3. Check the first high-risk transition.

Expansion, denominator, percentage base, theorem condition, calculator entry.

4. Use an independent check if possible.

Substitution, bounds, graph behaviour, units, reverse operation.

5. Only then consult more help.

This preserves problem-solving evidence.

Ben puts a small question mark next to the first line he distrusts.

He does not erase the entire solution.

Aisha used to mark only right/wrong.

Now she classifies:

correct and independent;

correct after hint;

wrong method;

right method, execution error;

incomplete answer;

or unresolved.

Jo sometimes obtains the correct answer by a route different from the textbook solution.

She checks validity rather than assuming difference means wrong.

Mathematics can admit multiple valid routes.

The answer key should not turn the learner into a copier of one canonical surface sequence.

The role of an answer is to give evidence about the result.

The learner must still judge the route.

13. Use full solutions to diagnose a specific gap, then close them and reconstruct the route

Some textbooks provide full solutions.

Others provide selected solutions or online solution manuals.

Full solutions are powerful because they reveal intermediate states.

They are risky because reading a solution can feel almost identical to understanding it.

Use full solutions only after locating the question you need them to answer.

Examples:

Why did my equation differ?

Why is this theorem legal?

Where did my sign error begin?

Why did the solution reject one root?

Why is the graph answer approximate?

Why did the author choose substitution instead of elimination?

Read the solution with that question in mind.

Then close it.

Reconstruct the route.

If reconstruction fails at the same point, the gap has not yet been repaired.

Suppose Jo attempts:

x² − 5x − 14 = 0.

She cannot factorise it.

The solution shows:

(x − 7)(x + 2) = 0.

Reading that line is not enough.

Close the solution.

Ask:

What two numbers multiply to −14 and add to −5?

−7 and 2.

Now reconstruct:

(x − 7)(x + 2).

Then use a changed problem:

x² − x − 20 = 0.

If the learner cannot transfer, the solution only provided recognition.

Now geometry.

The solution uses:

alternate angles → AA similarity → proportional sides.

Do not copy the chain.

Close the solution and identify the diagram evidence that authorises alternate angles.

Then reproduce the similarity proof.

Full solutions can also teach method comparison.

Your route may be longer but valid.

The textbook may choose a shorter representation.

Ask:

What did the author notice earlier?

What structure made the shorter route available?

What would make my route preferable in another problem?

Ben keeps a strict rule:

solution exposure must be followed by a no-solution reconstruction.

This converts help into learning.

Otherwise the solution manual becomes a second textbook that the learner reads instead of solving.

14. When an exercise is wrong, repair the first wrong or missing decision rather than redoing the whole chapter

Textbook marking is useful when it changes the next task.

A wrong answer should become diagnostic evidence.

Locate the first failure.

Suppose:

5 − 2(x − 3) = 11.

The learner writes:

5 − 2x − 6 = 11.

The first invalid transition is expansion.

Later equation errors are downstream.

Repair:

−2(x − 3) = −2x + 6.

Then retest with:

7 − 3(x − 4).

Do not assign an entire new equation chapter unless broader evidence requires it.

Now:

after a 25% discount, final price is 90.

The learner writes:

90 × 0.75.

The first failure is percentage direction.

Repair:

90 = 75% of original;

0.75P = 90.

Then changed retest.

Now graph:

learner uses correct gradient formula on incorrectly read coordinates.

The first failure is scale reading.

Do not practise more formula substitutions.

Repair the representation.

Now probability:

without replacement, learner uses unchanged denominator.

First failure is state update.

Repair the event state.

The textbook can become a source of micro-diagnoses.

Use an error note:

Question.

First failure.

Repair.

Changed retest.

Return later.

This is enough.

Do not build an enormous error journal if it becomes another administrative task.

Aisha used to write:

“careless mistake”.

That label changes nothing.

Now she writes:

“copied −3 as +3 when substituting into formula.”

The next control is clear:

write coefficients before calculator entry.

Jo writes:

“wrong theorem—no right-angle evidence.”

The next control:

condition check before Pythagoras.

Ben writes:

“right method, answer incomplete—forgot units/precision.”

The next control:

answer-contract check.

The textbook should not produce a pile of red crosses.

It should produce better decisions on the next problem.

15. Take textbook notes only when the notes will do a job the textbook itself cannot do for you

Copying a textbook into a notebook creates a second textbook.

That may feel productive.

It often consumes time without increasing retrieval or transfer.

Notes should compress, reorganise or personalise.

Good textbook notes can record:

a definition in the learner’s own accurate words;

a key condition;

a relationship map;

a method decision rule;

a common near-neighbour contrast;

an error pattern;

a one-line worked skeleton;

a retrieval question;

or a connection to another chapter.

Example:

Instead of copying three pages on direct proportion, write:

Direct proportion:

y = kx.

Ratio y/x constant.

Graph through origin.

Near neighbour:

y = kx + c with c≠0 is linear but not direct proportion.

Retrieval question:

What property distinguishes direct proportion from a general straight-line relationship?

That note is smaller than the textbook and more useful for later recall.

For quadratic factorisation:

Do not copy five full worked examples.

Write:

goal:

product/sum structure;

why:

factor form exposes zero-product roots;

check:

expand back.

Then list one error you personally make.

For geometry:

notes might be a condition map:

right angle → Pythagoras/right-triangle trig candidates;

parallel lines → specific angle relations;

similarity established → corresponding sides proportional.

The note should help future decision-making.

Aisha initially copies every boxed example.

She now writes one skeleton example and several questions to retrieve from memory.

Jo takes no notes and repeatedly rereads pages.

She now records only the definitions/conditions she repeatedly forgets.

Ben’s rule is:

If I can find it instantly in the textbook and do not need to reorganise it, I probably do not need to copy it.

Notes become valuable when they transform the book into a learner-specific control system.

16. Use formula boxes and reference pages as maps of relationships, not as substitutes for knowing what the symbols mean

Textbooks often contain formula summaries.

They are useful for orientation and checking.

They can also encourage formula hunting.

The learner sees a page of symbols and asks:

Which formula has the same letters as my question?

That is fragile.

A formula is a compressed relationship with conditions.

Take:

speed = distance/time.

That relationship tells the learner what each quantity means and how units interact.

It can be rearranged:

distance = speed × time;

time = distance/speed.

Memorising all three separately is possible.

Understanding one relationship is stronger.

Now area of a triangle:

A = 1/2 bh.

The symbols b and h are not any two side lengths.

Height must be perpendicular to the chosen base.

The formula box may be short.

The condition matters.

Now the quadratic formula.

It applies to equations that can be represented as:

ax² + bx + c = 0

with a ≠ 0.

The coefficients a, b, c carry signs.

The discriminant sits inside the square root.

A formula page is useful only if the learner can map the actual problem into this structure.

Use a formula routine.

Name.

What relationship is this?

Variables.

What does each symbol mean?

Conditions.

When is it legal or useful?

Direction.

Can the relationship be rearranged for another unknown?

Units/domain.

What kinds of values are sensible?

Check.

How could the result be verified?

Aisha makes a formula card for:

gradient = change in y/change in x.

Then adds:

same point order in numerator/denominator;

interpret units;

graph must be read from actual axis scale.

The formula is now connected to the representation.

Jo uses formula pages only after she has attempted to retrieve the relationship.

That prevents the textbook from becoming an external memory she consults before thinking.

Ben uses a reference page during early learning, then gradually closes it.

Later, he checks the page after attempting.

Formula sheets are valuable when they confirm and organise.

They become a problem when they replace relationship recognition.

17. Treat textbook diagrams, tables and graphs as mathematical statements that must be read as carefully as prose

Many learners read text carefully and glance at diagrams.

In Mathematics, the diagram may contain the decisive information.

Right-angle marks.

Equal-length ticks.

Parallel-line arrows.

Axis labels.

Scale.

Frequency headings.

Units.

Corresponding vertices.

These are part of the problem statement.

When the textbook introduces a diagram, ask:

What is explicitly given?

What is derived?

What is visual appearance only?

What label defines the target?

What condition authorises the method?

Suppose a geometry example shows two triangles.

One is rotated.

The textbook states:

△ABC ~ △DEF.

The learner should not match sides by page position.

Read correspondence from the named order:

A↔D;

B↔E;

C↔F.

Then:

AB↔DE;

BC↔EF;

AC↔DF.

Now a graph.

Before reading “steep” or “flat”, identify:

x-variable;

y-variable;

units;

scale;

domain;

legend if multiple series exist.

A horizontal segment on a distance-time graph means something different from a horizontal segment on a speed-time graph.

The axes define meaning.

Now a frequency table.

Rows or columns may represent values, frequencies, intervals or categories.

Do not assume every table is an x-y function table.

Textbooks often compress explanation into visual structure.

Ben reads captions and labels before interpreting the image.

Aisha used to redraw every diagram perfectly.

She now redraws only when simplification helps expose the relationship.

Jo marks only the relevant subfigure in a crowded diagram.

The goal is not beautiful copying.

It is accurate mathematical extraction.

A textbook figure has been understood when the learner can reconstruct its essential relationships without depending on its exact visual layout.

18. Reconcile the textbook with class notes by looking for the same Mathematics in different representations

Students sometimes treat textbook and teacher notes as competing sources.

“Which method should I follow?”

Often the two sources express the same underlying relationship differently.

The textbook may derive a formula carefully.

The teacher may provide a compact classroom method.

The textbook may use one notation.

The teacher may use another accepted convention.

The learner’s task is to identify the invariant.

Example:

For linear equations, a teacher may say:

“move +5 to the other side and it becomes −5.”

The textbook may write:

subtract 5 from both sides.

These can represent the same equivalence-preserving operation.

If the shortcut creates sign mistakes, return to the textbook relationship.

Now percentage.

The textbook may use multiplier language:

final = 0.8 × original.

Class notes may use:

80% corresponds to the final quantity.

These are compatible.

Now geometry.

One source may write a theorem name.

Another may use symbolic notation.

The learner should ask:

What condition and relationship are common to both?

Use a source-comparison table when confusion arises.

Textbook says: exact line.

Teacher says: classroom version.

Same underlying relationship: learner explanation.

Difference that matters: notation, condition, level of detail or genuinely different method.

Do not create this table for everything.

Use it only where sources seem inconsistent.

Sometimes the sources genuinely differ.

A textbook may use a method not required in the local course.

A teacher may restrict a method because of syllabus expectations.

A notation convention may differ.

In those cases, follow the course/teacher/official assessment requirements where relevant, while understanding the broader Mathematics.

Aisha used to copy class notes and textbook notes separately.

She had duplicate information everywhere.

Now she keeps one relationship map and notes source-specific conventions only when needed.

Ben uses the textbook to deepen the explanation behind a classroom shortcut.

Jo uses class examples to see how the textbook relationship appears in the actual course.

The two sources should reinforce each other rather than multiply paperwork.

19. Use AI to interrogate the textbook, not to replace the textbook’s learning sequence with instant solutions

AI can make textbook study more interactive.

It can also destroy the textbook’s instructional structure if used too early.

The risky workflow is:

see difficult exercise;

paste into AI;

receive full solution;

read;

continue.

This skips the exact decisions the exercise was designed to transfer.

A stronger workflow preserves the textbook’s role.

1. Use the textbook first.

Read the definition, explanation and worked example.

2. Attempt the exercise.

Write target, representation and first move.

3. Ask AI a bounded question.

Examples:

“Explain why this textbook line is valid without solving the rest.”

“Give me a simpler numerical example of the same relationship.”

“Compare my method with the textbook method and identify the first difference.”

“Do not solve this exercise; ask me one question that will help me decide the next step.”

“Generate two near-neighbour problems where one condition changes the method.”

4. Return to the book problem.

Complete it independently.

5. Close both textbook solution and AI output.

Reconstruct.

6. Test transfer.

Use a fresh problem.

AI can be especially useful when the textbook explanation is too compressed.

Suppose the book moves from:

x² − 6x + 5

to

(x − 3)² − 4

with little explanation.

The learner can ask:

“Explain how completing the square transforms this expression, but do not give me a new problem solution.”

Then return to the textbook and annotate the transformation.

AI can also create extra variation when the book has too few examples.

But the learner should verify generated Mathematics.

AI can make errors or use conventions inconsistent with the course.

The textbook, teacher and official curriculum remain the authority for course-specific expectations.

Ben treats AI as a tutor beside the book.

It can clarify.

It cannot be allowed to carry every first attempt.

Jo uses AI after marking, not before attempting.

Aisha uses AI to generate retrieval questions from her textbook notes rather than to summarise the chapter for her.

The principle is:

AI should increase the learner’s interaction with the Mathematics, not reduce the need to perform it.

20. Use the calculator to execute textbook Mathematics after the expression is understood, not to discover what the Mathematics is supposed to be

Textbook chapters often introduce calculator-supported techniques.

Trigonometry.

Statistics.

Roots and powers.

Scientific notation.

Iterative processes.

Graphing features in some courses.

The calculator is valuable.

It should not hide the mathematical expression.

Suppose the textbook example gives:

x = 10 sin 35°.

The learner should understand:

why sine;

why 35°;

why multiply by 10;

and what unit x should have.

Then the calculator evaluates.

If the answer is 5.74, the learner can compare with the diagram:

Is x shorter than hypotenuse 10?

Yes.

Now the quadratic formula.

Write:

x = [9 ± √13]/2.

Then use calculator output if decimals are required.

This preserves exact structure and makes entry easier to check.

Now statistics.

If a calculator computes mean from a data list, the learner should still know:

what data were entered;

whether frequencies were represented correctly;

what statistic is being reported;

and how to judge plausibility.

Textbooks sometimes show button sequences.

Those sequences can vary by device.

Learn the mathematical job first.

Then learn the device procedure appropriate to the calculator in use.

Aisha copies button sequences and forgets them.

She now writes the mathematical expression beside the key sequence.

Jo obtains a calculator answer and trusts it automatically.

Her new habit is estimate → enter → interpret.

Ben uses the textbook answer to check his output but also uses sign, magnitude, units and domain.

The calculator should reduce arithmetic cost.

It should not replace representation, method choice or interpretation.

21. Use the textbook differently when catching up: diagnose dependencies, skip secure material and reconnect repairs to current chapters

A learner who is catching up should not necessarily study the textbook from page one.

The book’s printed sequence represents a curriculum path.

The learner’s repair path may be different.

Suppose the current chapter is simultaneous equations.

The learner’s errors reveal weak single-equation control and negative-sign handling.

The textbook may contain those prerequisites fifty pages earlier.

Use the book as a dependency library.

Return to the exact earlier section.

Read the definition/explanation.

Study one representative worked example.

Attempt several targeted exercises.

Then return to the current chapter quickly.

Do not automatically redo every page between the prerequisite and current topic.

The worldwide owner How to Catch Up in Maths develops the dual-frontier system fully.

Inside textbook study, use a catch-up map:

Current chapter: what the class is doing.

Visible failure: what goes wrong.

Earlier textbook section: where the prerequisite is taught.

Minimum repair: which examples/exercises are needed.

Return point: which current exercise will prove the repair transferred.

Example:

Current:

linear graphs.

Failure:

cannot rearrange 2x + y = 7 into y = −2x + 7.

Earlier section:

changing subject / equation rearrangement.

Repair:

three direct rearrangements;

one sign-sensitive example;

one equation where the target variable appears twice.

Return:

graph the original equation.

Another example:

Current:

trigonometry.

Failure:

cannot solve sin35° = 7/h for h.

Earlier section:

formula rearrangement / equations with fractions.

Repair.

Return to trig.

A textbook makes catch-up efficient because the prerequisite explanation already exists.

But only if the learner uses it surgically.

Aisha wants to restart the whole book to feel safe.

Her rule becomes:

go back to the earliest active blocker, not the earliest chapter.

Jo wants to ignore old material and survive the current chapter.

Her rule becomes:

if the same older failure appears repeatedly, open the exact prerequisite section and repair it.

Ben tracks repaired sections with a small mark in the contents page.

Once the skill survives current work and delayed return, he stops treating that old chapter as an active gap.

22. Use the textbook for revision by retrieving before rereading and sampling across chapters instead of reading the book from the beginning again

Revision changes the textbook’s job.

During first learning, the book may explain.

During revision, the learner should increasingly generate before consulting it.

A weak revision workflow is:

read Chapter 1;

read Chapter 2;

read Chapter 3;

highlight familiar material;

feel reassured.

Recognition rises.

Retrieval may not.

Use a closed-book first pass.

From the chapter title, write:

definitions remembered;

main relationships;

conditions;

common representations;

one example;

one common error.

Then open the textbook.

Compare.

What was missing?

What was inaccurate?

What could not be reconstructed?

That difference defines revision.

Now sample exercises.

Do not redo every question.

Choose one or two from different layers:

direct;

changed;

mixed;

transfer.

If all are secure, move on.

If one fails, reopen that layer.

Use cumulative sampling.

For example:

one algebra question;

one ratio/percentage;

one graph;

one geometry;

one statistics/probability

where relevant to the learner’s course.

The textbook’s end-of-chapter and review sections are useful here because they often remove immediate example cues.

But learners should still mix across chapters.

The worldwide revision owner How to Revise for Maths develops finite-horizon revision more deeply.

The textbook-specific rule is:

during revision, the book should increasingly verify memory rather than create it from zero every time.

Aisha now covers the summary page and tries to reconstruct it.

Jo uses chapter exercises only after attempting a mixed diagnostic.

Ben turns worked-example titles into retrieval questions:

“How do I recognise reverse percentage?”

“What condition authorises Pythagoras?”

“How is combined mean rebuilt from totals?”

Then he opens the book to verify.

This makes revision active.

23. Treat end-of-chapter review as a transfer test, not as another place to copy the chapter’s method sequence

End-of-chapter review is often the first point where a textbook tests whether the learner can choose among several methods.

Chapter exercises may be blocked.

Review sets may mix.

This changes the task.

During blocked work, the learner knows:

“I am in the section on factorisation.”

During review, they must decide:

factorise?

complete the square?

use a formula?

rearrange?

draw a graph?

model from words?

This is a method-selection test.

Use review sections without peeking at section headings or example labels.

For each question, write only the first decision before solving:

target;

representation;

candidate method;

condition that supports it.

Then solve.

Mark errors by layer.

Reading error.

Question was mistranslated.

Selection error.

Known method not chosen or wrong method chosen.

Execution error.

Route was correct; algebra/arithmetic failed.

Completion error.

Final answer form was incomplete.

Knowledge error.

Relationship genuinely unavailable.

This tells the learner whether to:

reread explanation;

review example;

do more direct practice;

compare methods;

or work on accuracy.

End-of-chapter review should also be delayed.

If the learner does it immediately after the last section, chapter context is still strong.

Return later.

Then mix the chapter review with older topics.

Ben marks review questions with a code:

D = direct execution issue;

S = selection issue;

R = reading issue;

K = missing knowledge.

He does not need a complicated spreadsheet.

The code changes what he revisits.

Jo used to look back at the chapter whenever she forgot the method during review.

She now waits longer and attempts an alternative representation first.

Aisha used to judge the review only by percentage correct.

She now asks whether the errors cluster around one dependency.

A review section is valuable because it removes some of the textbook’s scaffolding.

Treat it as a bridge toward independent Mathematics.

24. Use cumulative textbook reviews to test whether older Mathematics remains available after newer chapters have displaced it from attention

Cumulative review is different from chapter review.

It reaches further back.

That matters because Mathematics is cumulative.

A learner may become competent in one chapter and then forget it while concentrating on the next.

Cumulative review asks:

Can earlier relationships be retrieved?

Can methods be selected without chapter context?

Can prerequisites re-enter later topics?

Suppose the book has reached trigonometry.

A cumulative set may include:

fractions;

equations;

graphs;

ratio;

geometry;

and trig.

This is valuable because trig itself may depend on ratio and equation rearrangement.

Use cumulative review diagnostically.

If an old topic fails once, do not immediately launch a full re-teach.

Test a second changed example.

If retrieval returns quickly after a small cue, the issue may be temporary access.

If the learner cannot explain or execute the relationship, reopen repair.

Cumulative review also reveals interference.

Two methods may be confused because they look similar.

Examples:

ordinary percentage versus reverse percentage;

direct proportion versus linear with fixed fee;

Pythagoras versus trigonometry;

mean versus median;

with versus without replacement.

Mixed cumulative questions force discrimination.

Aisha used to avoid old topics because failures felt discouraging.

She now treats them as maintenance signals.

Jo revises only the newest chapter.

She now includes small older samples.

Ben uses the textbook’s cumulative review but supplements it with one question from a much earlier chapter if the book’s review is too narrow.

The important principle is:

learning is not durable because a chapter was once completed; it is durable when it remains usable after attention has moved elsewhere.

25. When a textbook explanation does not click, change the representation before concluding that the topic is impossible

No textbook explanation works equally well for every learner.

A book may be:

too compressed;

too formal;

too procedural;

too verbal;

too dependent on diagrams;

or simply assume a prerequisite the learner lacks.

If an explanation does not click, diagnose why.

Vocabulary problem.

A mathematical term is not understood.

Notation problem.

Symbols cannot be translated.

Prerequisite problem.

The new explanation uses an earlier relationship that is unstable.

Representation mismatch.

The book gives algebra but the learner needs a diagram or table first.

Too-large step.

The worked example compresses several decisions.

Too many examples, not enough explanation.

The learner can imitate but not explain.

Use representation changes.

Words → diagram.

Diagram → equation.

Equation → table.

Table → graph.

Abstract rule → simple numerical case.

General theorem → concrete example and non-example.

Suppose the textbook explains direct proportion algebraically:

y = kx.

The learner remains uncertain.

Use a table:

x: 1, 2, 3;

y: 4, 8, 12.

Ratio y/x remains 4.

Graph passes through origin.

Now return to y = 4x.

Suppose completing the square feels arbitrary.

Use algebra tiles or geometric square-completion imagery if appropriate to the learner’s level, then return to symbolic form.

Suppose probability trees feel confusing.

Start with a state table listing bag contents after each draw.

Then rebuild the tree.

Ben asks another source for one alternative representation, then returns to the textbook.

He does not permanently abandon the book after one difficult paragraph.

Jo reads a teacher explanation or asks AI for a simpler example.

Aisha goes back one prerequisite section.

If several representations still fail, the learner may need direct teaching rather than more solitary reading.

A textbook is a tool.

Difficulty using one explanation does not prove inability in the Mathematics.

26. Do not collect multiple textbooks unless each one has a different job

Students sometimes respond to difficulty by collecting resources.

One school textbook.

One tuition book.

One revision guide.

One advanced problem book.

Several online worksheets.

The library grows.

The learning system becomes fragmented.

More books are useful only when they provide different functions.

One book may explain clearly.

Another may have stronger mixed exercises.

Another may provide assessment-style questions.

Another may be a concise reference.

Assign roles.

Core learning text.

Primary source for definitions, chapter sequence and worked examples.

Practice source.

Additional targeted questions where the core book has too little volume or variation.

Mixed/assessment source.

Questions where topic labels disappear and performance conditions matter.

Reference source.

Quick lookup for notation, formulas, theorems or summaries.

Do not read four explanations of the same easy method just because four books are available.

That can create recognition without practice.

Likewise, do not switch books every time one example feels hard.

First diagnose why it is hard.

If the issue is a missing prerequisite, changing books may merely change the wording of the same gap.

Aisha keeps buying revision guides because each new one creates the feeling of a fresh start.

Her new rule is:

one core text;

one supplementary source only if a specific job is missing.

Jo uses a difficult advanced book before her core chapter is stable.

She now uses it only for transfer after the textbook relationship is secure.

Ben compares books occasionally to see alternative methods.

He does not maintain separate complete note systems for each.

When two books use different notation or conventions, record the equivalence once.

Then follow the course-standard form where assessment requires it.

The principle is:

resources should have non-overlapping jobs; otherwise they multiply pages faster than they multiply learning.

27. A tutor should use the textbook to reveal learner decisions, not simply work through the book on the learner’s behalf

Tutors often use textbooks because they provide sequence and exercises.

The risk is that the tutor becomes the narrator of every page.

The learner watches.

Homework feels easier.

Independent use of the book does not improve.

A tutor can teach textbook independence.

At the start of a chapter, ask the learner to map it.

Which headings look new?

What prerequisite does the chapter appear to assume?

Which definition looks decisive?

Before a worked example, ask the learner to predict.

After the example, close it and give the paired exercise.

During marking, ask the learner to locate the first disagreement.

Before a solution is shown, ask what kind of help would be sufficient.

After a correct exercise, ask whether another question is needed or whether the learner should move to variation.

The tutor should also model resource judgement.

“This section is secure; skip five routine questions and try the mixed one.”

“This definition matters; do not skip it.”

“The answer key tells us the result is wrong, but we need to find the first invalid line.”

“This worked example is too close to your exercise; close it before trying.”

“Your algebra is fine. Go back to the diagram condition, not the algebra chapter.”

This teaches the learner how to use the book after tuition ends.

A tutor should avoid making every textbook decision for the learner.

Let the learner choose which exercise to attempt next and justify why.

Let them decide whether they need another direct example or a mixed question.

Let them identify which earlier section might repair a gap.

Then correct the resource decision if needed.

The long-term goal is a learner who can open an unfamiliar chapter and organise their own learning without waiting for an adult to curate every page.

28. Parents can support textbook study by asking about the learning job, not by measuring progress only in pages completed

Page counts are visible.

“How many pages did you do?”

“How many questions did you finish?”

“Did you complete the chapter?”

These questions are understandable.

They can push the learner toward quantity rather than control.

Better home questions include:

What was the main idea of the chapter?

Which definition mattered?

What type of example was new?

Which exercise first became difficult?

What was the first reason it failed?

Can you do one changed question without the book open?

What will you return to later this week?

These questions do not require the parent to know the Mathematics.

They make the study process visible.

Parents can also protect the learner from resource overload.

Before buying another book, ask:

What job is missing from the current book?

Does the learner need more explanation?

More direct practice?

More mixed questions?

More exam-style problems?

If the current book already contains unused material for that job, adding another resource may not help.

Parents can look for evidence of healthy textbook use.

The learner attempts before checking answers.

They skip secure repetitive questions without guilt when appropriate.

They return to definitions when language matters.

They can explain why an example works.

They use solutions to repair one gap rather than copy the route.

They can close the book and retrieve.

They can leave the textbook and solve a changed problem.

Aisha’s parent used to reward chapter completion.

Now they ask for one no-book problem after study.

Jo’s parent worried when she skipped easy questions.

Now the question is whether she has evidence that the direct layer is secure.

Ben’s parent does not monitor every answer.

They help maintain study consistency and resource discipline.

The strongest home support helps the textbook become a learner tool rather than a family supervision device.

29. Twelve common textbook-study failure modes and their repairs

Failure 1 — Start with exercises, skip definitions

Symptom: learner knows procedures but misuses terms and conditions.

Repair: read definitions with example/non-example before high-volume practice.

Failure 2 — Read every example without attempting

Symptom: chapter feels easy; independent exercise feels unfamiliar.

Repair: cover-and-predict, then attempt a paired problem after each representative example.

Failure 3 — Keep the worked example visible during every exercise

Symptom: high supported accuracy, poor no-cue performance.

Repair: fade support; close example before attempt.

Failure 4 — Do every question in order regardless of evidence

Symptom: hours of repetitive work with little new information.

Repair: sample direct questions, then select variation/mixed work; increase volume only where errors justify it.

Failure 5 — Choose only the hardest questions

Symptom: repeated noisy failure; unclear whether the new concept or prerequisites are weak.

Repair: climb a difficulty ladder so each layer becomes diagnosable.

Failure 6 — Check answers too early

Symptom: learner continues only when the book confirms each state.

Repair: make a meaningful attempt and mark uncertainty before checking.

Failure 7 — Read full solutions immediately after a wrong answer

Symptom: learner recognises corrections but repeats the same gap.

Repair: diagnose first; read only enough solution to repair; close and reconstruct.

Failure 8 — Copy notes from the textbook word for word

Symptom: beautiful notes, weak retrieval.

Repair: compress to definitions, conditions, relationship maps, contrasts and retrieval prompts.

Failure 9 — Use formula pages before attempting retrieval

Symptom: learner finds formulas but cannot recognise when they apply.

Repair: attempt to name relationship and condition before opening the reference page.

Failure 10 — Treat chapter completion as mastery

Symptom: learner finishes every exercise but old chapters disappear in mixed work.

Repair: delayed retrieval, cumulative review and changed-surface transfer.

Failure 11 — Collect multiple books without assigning roles

Symptom: repeated explanations, unfinished exercise sets, fragmented notes.

Repair: one core text plus supplementary sources only for specific missing functions.

Failure 12 — Never leave the textbook

Symptom: learner succeeds only on book-shaped questions.

Repair: use teacher questions, mixed sets, unfamiliar contexts and no-book retrieval to test whether the Mathematics travels.

These failures share one theme.

The textbook has begun carrying a responsibility the learner should eventually own.

The repair is to return that responsibility gradually.

30. The complete textbook system: use the book intensively, then make it progressively less necessary

Return to Jo, Aisha and Ben.

Jo began by treating the textbook as an exercise warehouse.

She became better at reading definitions, seeing chapter architecture and using answers diagnostically.

Aisha began by reading and copying everything.

She learned that recognition and beautiful notes are not the same as retrieval and transfer.

Ben began with a stronger system but still learned an important final rule:

the textbook should become less necessary as control increases.

The complete workflow is:

MAP THE CHAPTER → IDENTIFY THE BIG IDEA → LOCATE PREREQUISITES → READ DEFINITIONS FOR LEGAL MEANING → TRANSLATE NOTATION → READ EXPLANATIONS FOR RELATIONSHIPS → COVER AND PREDICT WORKED EXAMPLES → SELF-EXPLAIN STRATEGIC STEPS → ATTEMPT A PAIRED PROBLEM WITH THE EXAMPLE CLOSED → SELECT EXERCISES BY PURPOSE → CLIMB FROM DIRECT TO VARIED TO MIXED → CHECK ANSWERS AFTER A REAL ATTEMPT → DIAGNOSE THE FIRST FAILURE → USE FULL SOLUTIONS ONLY TO REPAIR A SPECIFIC GAP → RECONSTRUCT WITHOUT THE SOLUTION → TAKE ONLY FUNCTIONAL NOTES → RETRIEVE BEFORE REREADING → USE CHAPTER AND CUMULATIVE REVIEWS → CHANGE THE SURFACE → MIX WITH NEIGHBOURS → RETURN AFTER DELAY → LEAVE THE BOOK AND TEST WHETHER THE MATHEMATICS SURVIVES.

Frequently asked questions follow from this system.

Should I read the whole chapter before doing questions?

Usually not. Map the chapter, read enough definition/explanation to understand the new relationship, then alternate worked-example analysis with attempts. Long passive reading can create familiarity without performance.

Should I do every textbook question?

Not automatically. Use enough direct practice to stabilise the method, then sample variation, near neighbours, mixed and transfer questions. Increase volume where evidence shows a real weakness.

Is it cheating to look at the worked example?

No. Worked examples are instructional tools. The issue is whether support fades. Use the example to learn the route, then close it and reconstruct on a changed problem.

When should I check the answer?

After a meaningful attempt. If you check too early, you lose evidence about whether you could continue independently.

What if the answer key says I am wrong but I cannot see why?

Re-read the target, check transcription, identify the first high-risk transition and try an independent verification. If still unresolved, consult a solution or teacher for the specific disagreement.

Should I copy worked examples into my notes?

Only when the copy has a function. A short skeleton example plus the key decision and common error is often more useful than reproducing the entire printed solution.

How do I know I understand a definition?

You should be able to paraphrase it accurately, give an example, reject a tempting non-example and use the definition to justify a later decision.

Can a textbook replace a teacher or tutor?

Sometimes a learner can study substantial material independently. Other topics or learner states benefit from explanation, feedback or diagnosis. The relevant question is whether the book is transferring control or whether the learner remains stuck despite careful use.

How do I use a textbook when I am behind?

Use it surgically as a dependency library. Return to the earliest active blocker, repair only what current work requires, then reconnect to the current chapter rather than restarting the whole book by default.

How should I revise from a textbook?

Retrieve before rereading. Use summaries to check what you forgot, sample exercises across difficulty levels, and mix older chapters rather than reading everything again from the beginning.

Should I use AI with the textbook?

Yes, if AI clarifies, asks questions, generates variation or critiques your attempt. Avoid making it produce every solution before you have translated and attempted the problem yourself.

How do I know when to stop using the textbook for a topic?

When you can retrieve the relationship after delay, solve changed and mixed problems, select the method without chapter cues, check your work and transfer the Mathematics to questions outside the book. At that point the book moves from active teaching tool to occasional reference.

This page stops at textbook learning orchestration.

For the wider study architecture, use How to Study Maths Effectively.

For deliberate practice, use How to Practise Maths Effectively.

For worked-example fading, use How to Use Worked Examples in Maths.

For retrieval and spacing, use How to Remember Maths.

For accumulated gaps, use How to Catch Up in Maths.

The textbook has succeeded when the learner can close it and still carry the Mathematics.

Public references: What Works Clearinghouse — Organizing Instruction and Study to Improve Student Learning; What Works Clearinghouse — Teaching Strategies for Improving Algebra Knowledge in Middle and High School Students; NCETM — Five Big Ideas in Teaching for Mastery.

Appendix A — Textbook-use checkpoint: twenty-four study decisions

This checkpoint is original instructional material. It is not a standardised assessment. Its purpose is to reveal whether the learner is using the textbook actively or allowing the book to carry decisions that should gradually become independent.

For each case ask:

What is the textbook doing? What is the learner doing? Which responsibility should transfer next?

Case 1 — The learner skips definitions and starts Exercise 1

Chapter: direct proportion.

The learner can substitute into y = kx but later treats y = 3x + 5 as direct proportion.

Diagnosis: procedural work began before the legal definition was secure.

Repair: return to the definition; create one example and one non-example; explain why nonzero intercept changes the relationship.

Next textbook action: attempt two classification questions before continuing the exercise set.

Transfer test: recognise direct proportion from table, equation and graph.

Case 2 — The learner copies every definition word for word

The notebook contains exact textbook sentences.

When asked for an example or non-example, the learner hesitates.

Diagnosis: copying has preserved text but not classification power.

Repair: paraphrase accurately, then add example/non-example and the decisive condition.

Next textbook action: close the definition and classify unfamiliar cases.

Transfer test: use the definition in a later problem without reopening the book.

Case 3 — The learner reads a worked example straight through

The example feels obvious while visible.

The paired exercise feels impossible.

Diagnosis: recognition without route generation.

Repair: cover the next line and predict strategic steps.

Next textbook action: repeat the example with lower lines hidden.

Transfer test: solve a changed paired problem with the example closed.

Case 4 — The learner pauses after every arithmetic line

Textbook study becomes extremely slow even though arithmetic is secure.

Diagnosis: prediction is being applied at too fine a resolution.

Repair: predict only representation, method choice, condition use, major transformation and interpretation.

Next textbook action: mark high-value transitions in one example.

Transfer test: reproduce the route skeleton from memory.

Case 5 — The worked example remains open beside the exercise

Accuracy is high; no-example accuracy is low.

Diagnosis: example dependence.

Repair: close the example after reading; if stuck, look only at the first missing decision, then close it again.

Next textbook action: use a hint ladder rather than permanent visibility.

Transfer test: changed question after delay.

Case 6 — The learner finishes every direct exercise but avoids review questions

Diagnosis: execution is being trained; method selection is not.

Repair: reduce routine volume once accuracy is stable and add mixed/unlabelled questions.

Next textbook action: select one near-neighbour and one review problem.

Transfer test: identify method before solving a mixed question.

Case 7 — The learner jumps straight to the hardest problem

They fail and cannot tell whether the concept, notation, prerequisite or transfer demand caused the problem.

Diagnosis: difficulty is too entangled for clean feedback.

Repair: climb direct → varied → mixed → transfer.

Next textbook action: choose one problem at each relevant level.

Transfer test: move upward only when the previous layer is stable.

Case 8 — The learner checks the answer after every line

Diagnosis: external confirmation has replaced self-monitoring.

Repair: finish a meaningful attempt before checking the answer key.

Next textbook action: mark the exact point of uncertainty instead of checking continuously.

Transfer test: complete a short set with answers hidden until the end.

Case 9 — The answer matches, but the route used an invalid theorem

Example: Pythagoras was used on a triangle with no right-angle evidence, but numbers happened to produce the textbook answer.

Diagnosis: result agreement has been mistaken for method validity.

Repair: audit theorem conditions independently of the answer.

Next textbook action: identify given/derived/looks-like information.

Transfer test: near-neighbour diagrams where only some authorise the theorem.

Case 10 — The answer is wrong, so the learner reads the full solution immediately

Diagnosis: diagnostic evidence is being discarded too early.

Repair: re-read target, check transcription, inspect first high-risk transition, then consult solution only if needed.

Next textbook action: locate the first disagreement before revealing the route.

Transfer test: repair and solve a changed version without solution access.

Case 11 — The full solution is read and copied into notes

The learner understands it while looking.

Later they cannot reconstruct it.

Diagnosis: solution exposure without retrieval.

Repair: close solution and rebuild the route from a blank page.

Next textbook action: compare only after reconstruction.

Transfer test: new problem with the same invariant.

Case 12 — The learner copies the whole chapter summary

Diagnosis: notes duplicate the book rather than transforming it.

Repair: compress to definitions, conditions, relationship maps, contrasts and retrieval questions.

Next textbook action: limit the summary to one page or one functional map.

Transfer test: close both book and notes and retrieve from the map prompts.

Case 13 — Formula page is opened before the learner tries to remember the relationship

Diagnosis: textbook reference is functioning as permanent external memory.

Repair: attempt retrieval first: name the formula, variables and condition from memory.

Next textbook action: open the reference page only to verify or repair what was missing.

Transfer test: mixed question where the learner must decide which relationship applies without formula-page cues.

Case 14 — The formula is copied correctly but the variables are mapped wrongly

Quadratic formula or trigonometric relationship is written correctly; signs or side roles are substituted incorrectly.

Diagnosis: formula memory is stronger than problem-to-variable mapping.

Repair: identify coefficients/side roles before substitution.

Next textbook action: annotate one example by mapping every symbol to the problem.

Transfer test: changed coefficients, signs or rotated diagram.

Case 15 — The learner ignores the textbook’s diagrams and reads only prose

Diagnosis: part of the mathematical statement is being skipped.

Repair: identify markings, axes, scale, labels and stated correspondence before reading the solution.

Next textbook action: explain what the diagram contributes that the prose does not.

Transfer test: reconstruct the essential diagram from the written conditions.

Case 16 — The learner redraws every diagram perfectly

Study time is consumed by copying rather than reasoning.

Diagnosis: visual reproduction is not selective.

Repair: redraw only when it simplifies or isolates the relevant substructure.

Next textbook action: create a minimal diagram containing only target, conditions and needed labels.

Transfer test: solve from the simplified redraw.

Case 17 — Textbook and teacher notes use different-looking methods

Diagnosis: learner treats representation difference as mathematical contradiction.

Repair: identify the invariant relationship and compare the operations each source is performing.

Next textbook action: create a one-time source comparison: textbook method, teacher method, common structure.

Transfer test: use either valid method on a fresh problem and explain why it works.

Case 18 — The learner opens AI before attempting the textbook exercise

Diagnosis: the textbook’s intended transfer has been bypassed.

Repair: learner-first attempt with target, representation and first move.

Next textbook action: use AI only for bounded clarification or critique.

Transfer test: changed problem completed without AI route generation.

Case 19 — AI explanation conflicts with the textbook’s notation or course convention

Diagnosis: source authority is unclear.

Repair: compare mathematical meaning; follow official course/textbook/teacher conventions where assessment requires them.

Next textbook action: annotate the convention once and continue consistently.

Transfer test: produce the course-appropriate notation independently.

Case 20 — Calculator output is copied without a visible expression

Diagnosis: device execution is hiding mathematical structure.

Repair: write the intended expression or substituted formula before evaluation.

Next textbook action: compare output with sign, magnitude, units and domain.

Transfer test: another calculator-supported problem where grouping matters.

Case 21 — The learner is behind and restarts the textbook from Chapter 1

Diagnosis: catch-up is organised by book order rather than dependency evidence.

Repair: identify the earliest active blocker and return only to that section.

Next textbook action: repair a narrow prerequisite, then reconnect to current chapter.

Transfer test: use the repaired skill inside present schoolwork.

Case 22 — Revision means rereading textbook summaries

Diagnosis: recognition is being mistaken for retrieval.

Repair: reconstruct chapter summary before opening it.

Next textbook action: compare memory with printed summary and target omissions.

Transfer test: mixed exercise after delay.

Case 23 — The learner finishes the chapter review with a high score and stops

Diagnosis: chapter context may still be carrying method selection.

Repair: return later and mix with older chapters.

Next textbook action: use cumulative review or self-selected cross-chapter set.

Transfer test: identify methods without the chapter title as a cue.

Case 24 — The learner can solve textbook questions but not teacher/unseen questions

Diagnosis: transfer is textbook-bound.

Repair: leave the book after direct competence and use changed wording, representation and context.

Next textbook action: use the book only as reference when a specific relationship needs review.

Transfer test: unfamiliar question with the same mathematical invariant.

How to interpret the checkpoint

Do not total the cases into a textbook-use score.

Identify the first responsibility the learner is not yet carrying.

Meaning responsibility: definitions and notation remain external.

Route responsibility: examples supply every decision.

Selection responsibility: exercise headings supply the method.

Feedback responsibility: answer keys certify every step.

Memory responsibility: formula pages and summaries remain open.

Transfer responsibility: textbook surface features remain necessary.

The next study change should return one responsibility to the learner at a time.

Appendix B — Textbook routing matrix: study symptom → likely misuse → smallest repair → independence test

This matrix is not a universal rulebook. It is a way to turn vague textbook frustration into a specific learner action.

Symptom: “I read the chapter but nothing sticks”

Likely misuse: reading is recognition-heavy and retrieval-light.

Smallest repair: after each short explanation, close the page and state the relationship from memory.

Independence test: reconstruct the chapter’s big idea before reopening the book the next day.

Symptom: “The example makes sense but I cannot do the exercise”

Likely misuse: worked example has remained too complete.

Smallest repair: cover and predict the next strategic line, then attempt one paired problem with the example closed.

Independence test: solve a changed problem after a short delay.

Symptom: “I keep checking the answer because I don’t trust myself”

Likely misuse: answer key is functioning as continuous reassurance.

Smallest repair: define a minimum attempt boundary—target, representation, method and several lines—before checking.

Independence test: complete a three-question set with answers hidden until the end.

Symptom: “I got the right answer but the teacher says my method is wrong”

Likely misuse: result agreement is being treated as proof of valid reasoning.

Smallest repair: audit conditions, notation and transformations independently of the answer.

Independence test: explain why each strategic step is legal on a new question.

Symptom: “I copy full solutions but make the same mistake later”

Likely misuse: solution reading without reconstruction.

Smallest repair: after reading only enough to locate the gap, close the solution and rebuild from a blank page.

Independence test: changed example with no solution access.

Symptom: “My notes are huge and revision still takes forever”

Likely misuse: notes duplicate the textbook.

Smallest repair: keep only definitions, conditions, relationship maps, error patterns and retrieval prompts.

Independence test: use one-page notes to reconstruct the chapter without reopening the book.

Symptom: “I know formulas but never know which one to use”

Likely misuse: formula lookup has replaced condition-to-method mapping.

Smallest repair: for each formula, record the target it helps with and the conditions that authorise it.

Independence test: classify mixed problems before calculating.

Symptom: “I always run out of questions in the textbook”

Likely state: the book may provide too little variation for the learner’s current need.

Smallest repair: generate or source changed-surface and mixed problems rather than rereading examples.

Independence test: transfer to a new source with the book closed.

Symptom: “I have three textbooks and don’t know which one to use”

Likely misuse: resources have overlapping, undefined roles.

Smallest repair: assign one core-learning text, one supplementary practice role if needed and one assessment/mixed source if needed.

Independence test: complete a study week without switching sources impulsively.

Symptom: “I restart the chapter every time I get stuck”

Likely misuse: local gap is being treated as total chapter failure.

Smallest repair: identify the first failed dependency and return only to that explanation/example.

Independence test: repair and re-enter the original exercise without rereading the whole chapter.

Symptom: “The chapter is easy, but mixed tests are hard”

Likely misuse: blocked chapter context supplies method selection.

Smallest repair: use chapter reviews and cross-chapter mixed sets with headings hidden.

Independence test: state method and reason before solving each mixed question.

Symptom: “The textbook is too wordy”

Likely issue: learner may need to extract the mathematical relationship rather than read every sentence equally.

Smallest repair: after each paragraph, write one sentence: “This is really saying that…”

Independence test: explain the relationship using a different example.

Symptom: “The textbook is too terse”

Likely issue: several decisions are compressed into one line.

Smallest repair: expand the missing steps using another representation, teacher explanation or bounded AI clarification.

Independence test: compress back to the textbook’s concise form only after the route is understood.

Symptom: “I highlight everything”

Likely misuse: highlighting is replacing discrimination.

Smallest repair: replace highlight with a question or one-line relationship note.

Independence test: retrieve the point without looking at the highlight.

Symptom: “I do well when the worked example is nearly identical”

Likely misuse: surface matching rather than structural recognition.

Smallest repair: change numbers, wording, orientation and unknown direction one at a time.

Independence test: identify the invariant before solving the changed version.

Symptom: “I can solve examples but not word problems”

Likely issue: translation/model formation, not necessarily execution.

Smallest repair: use textbook word problems only to define variables and form equations at first.

Independence test: changed context with the same algebraic structure.

Symptom: “I can do exercises but forget after a week”

Likely misuse: chapter completion without delayed retrieval.

Smallest repair: return after delay before rereading.

Independence test: mixed old-topic problem later in the month.

Symptom: “The textbook answer is different from mine, so I assume I’m wrong”

Likely misuse: one printed route is being treated as the only valid route.

Smallest repair: verify your method against definitions, conditions and the original problem.

Independence test: compare two valid methods and explain trade-offs.

Symptom: “My calculator says the textbook answer is wrong”

Likely issue: device entry, mode, rounding or transcription may differ.

Smallest repair: write and compare the mathematical expression before blaming either source.

Independence test: estimate magnitude and verify with an alternative calculation where possible.

Symptom: “The textbook uses a method my teacher does not”

Likely issue: different valid representations, or course-specific constraints.

Smallest repair: identify whether methods are mathematically equivalent and whether local assessment expects one form.

Independence test: use the course-approved route independently while understanding the alternative.

Symptom: “I don’t know which questions to skip”

Likely misuse: exercise selection has no evidence rule.

Smallest repair: sample two or three direct questions; if accurate and independent, move to variation/mixed. If not, increase direct volume.

Independence test: justify why the next selected question adds new evidence.

Symptom: “I skip all easy questions and make many mistakes later”

Likely misuse: learner assumes conceptual familiarity equals stable execution.

Smallest repair: use a small direct diagnostic before harder questions.

Independence test: advance only if execution is genuinely stable.

Symptom: “I only do the easy questions”

Likely misuse: textbook is being used for reassurance rather than boundary growth.

Smallest repair: add one changed or mixed question beyond the direct layer.

Independence test: recover from one unfamiliar variation without immediate help.

Symptom: “I cannot study from the textbook unless someone sits beside me”

Likely issue: resource-use decisions remain externally owned.

Smallest repair: learner chooses one chapter objective, one example, one paired exercise and one review question before adult help begins.

Independence test: run a short textbook study block alone and bring only specific unresolved questions afterward.

Symptom: “I understand the chapter but don’t know when I’m done”

Likely issue: no exit criteria.

Smallest repair: define a chapter receipt: retrieve definition, solve changed problem, solve mixed problem, explain one condition, return after delay.

Independence test: succeed outside the textbook on a fresh problem.

The compact textbook-routing rule

Ask:

Do I know the meaning?

Can I predict the route?

Can I solve with the example closed?

Can I choose the method without the section heading?

Can I mark and repair from evidence?

Can I retrieve after delay?

Can I solve beyond the textbook?

The first “no” tells you what the book should do next.

Appendix C — Eight composite textbook laboratories: from printed chapter to independent Mathematics

These laboratories show how the full textbook workflow changes by topic and learner state. Each one begins with a typical chapter structure, then follows the learner through mapping, definition, worked example, exercise selection, marking, repair, retrieval and transfer.

Laboratory 1 — Algebra chapter: expansion, factorisation and equations

Chapter map.

The learner scans the chapter and identifies:

expanding single brackets;

double brackets;

factorising common factors;

factorising simple quadratics;

solving equations from factor form;

mixed review.

The map reveals a dependency:

expansion and factorisation are inverse form changes.

Factor form can expose roots when an equation equals zero.

Definition/notation pass.

The learner clarifies:

term;

factor;

coefficient;

expression;

equation.

This matters because the book later says “factorise the expression” and “solve the equation”.

Those are different jobs.

Worked example.

The textbook shows:

x² − 7x + 12

= (x − 3)(x − 4).

The learner self-explains:

need numbers with product +12 and sum −7;

−3 and −4.

Then predicts the next equation example:

x² − 7x + 12 = 0

becomes

(x − 3)(x − 4) = 0

so x = 3 or 4.

Paired problem.

x² − 9x + 20 = 0.

Book closed.

The learner finds 4 and 5, factors and solves.

Exercise selection.

Instead of doing twenty near-identical questions, select:

two direct positive-constant quadratics;

two with negative constant;

one where leading coefficient is not 1 if the course includes it;

one equation where factorisation must be recognised rather than requested;

one word/model problem producing a quadratic.

Marking.

One error appears:

x² + x − 12 becomes (x + 4)(x + 3).

Expand to check:

x² + 7x + 12.

Mismatch.

First failure:

sign/product-sum pairing.

Repair.

Use factor-pair table for −12:

1,−12;

−1,12;

2,−6;

−2,6;

3,−4;

−3,4.

Need sum +1 → −3 and 4.

So:

(x − 3)(x + 4).

Changed retest.

x² − x − 20.

Delayed retrieval.

Several days later:

factor and solve x² + 2x − 15 = 0.

Transfer beyond textbook.

Use a rectangle-area word problem that produces the same quadratic structure.

The learner must form the equation first.

Exit receipt.

Factorisation is no longer just a chapter exercise. The learner can choose factor form when it exposes roots and verify by expansion.

Laboratory 2 — Percentage chapter: direct, reverse and compound change

Chapter map.

Sections:

percentage of a quantity;

increase/decrease;

multipliers;

reverse percentage;

compound change;

applications.

Big idea:

percentage is a relationship to a reference quantity.

Definition pass.

The learner distinguishes:

percentage;

percentage point;

original/reference value;

multiplier.

Worked example.

Original price 200, discount 15%.

Final = 0.85×200 = 170.

Cover and predict reverse version:

final 170 after 15% discount.

0.85P = 170.

P = 200.

Self-explanation.

Why divide by 0.85?

Because 170 represents 85% of the unknown original.

Paired problem.

After 20% increase, final value 96.

1.2P = 96.

P = 80.

Exercise selection.

two forward;

two reverse;

one percentage-point contrast;

two compound changes;

one ratio/percentage mixed question.

Marking.

Learner writes:

increase 20%, then decrease 20% → no change.

Textbook answer disagrees.

Diagnose before solution.

Multiplier:

1.2×0.8=0.96.

Final is 96% of original.

First failure:

treating percentage changes as additive on one fixed base.

Repair.

Stage map with changing bases.

Changed retest.

increase 10%, then decrease 10%.

1.1×0.9=0.99.

Delayed retrieval.

Book closed: explain why equal percentage increase/decrease do not normally cancel.

Transfer beyond textbook.

Population, salary and concentration contexts.

Exit receipt.

Learner identifies the 100% base and direction before arithmetic.

Laboratory 3 — Geometry chapter: conditions, diagrams and theorem use

Chapter map.

Angle facts;

parallel lines;

triangles;

Pythagoras;

similarity;

applications.

Big idea:

geometric methods are authorised by conditions, not appearance.

Definition/notation pass.

Parallel markings;

right-angle mark;

congruent/equal-length ticks;

similarity notation;

vertex correspondence.

Worked example.

A diagram shows AB ∥ CD.

The book finds an angle using alternate angles.

Cover the solution and ask:

What condition makes alternate angles available?

Parallel lines.

Which lines are parallel?

AB and CD.

Paired problem.

Rotate the diagram.

Keep markings.

The learner must identify the same relationship from conditions rather than page orientation.

Similarity example.

△ABC ~ △DEF.

Predict corresponding sides from vertex order.

Do not use visual position.

Exercise selection.

one direct angle fact;

one multi-step angle chain;

one misleading not-to-scale diagram;

one similarity correspondence;

one length scale;

one area-scale question.

Marking.

Learner uses Pythagoras on a triangle that merely looks right-angled.

The numerical answer happens to match a distractor.

Answer key alone cannot diagnose the theorem error if only final answers are shown.

Audit condition.

No right-angle evidence.

Repair.

Create three diagram categories:

right angle given;

right angle derivable;

right-looking only.

Changed retest.

New triangle with explicit right-angle mark in an unusual orientation.

Delayed retrieval.

State the condition needed for Pythagoras and similarity before solving.

Transfer beyond textbook.

Teacher-created unfamiliar geometry with same conditions but different layout.

Exit receipt.

Learner chooses geometry methods from evidence, not appearance.

Laboratory 4 — Graph chapter: axes, gradient, intercept and modelling

Chapter map.

coordinates;

straight-line graphs;

gradient;

intercept;

equation y=mx+c;

applications;

intersections.

Big idea:

a graph represents relationships among quantities; axes and scale create meaning.

Worked example.

Points (2,5) and (6,13).

Gradient:

(13−5)/(6−2)=8/4=2.

Cover and ask:

What must remain consistent?

Point order in numerator and denominator.

Paired problem.

(−1,7) and (4,−3).

Gradient:

(−3−7)/(4−(−1)) = −10/5 = −2.

Interpretation.

Book gives a cost-distance graph.

Gradient = 3 dollars/km;

intercept = fixed 5-dollar charge.

Ask:

Would the same 3 and 5 mean the same thing on a temperature-time graph?

No.

Exercise selection.

one coordinate-reading problem with non-unit scale;

one gradient calculation;

one equation from graph;

one graph from equation;

one contextual interpretation;

one intersection.

Marking.

Learner calculates gradient from square counts, not axis labels.

First failure:

scale reading.

Repair.

Read axes → units → scale → coordinates before formula.

Changed retest.

Same visual slope, different axis scale.

Delayed retrieval.

Explain gradient units and intercept meaning on a new context.

Transfer beyond textbook.

Compare two real-world linear models from a teacher worksheet.

Exit receipt.

The learner no longer reads graphs as pictures; they read represented quantities.

Laboratory 5 — Probability chapter: event language, changing state and efficient representation

Chapter map.

single-event probability;

complements;

combined events;

tree diagrams;

with and without replacement;

conditional probability;

mixed review.

The learner identifies the big idea:

probability depends on the event definition and the sample state.

Definition pass.

Event.

Outcome.

Sample space.

Independent.

Mutually exclusive where included by the course.

Conditional language.

Without replacement.

The book may define these concisely.

The learner creates examples and non-examples.

Worked example.

Bag contains 5 red, 3 blue.

Two counters drawn without replacement.

Example starts:

P(red first)=5/8.

Cover next line.

Predict state after red:

4 red, 3 blue, total 7.

So P(red second | red first)=4/7.

Then:

P(two reds)=5/8×4/7.

Self-explanation.

Why did denominator change?

Because one counter left the sample space.

Why did red numerator change?

Because the removed counter was red.

Paired problem.

Same bag, red then blue.

5/8×3/7.

Near-neighbour.

Same bag, with replacement.

5/8×5/8.

The one phrase changes the state model.

Exercise selection.

one direct event;

one complement;

one with replacement;

one without replacement;

one “at least one”;

one conditional denominator problem;

one mixed event representation.

Marking.

Learner writes “at least one success” as exactly one success.

First failure:

event language.

Do not practise more tree arithmetic yet.

Repair event set:

at least one = one or more;

at most one = zero or one;

exactly one = one.

Changed retest.

Three independent trials; probability of at least one success.

Recognise complement:

1 − P(no success).

Delayed retrieval.

Book closed, explain when branch probabilities change and when they stay the same.

Transfer beyond textbook.

Two-way table conditional probability.

The representation changes but denominator reasoning remains.

Exit receipt.

Learner defines the event and state before calculation.

Laboratory 6 — Statistics chapter: meaning before formula, exact versus estimated summaries

Chapter map.

mean;

median;

mode;

range;

quartiles/IQR where relevant;

frequency tables;

grouped data;

comparison of distributions.

Big idea.

Different statistics summarise different features of data.

Definition pass.

The learner writes one sentence for each measure:

mean uses all values through total/count;

median depends on ordered position;

range uses extremes;

IQR measures spread of the middle half under the course’s convention.

Worked example.

Two groups:

A: 10 students, mean 12.

B: 20 students, mean 18.

Textbook combines:

(10×12 + 20×18)/(10+20).

Cover and ask:

Why multiply mean by count?

Because total = mean×count.

Paired problem.

A: 8 students, mean 15;

B: 12 students, mean 21.

Reconstruct totals and combine.

Grouped-data example.

Intervals with frequencies.

The book uses midpoints to estimate a mean.

Ask:

Why estimate?

Because exact values inside each interval are unknown.

Exercise selection.

one raw-data mean/median;

one frequency table;

one combined mean;

one grouped estimated mean;

one comparison of centre/spread.

Marking.

Learner averages two group means equally despite unequal group sizes.

First failure:

mean relationship, not arithmetic.

Repair.

mean = total/count;

therefore total = mean×count.

Changed retest.

Use very unequal group sizes so simple averaging becomes clearly implausible.

Delayed retrieval.

Before opening textbook, explain why grouped mean is estimated and why combined mean may need weighting.

Transfer beyond textbook.

Teacher-provided data display or two-way table.

Exit receipt.

Learner chooses summaries from target/meaning rather than formula memory alone.

Laboratory 7 — Catch-up from the textbook: repairing one prerequisite without restarting the book

Current frontier.

The class is studying straight-line graphs.

Ben can interpret gradient and intercept but repeatedly fails to rearrange equations such as:

2x + y = 9

into

y = −2x + 9.

He considers restarting the algebra unit from the beginning.

Dependency map.

The visible failure is equation rearrangement.

Check earlier textbook section “Changing the subject / Rearranging formulas”.

Confirming tasks:

a + b = c, make b subject;

3x + y = 7, make y subject;

2y − x = 5, make y subject.

Results:

simple one-step rearrangement secure;

sign-sensitive multi-step rearrangement weak.

Minimum repair.

Read the earlier explanation of preserving equality.

Study one worked example.

Predict transformations.

Attempt four selected problems:

two direct;

one negative coefficient;

one division at the end.

Do not:

redo thirty pages of unrelated algebra.

Current reconnection.

Return immediately to:

2x + y = 9.

Subtract 2x:

y = −2x + 9.

Then graph.

Changed retest.

3x − 2y = 12, rearrange for y.

Delayed retrieval.

Several days later, rearrange a formula from another topic.

Transfer.

Use rearrangement in trigonometry or physics-style rate formula where relevant.

Exit receipt.

The old textbook section becomes reference/maintenance, not a restarted course.

This laboratory shows why textbook catch-up should follow dependencies rather than page order.

Laboratory 8 — Advanced chapter independence: when the textbook should be closed

Starting state.

Aisha has completed a chapter on quadratic graphs.

She can:

expand;

factorise;

solve roots;

identify intercepts;

complete the square;

find vertex form.

Her textbook review score is high.

Does she still need the chapter open?

Use an independence sequence.

Step 1 — closed-book chapter map.

Write:

what forms of a quadratic reveal;

expanded form;

factor form;

completed-square form.

She states:

expanded form shows coefficients;

factor form can expose roots;

completed-square form can expose vertex.

Step 2 — no-book definitions/conditions.

What are roots/x-intercepts?

What does discriminant tell you if included in course?

What does leading coefficient do to opening direction?

Step 3 — changed problem.

Given y = x² − 6x + 5, find roots and vertex using suitable forms.

No chapter cues.

Step 4 — representation transfer.

Given a graph, reconstruct likely algebraic information.

Given roots 1 and 5 with leading coefficient 1, form the quadratic.

Step 5 — mixed selection.

Put the quadratic among linear, exponential or other course-appropriate questions.

Ask which route applies and why.

Step 6 — delayed return.

One week later, solve a new problem without reviewing the chapter first.

Step 7 — unfamiliar context.

Use a rectangle or projectile/model context if appropriate to the curriculum.

Form and interpret the quadratic.

Aisha succeeds.

The textbook chapter has finished its active-teaching job.

It can now become:

reference;

maintenance exercise source;

or occasional explanation source if a new edge case appears.

Exit receipt.

“Quadratics: retrieve forms, choose representation, solve changed problems, transfer to context, one-week return secure. Move chapter to maintenance.”

This is the desired endpoint of textbook study.

The learner should not remain permanently attached to the pages that first taught the Mathematics.

What the eight laboratories reveal

A textbook chapter is not complete when the final exercise is ticked.

It is complete when the learner can:

state the definitions;

explain the relationships;

reconstruct worked routes;

select methods;

repair errors;

retrieve after delay;

and transfer outside the book.

The printed chapter is the launch environment.

Independent Mathematics is the destination.

Appendix D — The one-page textbook chapter dashboard

A dashboard should not become another large note system. Its purpose is to keep the learner oriented while the chapter is active and to make later retrieval more efficient. If the dashboard takes longer to build than the learning it supports, it is too large.

Chapter title

Write the official chapter name so the dashboard remains easy to reconnect to the textbook.

One-sentence big idea

Complete:

“This chapter is mainly about…”

Examples:

“representing constant-rate relationships with y = kx and distinguishing them from general linear relationships”;

“rewriting quadratics into forms that expose roots or vertex structure”;

“using event states and conditional information to model probability.”

If the learner cannot write the big idea after studying the chapter, they may have collected procedures without seeing the system.

Prerequisites

List only the earlier Mathematics actively required.

For trigonometry:

right-triangle recognition;

ratio meaning;

equation rearrangement;

calculator angle functions.

For combined mean:

mean = total/count;

multiplication;

fraction/decimal control.

For quadratic graphs:

coordinates;

substitution;

expansion;

factorisation where used.

Prerequisites should help diagnose later failures.

Definitions that change what is legal

Do not list every word.

Choose terms where an inaccurate definition would change method or classification.

For example:

direct proportion;

similar figures;

independent events;

function;

irrational number;

gradient;

median.

Write:

definition;

one example;

one near non-example.

Core relationships

Write formulas or structural statements with meaning.

Weak:

m = (y₂−y₁)/(x₂−x₁).

Stronger dashboard entry:

gradient = change in vertical quantity / change in horizontal quantity;

keep point order consistent;

interpret units.

Weak:

mean = Σx/n.

Stronger:

mean = total/count;

therefore total = mean×count;

use this to combine groups.

Representation map

What forms can the chapter appear in?

Words.

Equation.

Table.

Graph.

Diagram.

Tree.

Frequency table.

State which translations matter.

Example for linear models:

words “fixed fee plus rate” ↔ equation C = fixed + rate×usage ↔ straight-line graph with nonzero intercept.

Worked-example families

Do not list every example number.

Identify distinct learning jobs.

Example for percentage:

direct percentage;

percentage change;

reverse percentage;

compound change;

percentage-point comparison.

Example for geometry:

direct theorem use;

multi-step angle chain;

similarity correspondence;

length scale;

area/volume scale.

Current personal failure

Write one or two mechanisms only.

Examples:

“reverse percentage: I treat final as 100%”;

“similarity: I match by visual position after rotation”;

“equations: negative distribution fails”;

“graphs: I read square count instead of axis values.”

This makes the dashboard learner-specific.

Checking method

What is the cheapest useful check?

Equation:

substitute into original.

Factorisation:

expand.

Percentage:

run change forward/backward.

Graph:

sign/magnitude/intercept behaviour.

Geometry:

condition, angle sum, scale plausibility.

Exit test

Before moving the chapter to maintenance, the learner should be able to:

state key definitions;

retrieve the core relationship;

solve a direct problem;

solve a changed problem;

choose the method in mixed work;

return after delay;

and solve one problem outside the textbook.

Example dashboard — reverse percentage

Big idea: final amount is a percentage/multiplier of original; direction of unknown controls rearrangement.

Prerequisite: percentage multiplier, equation rearrangement.

Definition: original/base quantity = 100% reference.

Relationship: final = multiplier×original.

Representations: price, population, salary, growth graph.

Failure: I multiply final by multiplier instead of solving backward.

Check: apply percentage change forward to reconstructed original.

Exit: direct and reverse mixed correctly after delay.

Example dashboard — similar triangles

Big idea: established similarity creates equal corresponding angles and proportional corresponding lengths.

Prerequisite: ratio, angle facts, diagram reading.

Condition: similarity must be given or established under course-appropriate criteria.

Relationship: corresponding side ratios equal.

Representation: rotated/reflected diagrams.

Failure: I match sides by page position.

Check: vertex correspondence and consistent scale factor.

Exit: rotated mixed geometry without example support.

How the dashboard should fade

During first learning, the learner may fill most fields.

During revision, the dashboard becomes a retrieval prompt.

Later, only the personal failure and maintenance items may remain relevant.

The dashboard is successful when the learner can reconstruct the chapter architecture without rereading it.

Appendix E — Example-to-independence ladder: seven stages for transferring control from the textbook to the learner

This ladder is a practical way to see whether a worked example has finished teaching. It is not a validated scale. The learner can move up and down as needed. Different topics may sit at different stages at the same time.

Stage 1 — Recognition

The learner reads the worked example and can follow every line.

This is useful.

It is also the weakest form of ownership.

Questions to ask:

What is the target?

What representation did the author choose?

Which relationship is being used?

What conditions make it legal?

What is the final answer form?

Do not leave the learner at recognition.

Move immediately to prediction.

Stage 2 — Prediction

The learner can predict the next major step before reading it.

Examples:

factorise before solving roots;

update denominator after a without-replacement draw;

use a ratio because corresponding lengths are involved;

convert minutes to seconds because the rate is per second.

Prediction shows that some route generation has begun.

If prediction fails, use the example to clarify the relationship rather than simply reading onward.

Stage 3 — Completion

Part of the worked route is hidden.

The learner supplies missing lines or decisions.

Examples:

x² − 6x + 5

= x² − 6x + ___ − ___ + 5.

Or:

After red is drawn from 5 red and 3 blue without replacement:

P(second red) = ___.

Completion keeps the example structure visible while returning local responsibility.

Stage 4 — Paired independent problem

The learner solves a fresh problem with the same invariant and the example closed.

Surface features change:

numbers;

signs;

labels;

orientation;

context.

The method remains recognisable.

If the learner immediately reopens the example, the support has not faded enough.

Use a small hint rather than full route exposure where possible.

Stage 5 — Changed representation

The same relationship appears in another form.

Examples:

percentage relationship moves from money to population;

linear equation moves from symbols to word problem;

ratio moves from recipe to similar triangles;

mean moves from raw list to frequency table;

quadratic moves from equation to graph.

The learner must recognise structure rather than surface.

This is a crucial textbook exit stage because books often group similar representations together.

Stage 6 — Mixed selection

The learner encounters several neighbouring methods and must choose.

Examples:

Pythagoras versus trigonometry;

forward versus reverse percentage;

mean versus median;

factorisation versus another quadratic method;

with versus without replacement.

Now the learner has to answer:

what conditions and target make this route appropriate?

If they can execute but cannot choose, the example taught procedure but not discrimination.

Stage 7 — Delayed transfer without textbook cues

The relationship returns after time has passed and outside the original textbook sequence.

The learner should be able to:

retrieve the relationship;

choose it;

execute;

check;

and interpret the result.

This is the strongest practical evidence that the worked example has finished its active teaching job.

Example ladder — solving linear equations

Recognition: follow 3x+5=20 → 3x=15 → x=5.

Prediction: after 3x+5=20, predict subtract 5.

Completion: 4x−7=21 → 4x=___ → x=___.

Paired: 5(x−2)=30.

Changed representation: fixed-fee word problem produces 4h+7=31.

Mixed selection: equations mixed with ratio and percentage problems.

Delayed: use equation rearrangement inside a later graph or formula problem.

Example ladder — reverse percentage

Recognition: final 80 after 20% discount → 0.8P=80.

Prediction: next step P=80/0.8.

Completion: final 153 after 15% discount → 0.85P=___.

Paired: final 276 after 15% increase.

Changed representation: population or salary context.

Mixed selection: ordinary percentage, reverse percentage and percentage-point questions.

Delayed: return a week later with no percentage heading.

Example ladder — similarity

Recognition: follow side ratio in similar triangles.

Prediction: identify correspondence before ratio.

Completion: missing scale factor step.

Paired: different side lengths.

Changed representation: rotated or reflected triangles.

Mixed selection: similarity mixed with Pythagoras/trig.

Delayed: same invariant inside map/scale or geometry synthesis.

When to move down a stage

If transfer fails, do not shame the learner or repeat the whole chapter.

Move to the nearest supporting stage.

If mixed selection fails but paired problems are strong, practise near-neighbour discrimination.

If paired problems fail, return to completion or prediction.

If prediction fails, reopen the relationship explanation.

The ladder is a control system.

Its purpose is to make support proportional to the missing responsibility.

When to leave the textbook example behind

Leave active example support when the learner can solve changed and mixed problems, return after delay, and explain why the method applies.

The example may remain useful as occasional reference.

It should no longer be necessary for ordinary execution.

A worked example has finished teaching when the learner can generate the mathematical decisions the example used to display.

Appendix F — Exercise selection matrix: choose the next textbook question from evidence rather than habit

A fixed exercise set can still become adaptive if the learner chooses the next problem for a reason. The matrix below is a decision aid, not a compulsory sequence.

If direct accuracy is below about “mostly secure”

Do not rush into complex transfer.

Choose more direct questions with small variation.

Focus on one relationship at a time.

Use immediate feedback after a real attempt.

Example:

linear equations with one new feature at a time—brackets, negatives, variables on both sides.

The learner should be able to explain why each transformation preserves the equation.

If direct accuracy is high but one execution error repeats

Choose questions that isolate that transition.

If negative distribution is weak, select sign-sensitive brackets rather than more easy positive examples.

If calculator grouping is weak, select expressions where brackets matter.

If units are weak, use rate questions with conversions.

The purpose is not more volume everywhere.

It is more evidence at the fault line.

If execution is secure but the learner cannot choose the method

Move to near-neighbour and mixed questions.

Examples:

forward versus reverse percentage;

Pythagoras versus trigonometry;

direct proportion versus fixed-fee linear;

with versus without replacement.

Ask for method/condition before calculation.

If the learner knows the method only in one representation

Select a changed representation.

Equation → graph.

Words → equation.

Table → formula.

Diagram → ratio.

Frequency table → mean.

This reveals whether the learner owns the relationship or only one surface.

If the learner is fast but careless

Do not simply add harder problems.

Select moderate questions with high-risk transitions.

Require one lightweight check at the transition that historically fails.

Examples:

negative signs;

percentage base;

units;

calculator entry;

domain restriction.

The goal is prospective control.

If the learner is accurate but very slow

Choose short focused clusters of stable routines.

Do not time a brand-new concept.

Examples:

five substitutions;

five factorisations;

five fraction simplifications;

five coordinate readings.

Then re-embed the routine in full problems.

If the learner gets stuck only on long word problems

Select translation-only questions.

Do not solve them completely.

For each:

state target;

define variables;

write relationships;

mark conditions;

stop.

If those states are correct, the gap may lie later in method selection or execution.

If the learner can solve but cannot explain

Select one problem where the task is to justify a step, compare methods or reject a near neighbour.

Ask:

why is this theorem legal?

why does this form expose the target?

why is this estimate rather than exact?

what would make this method invalid?

If the learner has just repaired a gap

Do not leave the repair inside a special worksheet.

Select one direct retest, one changed-surface problem and one current-topic application.

Then return later.

Example:

repair fraction division;

retest 9÷3/4;

then solve (3/4)x=9;

then use a fractional coefficient inside a current graph/equation problem.

If the learner has high chapter scores but poor test scores

Select cumulative and mixed review questions.

Hide chapter headings.

Add ordinary time constraints only after accuracy is stable.

Track whether failure begins at reading, method selection, fluency or pacing.

If the learner keeps choosing only familiar questions

Require one boundary problem per study block.

It should be difficult enough to reveal the edge of current control, but not so overloaded that the failure becomes uninterpretable.

A boundary problem might change:

unknown direction;

context;

representation;

or mix one near-neighbour.

If the learner keeps choosing only hard questions

Require a quick direct diagnostic first.

If direct performance is not stable, repair that layer.

Hard questions should test transfer, not hide missing foundations inside several simultaneous demands.

If the textbook has too much repetition

Sample strategically.

For example:

questions 1, 3, 6, 10, 14 if they represent increasing variation.

Do not use fixed question numbers across all books; inspect actual content.

If a sampled problem fails, return and increase volume around that type.

If the textbook has too little practice

Create additional changed examples from the structure.

Change:

numbers;

signs;

unknown;

context;

representation;

or method neighbour.

Use another source only if a clear practice function is missing.

A three-question minimum evidence set

When time is short, use:

Question 1 — direct.

Can I execute the relationship?

Question 2 — changed.

Can I preserve it after one surface change?

Question 3 — mixed/unlabelled.

Can I recognise when it applies?

This is not enough for every skill or learner.

It is a compact sampling frame.

If all three are strong, the learner may move to delayed retrieval or broader transfer.

If one fails, the failure tells you what to select next.

The matrix’s final rule

The next textbook exercise should answer a question about the learner’s state.

If you cannot say what evidence the next question will provide, you may be doing volume without direction.

Appendix G — A six-week textbook learning arc: from first chapter contact to independent transfer

This six-week arc is illustrative, not a universal calendar. A learner may move faster or slower. The purpose is to show how textbook dependence should reduce as control grows.

Week 1 — Map and decode

Choose one active chapter.

Build the one-page chapter dashboard.

Identify:

big idea;

prerequisites;

definitions;

notation;

core relationships;

worked-example families.

Do not attempt to finish the whole exercise set.

Study one or two representative examples using cover-and-predict.

After each, attempt a paired problem with the example closed.

At the end of the week, answer:

What is the chapter really about?

Which prerequisite is least stable?

Which definition or condition changes method legality?

Which example family still needs support?

If the learner cannot answer these, the map is not yet functional.

Week 2 — Stabilise direct control

Use direct and lightly varied exercises.

Target accuracy before speed.

Mark by first failure.

If one execution error repeats, isolate it.

If a definition or notation issue is causing mistakes, return to meaning rather than adding more calculation.

Use worked examples only where needed.

Fade them.

At the end of Week 2, the learner should usually be able to:

state the core relationship;

solve representative direct problems;

explain at least one strategic step;

and check a result.

If not, stay at this layer rather than pretending the chapter is ready for mixed review.

Week 3 — Vary and reverse

Change one feature at a time.

Examples:

unknown direction;

sign pattern;

diagram orientation;

context;

representation;

unit;

precision.

Use near-neighbour questions.

For percentage:

forward versus reverse.

For graphs:

direct proportion versus fixed-fee linear.

For geometry:

Pythagoras versus trigonometry.

For probability:

with versus without replacement.

Ask before solving:

What changed?

Why does the route change or remain the same?

This is the week where the chapter stops being one repeated procedure.

Week 4 — Mix and remove chapter cues

Use end-of-chapter review.

Hide section headings where possible.

Add questions from one or two neighbouring chapters.

Before each question, state:

target;

representation;

candidate method;

condition.

Mark selection errors separately from execution errors.

If the learner chooses correctly but calculates badly, the repair is different from choosing the wrong method.

Begin closed-book summary retrieval before opening chapter notes.

Week 4 evidence should answer:

Can the learner recognise the method without the textbook telling them what section they are in?

Week 5 — Delay and cumulative return

Reduce new chapter-specific work.

Bring back earlier sections after several days.

Use cumulative review.

Do not reread before attempting.

Ask the learner to reconstruct:

definition;

relationship;

condition;

one example;

one check.

Then sample direct, changed and mixed questions.

If retrieval is strong, reduce chapter-specific volume further.

If retrieval fails, reopen only the missing layer.

This week distinguishes durable learning from recent familiarity.

Week 6 — Leave the textbook

Use a different source.

Teacher worksheet.

Past assessment-style material where appropriate.

A tutor-created changed problem.

An AI-generated problem that has been checked for correctness.

Or a problem constructed by changing the textbook example’s surface.

The learner should no longer have the chapter open.

Test:

target recognition;

method selection;

execution;

checking;

interpretation;

and recovery.

Then return to the textbook only if a specific relationship needs review.

Six-week example — linear graphs

Week 1: map coordinates, gradient, intercept, y=mx+c, graph interpretation.

Week 2: direct gradient and line equations.

Week 3: negative gradients, changed scales, unknown intercept/gradient, equation↔graph translation.

Week 4: mixed graphs with direct proportion and simultaneous intersections.

Week 5: closed-book retrieval plus cumulative algebra.

Week 6: unfamiliar tariff and motion contexts outside the textbook.

Six-week example — geometry

Week 1: conditions, markings, angle facts.

Week 2: direct theorem use.

Week 3: rotated diagrams, correspondence, changed target.

Week 4: near-neighbour method choice.

Week 5: cumulative retrieval with older algebra.

Week 6: unfamiliar geometry synthesis.

When the six-week arc should shorten

If the learner already owns definitions and direct execution, begin later in the arc.

A chapter may need only variation, mixing and delayed transfer.

Do not force six weeks of direct work where evidence is already strong.

When the arc should lengthen

If prerequisites are missing, direct accuracy remains unstable or transfer repeatedly fails, extend the relevant stage.

Do not interpret a longer route as failure.

The calendar is not the goal.

Independent mathematical control is.

The arc’s one-direction principle

As weeks progress, the textbook should generally become less visible:

OPEN AND EXPLANATORY → PARTLY COVERED → REFERENCE ONLY → CLOSED DURING ATTEMPT → CLOSED DURING RETRIEVAL → ABSENT DURING TRANSFER.

If the book becomes more necessary over time, diagnose why.

Appendix H — Solution-manual discipline: twelve rules for getting help without surrendering the learning

A full solution is one of the most powerful and dangerous textbook supports. It can reveal a missing idea immediately. It can also remove the exact decision the learner needed to practise. The rules below keep solution use diagnostic.

Rule 1 — Attempt before exposure

Write something mathematically meaningful before opening the solution.

At minimum:

target;

givens;

representation;

first move or attempted route.

If nothing can be written because the concept is genuinely unknown, return to the explanation/worked example rather than using the exercise solution as first teaching.

Rule 2 — Name the exact question you want the solution to answer

Do not open a solution with the vague aim “see how to do it”.

Ask:

Why is my equation different?

Which theorem did I miss?

Where did the sign change?

Why was one root rejected?

What condition makes this method legal?

What representation made the route shorter?

This turns solution reading into targeted diagnosis.

Rule 3 — Reveal as little as necessary

If the first line is enough to restart independent work, stop reading.

If one diagram mark or variable definition resolves the gap, return to your attempt.

Do not consume the whole route merely because it is available.

Rule 4 — Compare states, not handwriting

Your solution may look different from the printed one.

Compare:

representation;

mathematical relationships;

conditions;

logical validity;

final answer.

Two routes can be different and both valid.

The goal is not to imitate the typography of the manual.

Rule 5 — Find the first divergence

Place your working beside the solution.

Move line by line until the routes first meaningfully differ.

Ask whether:

your route is still valid;

the solution is using a shorter alternative;

or your state became invalid at that point.

This prevents later symptoms from hiding the original cause.

Rule 6 — Explain the printed step in your own words

If the solution says:

(x−3)(x−4)=0

therefore x=3 or 4,

say why:

a product is zero only when at least one factor is zero.

If the solution divides by 0.8 in reverse percentage, explain why:

the known final is 80% of original.

If the solution changes probability denominator, explain the state change.

Rule 7 — Close the solution before continuing

Once the missing idea is understood, close the manual.

Restart from the last trusted state or from the beginning if needed.

Do not leave the route visible and copy line by line.

Rule 8 — Reconstruct from blank

After finishing, wait a short moment, then redo the core route without looking.

If the same gap reappears immediately, the solution produced recognition but not repair.

Return only to the missing step.

Rule 9 — Use a changed retest

Never let the repaired question be the only evidence.

Change:

numbers;

signs;

unknown direction;

context;

diagram orientation;

or representation.

The invariant should survive.

Rule 10 — Return later

Delayed retrieval distinguishes durable repair from same-session familiarity.

Before reopening the solution, attempt again.

If the route returns, support can fade.

Rule 11 — Do not punish yourself for needing a solution

A solution manual is an instructional resource.

The question is not whether it was used.

The question is what responsibility remained with the learner afterward.

Using one strategic line can be better learning than spending forty minutes repeating an unknown misconception.

Rule 12 — Track dependence, not frequency alone

A learner may use solutions often while first learning a difficult topic and still make strong progress if solution size decreases.

Another learner may open solutions only occasionally but copy entire routes whenever they do.

The better measure is:

how much help is needed;

how soon it is removed;

whether reconstruction succeeds;

and whether transfer occurs.

Solution-use receipt

After consulting a solution, record only:

What I could do before looking.

What exact step I needed.

Why that step works.

Changed retest result.

Whether I needed the solution again later.

Example:

“Could form simultaneous equations; did not see elimination after multiplying first equation. Solution showed coefficient alignment. Reconstructed independently. Changed retest succeeded. No further support needed.”

Another:

“Could calculate percentages; reverse question failed at base identification. Solution showed 0.85P=final. Changed retest still failed. Reopen percentage-base repair rather than reading more full solutions.”

The solution manual has done its job when the next similar problem requires less of it.

Appendix I — Textbook exit gate: twelve questions to decide whether the chapter can move from active study to maintenance

A textbook chapter should not remain permanently active. At some point, the learner should stop rereading explanations, stop depending on nearby worked examples, and stop treating the chapter as a special environment. The exit gate below asks whether that point has been reached.

1. Can you state the chapter’s big idea without opening the book?

Do not recite the title.

State the relationship the chapter built.

Examples:

“Direct proportion models a constant ratio y/x through y=kx.”

“Factorisation rewrites an expression into a product and can expose roots when a quadratic equals zero.”

“Similarity preserves shape through equal corresponding angles and proportional corresponding lengths.”

If the learner can perform procedures but cannot state the chapter’s central relationship, return briefly to explanation and connection rather than assigning more routine questions.

2. Can you reproduce the important definitions accurately enough to classify examples and non-examples?

A definition is not secure because it sounds familiar.

Test it.

For direct proportion, classify:

y=4x;

y=4x+1;

y=x².

For prime numbers, classify 1, 2, 9, 13.

For similar figures, reject a pair that merely looks similar but lacks sufficient evidence.

If definition use fails, keep the definition active in the chapter dashboard.

3. Can you translate the notation without relying on the textbook’s prose?

Read symbols as mathematical meaning.

Can you explain:

x≥5;

f(3);

AB∥CD;

3:5;

√18;

or the course-appropriate notation in the chapter?

Can you reverse the translation from words back to symbols?

If notation remains a visual pattern rather than a language, the chapter is not ready to leave active study.

4. Can you reconstruct one representative worked route with the example closed?

Choose a central worked example.

Do not memorise its numbers.

Reconstruct the decision sequence.

For a reverse-percentage example:

identify original/final;

convert percentage to multiplier;

write final=multiplier×original;

rearrange;

check forward.

For a geometry example:

identify target;

establish condition;

choose theorem;

calculate;

interpret.

If the route disappears when the printed example disappears, active worked-example support is still needed.

5. Can you solve a direct fresh problem accurately?

This is the basic execution gate.

The learner should be able to perform the new relationship on a clean question without help.

If direct work is unstable, do not hide the problem by moving immediately to transfer.

Repair the direct layer.

6. Can you solve a changed problem where one surface feature is different?

Change:

numbers;

signs;

unknown direction;

diagram orientation;

wording;

or context.

The relationship should remain available.

Example:

if the learner can solve ordinary percentage increase, reverse the known/unknown direction.

If the learner can solve similarity on upright triangles, rotate one.

If the learner can calculate gradient from positive coordinates, introduce negatives.

This gate tests whether learning is attached to structure rather than one printed surface.

7. Can you distinguish a near-neighbour that requires a different method?

Many textbook chapters become dangerous when the learner overgeneralises the method.

Test contrasts.

Direct proportion versus linear with a fixed intercept.

Pythagoras versus non-right triangle.

Forward versus reverse percentage.

With versus without replacement.

Exact versus approximate answer.

A learner who cannot reject the near neighbour has not yet learned the boundary of the method.

8. Can you choose the method when the chapter heading is hidden?

Use mixed questions.

The learner should identify the target, conditions and representation before naming a route.

If the method is available only because the page says “Factorisation” or “Trigonometry”, chapter context is still doing cognitive work.

Do not demand perfect mixed performance immediately.

Look for correct route selection on representative problems.

9. Can you mark and repair an error without needing the full solution?

Give the learner a wrong attempt or allow an ordinary error to occur.

Can they:

find the last trusted state;

locate the first wrong or missing decision;

repair locally;

and continue?

If every disagreement still requires opening the full solution, textbook dependence remains high.

10. Can you retrieve the chapter after a delay before rereading?

Wait.

Then ask for:

definition;

core relationship;

one direct example;

one changed example;

one check.

Attempt before opening the book.

If most of the system returns, the chapter is becoming durable.

If the relationship disappears, reopen only the missing layer and shorten the next retrieval interval.

11. Can you use the Mathematics on a problem that does not come from this textbook?

This is the critical transfer gate.

Use:

a teacher worksheet;

a tutor-created problem;

a past paper or assessment-style question where appropriate;

another textbook;

or a carefully checked generated problem.

The surface should differ enough that page memory cannot carry the route.

If performance collapses outside the book, return to changed-representation and mixed selection work rather than rereading the entire chapter.

12. Can the textbook now become reference instead of constant support?

Ask what the learner still needs the chapter for.

Occasional definition lookup?

A rare edge-case example?

Additional exercise volume?

A cumulative review source?

Those are normal maintenance roles.

If the learner still needs the chapter open for every routine problem, active study is not finished.

Exit classifications

Keep active:

definitions, direct execution or worked-route reconstruction are still unstable.

Transfer watch:

direct work is secure, but changed/mixed problems still fail.

Maintenance:

changed problems and delayed retrieval are secure; sample occasionally.

Reference only:

the learner can use the Mathematics independently and opens the chapter only for uncommon details or confirmation.

Example exit gate — linear equations

Big idea:

equivalent operations preserve equality while isolating an unknown.

Definition/notation:

equation, solution, equality.

Worked route:

reconstruct variable-on-both-sides example.

Direct:

3x+5=20.

Changed:

5−2(x−3)=11.

Near neighbour:

inequality where negative division changes order.

Mixed:

equation hidden inside a word problem.

Repair:

substitute solution back into original.

Delayed:

use equation rearrangement several days later.

Outside textbook:

teacher modelling question.

If these are secure, the linear-equation chapter can move to maintenance.

Example exit gate — probability

Big idea:

probability models events within a sample state.

Definitions:

event, sample space, independent/conditional language as appropriate.

Worked route:

without-replacement tree.

Direct:

one event probability.

Changed:

different counts.

Near neighbour:

with versus without replacement.

Mixed:

at least one using complement.

Repair:

check result lies in [0,1] and branch state is coherent.

Delayed:

return later.

Outside textbook:

two-way table conditional problem.

If these are secure, the chapter has travelled beyond its original representation.

Example exit gate — quadratic graphs

Big idea:

different algebraic forms expose different graph features.

Definitions:

root/intercept, vertex where course-appropriate.

Worked route:

factor form for roots, completed-square form for vertex.

Direct:

solve simple factorable quadratic.

Changed:

form equation from roots or graph.

Near neighbour:

choose factorisation versus completing square based on target.

Mixed:

quadratic among other function types.

Repair:

expand/factor to verify equivalence.

Delayed:

return a week later.

Outside textbook:

modelling problem.

Once secure, the chapter becomes reference and maintenance material.

The final principle

A learner does not “finish” a Mathematics textbook by turning the last page.

They finish the active job of a chapter when the book can close without the relationship disappearing.

Textbook independence means the printed page has transferred meaning, route, selection, feedback and memory responsibilities back to the learner.

Appendix J — Textbook maintenance receipts: short records that decide what happens next

A maintenance receipt is not a study diary. It is a compact record of evidence that changes the next textbook decision. Use it only when a chapter, method or recurring error needs tracking.

Receipt 1 — Definition secure, execution weak

Chapter: direct proportion.

Evidence: learner correctly distinguishes y=3x from y=3x+5 and explains why only the first is direct proportion.

Failure: arithmetic errors when finding k from data.

Next textbook action: skip more definition reading; choose short direct k-calculation exercises.

Exit condition: arithmetic becomes accurate enough that definition knowledge can be used fluently.

Receipt 2 — Execution secure, selection weak

Chapter: quadratic methods.

Evidence: factorisation and completing square both accurate when labelled.

Failure: learner cannot decide which form is useful from the target.

Next textbook action: use mixed review and method-comparison questions, not more blocked technique practice.

Exit condition: learner can justify route choice before calculation.

Receipt 3 — Worked-example dependence

Chapter: trigonometry.

Evidence: correct solutions with textbook example open.

Failure: route disappears when example closes.

Next textbook action: cover-and-predict, then one paired problem with only a small hint if needed.

Exit condition: learner selects ratio and rearranges without visible model.

Receipt 4 — Answer-key dependence

Chapter: algebra.

Evidence: learner checks every line and pauses until confirmed.

Failure: self-monitoring remains external.

Next textbook action: complete three questions before checking; mark uncertainty points.

Exit condition: learner can continue through ordinary uncertainty and check at the end.

Receipt 5 — Solution manual used well

Chapter: similarity.

Evidence: learner could establish similarity but mismatched corresponding sides.

Solution use: read only correspondence step, closed solution, reconstructed ratio.

Changed retest: rotated diagram correct.

Next textbook action: no more full solution support; move to mixed geometry.

Receipt 6 — Same-day success, delayed failure

Chapter: reverse percentage.

Evidence: five same-session questions correct.

Delayed evidence: two days later learner multiplies final by 0.8 again.

Next textbook action: reopen relationship briefly, then schedule closer retrieval return.

Exit condition: correct base/direction after delay and in mixed work.

Receipt 7 — Textbook surface dependence

Chapter: graphs.

Evidence: textbook graphs accurate.

Failure: unfamiliar teacher graph with different scale is misread.

Next textbook action: reduce book-specific practice; use changed axis scales and external graph sources.

Exit condition: axes/units/scale scan survives unfamiliar formatting.

Receipt 8 — Chapter ready for maintenance

Chapter: linear equations.

Evidence: direct, sign-sensitive, word-to-equation and mixed problems correct; substitution check used independently; one-week retrieval secure.

Next textbook action: stop dedicated chapter study.

Maintenance: occasional equation-rich mixed questions.

Reopen only if: recurring mechanism returns across several later tasks.

Receipt 9 — Chapter ready for reference-only status

Chapter: basic statistics.

Evidence: learner selects mean/median appropriately, handles frequency tables, explains estimated grouped mean, compares distributions and returns successfully after delay.

Next textbook action: use chapter only for uncommon conventions, notation or extra practice if needed.

Meaning: active teaching role is complete.

Receipt 10 — Another resource is genuinely needed

Chapter: probability.

Evidence: definitions and textbook exercises secure, but the book contains almost no mixed or transfer problems.

Missing job: method selection under unfamiliar wording.

Next action: add one supplementary mixed-problem source.

Do not: buy another full explanatory textbook covering the same direct material.

Receipt 11 — Textbook explanation genuinely insufficient

Chapter: completing the square.

Evidence: learner understands surrounding algebra but cannot explain why +9−9 is introduced in x²−6x+5 despite careful rereading.

Next action: use teacher/tutor/alternative representation for the missing conceptual bridge.

Return: reconnect the new explanation to the textbook example and complete paired problems.

Receipt 12 — Textbook lane complete

Evidence: chapter map can be reconstructed; definitions and notation are secure; worked routes can be generated independently; mixed method selection is accurate; delayed retrieval succeeds; external transfer succeeds.

Next action: close the textbook as an active study environment.

Future role: reference, maintenance or occasional additional practice.

The final receipt rule

A receipt should end with a decision.

Read again.

Attempt more direct questions.

Move to variation.

Move to mixed work.

Repair one prerequisite.

Use a small hint.

Return after delay.

Move to maintenance.

Leave the textbook.

If the receipt does not change what happens next, it is paperwork rather than learning evidence.

The textbook becomes powerful when every page produces a better next decision—and temporary when the learner can finally make those decisions without it.

Appendix K — A 15-minute textbook router for short study sessions

Sometimes the learner has only a short block of time. The answer is not to abandon the textbook workflow. Compress it around one clear learning job.

Minute 1–2 — Orient

State the chapter and the exact task.

Examples:

“I am repairing reverse percentage direction.”

“I am learning how gradient connects two coordinates.”

“I am checking whether I can factorise simple quadratics without a model.”

A short session needs one target.

Do not open three chapters.

Minute 3–5 — Retrieve before reading

Close the book first.

Write or say what you already remember:

definition;

relationship;

condition;

one example.

Then open the relevant page and compare.

Use the difference to decide what the next ten minutes should do.

Minute 6–8 — Use one explanation or worked example actively

If a relationship is missing, read the explanation.

If a route is missing, study one worked example.

Cover the next line and predict.

Ask why the key step is valid.

Do not read several examples simply because they are adjacent.

Minute 9–12 — Attempt one or two selected exercises

Choose a question that directly tests the target.

Then, if the first is secure, choose one changed version.

Keep the example closed.

If stuck, identify the exact decision missing before checking help.

Minute 13–14 — Mark and repair

Check the answer.

If wrong, locate the first wrong or missing decision.

Write a one-line repair:

“wrong 100% base”;

“misread graph scale”;

“negative distribution”;

“condition for theorem missing”;

“answer required exact form”.

If correct, ask whether the method was independent or supported.

Minute 15 — Close the book

State what you should still be able to do tomorrow.

Examples:

“Given a final amount after a percentage change, I can write final = multiplier×original before deciding the arithmetic.”

“I will read axis values before counting graph squares.”

“I will match similar-triangle sides from vertex correspondence, not visual position.”

Schedule a delayed return if the skill matters.

When a 15-minute block is not enough

A short session is not suitable when:

the concept is entirely new and requires substantial explanation;

several prerequisites are unstable;

the learner is repeatedly unable to interpret the notation;

or a long multi-step task needs sustained reasoning.

In those cases, use the router only to identify what the longer session should address.

Why the router works

Even a short textbook session can contain the essential learning cycle:

ORIENT → RETRIEVE → READ WITH PURPOSE → ATTEMPT → MARK → REPAIR → CLOSE → RETURN LATER.

The learner still does Mathematics rather than merely consuming pages.

That is more useful than spending fifteen minutes rereading a familiar chapter without testing whether anything can be generated independently.

Short study is productive when the learning job is narrow, the attempt is genuine and the session ends with a clear next state.

Appendix L — Final release check: what “using the textbook properly” should look like in ordinary study

Before this chapter system is considered complete, the learner should be able to describe their own textbook workflow without needing a checklist beside them.

They should know where to begin:

map the chapter, identify prerequisites and read definitions that control meaning.

They should know how to use examples:

predict before reading the next strategic step, explain why the route is valid, then close the example and attempt a paired problem.

They should know how to choose exercises:

enough direct work to stabilise the relationship, then variation, near neighbours, mixed selection and transfer.

They should know how to use answers:

after a real attempt, as evidence about the result rather than continuous reassurance.

They should know how to use full solutions:

to repair a specific missing decision, then close the solution and reconstruct.

They should know what notes are for:

compressing relationships, conditions, contrasts, personal error patterns and retrieval prompts—not reproducing the whole book.

They should know when to reread:

when retrieval reveals a genuine gap, not simply because rereading feels easier than attempting.

They should know when to leave the book:

when definitions, routes, method selection, checking and delayed transfer survive on problems that do not look like the chapter.

Most importantly, the learner should be able to tell the difference between:

“the textbook explained it”

and

“I can now do the mathematical work the textbook used to do for me.”

That difference is the entire purpose of the system.

The textbook is not the destination. It is a temporary structure for transferring mathematical meaning and control to the learner.

Appendix M — One last test

Close the textbook.

Without looking, state the chapter’s central relationship, one condition that controls when it applies, one example, one near non-example, one checking method and one question you could now solve that would have been difficult before studying.

If those six things can be produced accurately, the book has transferred more than information.

It has transferred a usable mathematical system.

If one item is missing, reopen only that part of the chapter, repair it, close the book again and retest.

The final study habit is selective return, not permanent dependence.