BTT Mathematics / Primary Mathematics Learning Hub / Working Mathematically

Primary Mathematics: Choosing Strategies, Tools and Efficient Methods | Worked Learning Guide

The best mathematical tool is not the most advanced one. It is the one that makes the important relationship visible, keeps error risk under control and earns its cost in time and effort.

Primary Mathematics offers many possible tools: mental calculation, written algorithms, number lines, arrays, bar models, tables, grids, rulers, protractors, manipulatives, graph paper, estimation, calculators or digital tools where the learning context permits them. A learner can know how to use several of these and still struggle with a more important question: which one should I choose now?

Strategy selection is not a race toward shorter working. A method can be short but fragile. Another can be slower but transparent. A third may be ideal for checking but poor for discovering the answer. The goal is to match method to mathematical job.

This guide develops that matching process. It asks what the problem needs, what information the tool reveals, what it hides, whether the learner can switch methods when the first choice becomes inefficient, and how to use a second method as an independent check.

Identify the job · Mental methods · Written methods · Models and diagrams · Measurement tools · Technology and limits · 24 questions · Worked solutions

1. Start with the mathematical job, not the favourite method

For 398+247, a learner might use:

• mental compensation: 400+245=645,
• written column addition,
• an open number line,
• place-value blocks.

All can be mathematically valid. If the learner sees 398 as two less than 400, mental compensation may be fastest. If regrouping itself is the learning target, place-value or written work may be more informative. If the answer looks suspicious, a different representation can serve as a check.

Method quality depends on the job.

2. Separate solving, explaining and checking

The same tool does not need to perform all three jobs.

For 49×18, mental calculation can solve efficiently: 50×18−18=900−18=882.

An area model can explain why the distributive step is valid.

An estimate 50×20≈1000 can check that 882 has a plausible magnitude.

A learner who insists on one tool for everything may work harder than necessary.

3. Choose mental methods when the structure is friendly

Mental calculation is especially useful when numbers are close to benchmarks, contain convenient factors, or can be decomposed simply.

Worked example: compensation

199+386 = 200+385 = 585.

The total is preserved by moving one unit from the second addend to the first.

Worked example: halving and doubling

25×48 = 50×24 = 100×12 = 1200.

Doubling one factor while halving the other preserves the product.

When mental calculation is a poor choice

For 4873−2968, some learners can calculate mentally, but a written method may reduce working-memory load and make regrouping easier to inspect. Efficiency includes reliability.

4. Choose written algorithms when place-value bookkeeping matters

Written methods externalise steps. They are valuable for long calculations, multiple regroupings, decimal alignment and work that must be checked later.

Worked example: decimal subtraction

8.05−3.786. Rename 8.05 as8.050 and align place values. The exact result is4.264.

A written algorithm protects the thousandths column. Mental estimation—about8−3.8≈4.2—checks the scale.

Do not use the algorithm before the model is understood

If a learner subtracts denominators in 3/4−1/8 because “subtraction means subtract top and bottom”, more written practice will not repair the conceptual error. First create common fractional units.

5. Choose a diagram when the relationship is harder than the arithmetic

Question: A is 18 more than B and together they total74.

A bar model makes the structure visible: two equal B-sized parts plus18 equal74. Remove18, split56 into two, giving B=28 and A=46.

The arithmetic is simple. The main difficulty is representing the relationship correctly.

Choose an array for equal groups

For 6×14, an array can be decomposed into6×10 and6×4, giving60+24=84.

Choose a number line for order or distance

To compare0.58 and0.6, place both between0 and1 or rename0.6 as0.60. The number line emphasises relative position.

Choose a table for repeated pairs

A taxi-style classroom model has cost $4 plus $2 per kilometre. A table reveals the fixed starting fee and the +$2 change per kilometre. It also shows why direct proportion fails: zero kilometres still costs $4.

6. Choose an organised list or tree when completeness matters

Using digits1,2,3,4 without repetition, find all two-digit numbers greater than30. An organised list by tens digit is better than random trial.

For a sequential two-stage choice where order matters, a tree can preserve chronology. For unordered selections, a triangular list may prevent duplicates.

This connects to Working Systematically.

7. Choose algebra when repeated structure needs compression

A row of n joined squares uses3n+1 matchsticks. For Figure100, drawing all squares is inefficient. The direct rule gives301 immediately.

But if the learner cannot explain where3n+1 comes from, the algebra is premature. A visual construction should establish the rule first.

8. Choose estimation before exact work when scale matters

Estimate 498×21 as about500×20=10,000. An exact answer of1,045 or104,580 should immediately be questioned.

Estimation is a strategic tool because it can reject impossible magnitudes before detailed rechecking.

Worked example: money

Six items cost$19.85 each. Estimate about6×20=$120. Exact total=$119.10. The estimate confirms the order of magnitude.

9. Choose inverse operations for checking

For 803−476=327, check476+327=803.

For 157÷12=13 remainder1, check12×13+1=157 and confirm remainder1 is less than12.

Inverse checking is stronger than repeating the same subtraction or division because it reorganises the relationship.

10. Choose a ruler only when length is the quantity to measure

A ruler measures linear distance. It does not directly measure area, angle or volume.

When measuring length, align the zero mark with the start of the object. If the zero edge is damaged, begin at another known mark and subtract the two readings.

Worked example: broken-zero ruler

An object begins at2.3 cm and ends at8.9 cm. Length=8.9−2.3=6.6 cm.

The tool provides two positions; subtraction provides the distance.

11. Choose a protractor only after identifying the angle correctly

A protractor is useful when an angle is not already determined by geometric relationships.

Before measuring, identify the vertex and two rays. Centre the protractor correctly, align one baseline, then read the scale that begins at zero on that ray.

If a diagram states that an angle is a right angle, measuring a slightly imperfect drawing as89° should not override the stated mathematical condition.

12. Choose grid paper when position and scale need control

Grid paper is useful for coordinates, area, transformations, scale drawings and systematic geometry counts. It reduces ambiguity about equal lengths and perpendicular directions.

Worked example: translation

Move point(2,3) four units right and two up. New point=(6,5). A grid makes both coordinate changes visible.

13. Choose manipulatives when the unit structure is not yet secure

Base-ten blocks can show one ten regrouping into ten ones. Fraction strips can show why1/2=2/4. Counters can model equal groups and remainders.

The object is not the mathematics. The learner must eventually explain which property of the object represents the mathematical relationship.

Move from concrete to drawn to symbolic

For 3/4+1/8, fraction strips can show that each quarter equals two eighths. A drawn bar can then represent6/8+1/8. Symbolically, the result is7/8.

14. Choose a graph when change across a variable matters

A graph is useful for trends, comparison and rates of change across ordered inputs. A table may be better when exact individual values matter more.

If a learner needs the precise value at time7, read the graph carefully. If the learner needs the overall relationship, the graph shape may communicate more than a long value list.

15. Use calculators or digital tools strategically when permitted

Technology can reduce routine arithmetic, generate many cases or visualise changing parameters. It can also hide input errors and produce precise-looking nonsense.

Where a calculator is permitted in a learning context, estimation should usually precede entry. If 48.7×19.8 is expected near1,000, a displayed answer96.426 signals a likely input or decimal mistake.

Do not infer examination permission from this article. Tool rules depend on the specific assessment and school instructions.

Worked example: spreadsheet-style table

To investigate perimeter of whole-number rectangles with area120, a digital table can list factor pairs and calculate perimeters quickly. The learner still needs to define the permitted factor pairs and explain why the search is complete.

Automation can accelerate a valid model; it cannot decide whether the model itself is valid.

16. A tool’s limitation is part of knowing how to use it

Tool	Useful for	Common limitation
mental calculation	friendly structure, fluency	working-memory overload
written algorithm	place-value bookkeeping	can hide conceptual misunderstanding
bar model	part-whole, comparison, ratio	wrong unit lengths create false relationships
number line	order, distance, fractions	poor scale can mislead
table	repeated relationships	does not prove causation or rule by itself
graph	trend and comparative shape	axis scale can distort visual impression
calculator	routine computation	accepts wrong input obediently

17. Compare methods on more than speed

For 25×36:

• Written multiplication is reliable.
• Quarter-of-100 reasoning: 25×36=(100÷4)×36=3600÷4=900.
• Half-and-double: 50×18=100×9=900.

The “best” method depends on fluency, explanation goal and error risk. A learner should be able to compare why one route is efficient for these numbers.

18. Switch strategies when the first route becomes expensive

Perseverance is not repeating an ineffective method indefinitely.

If random listing produces duplicates, switch to an ordered table. If a bar model becomes crowded in a many-stage process, switch to a stage table or equation. If exact arithmetic obscures a likely impossibility, estimate or use a bound first.

Strategic flexibility includes abandoning a legal but inefficient route.

19. Use a second tool as an independent check

Calculate 24×13 by written multiplication to get312. Check with area decomposition:24×10+24×3=240+72=312.

The second method is valuable because it is structurally different. Repeating the same written algorithm may reproduce the same transcription error.

20. Strategy choice under time pressure

Under a reasonable timed practice condition, use a decision hierarchy:

1. Is there an obvious mental shortcut?
2. If not, is a standard written method direct?
3. Is the main difficulty actually representation?
4. Would an estimate or bound simplify the search?
5. Would a table/list prevent omission?
6. Can a different method verify the result quickly?

Speed should emerge from method fit, not from skipping interpretation.

21. Common strategy-selection errors

• Using the same favourite method for every problem.
• Choosing a diagram after the arithmetic is already easier than the drawing.
• Using mental calculation when working memory is overloaded.
• Using a calculator without estimating the expected scale.
• Using an exact method when a bound already proves impossibility.
• Using random trial where an ordered search would guarantee completeness.
• Keeping a representation after it has become cluttered.
• Switching methods so often that no method is completed or checked.

22. A six-question strategy selector

1. What mathematical relationship is hardest here?
2. Which method makes that relationship visible?
3. What information will the method hide?
4. What is the likely error if I use it?
5. Is there a cheaper method for the same job?
6. Which different method can check the result?

23. Practice: 24 original questions

Questions 1–8: Choose a method

1. Calculate199+386 and explain why compensation is efficient.
2. Calculate25×48 using halving and doubling.
3. Calculate4873−2968. Choose mental or written work and justify your choice.
4. Solve A is18 more than B, total74. Choose a useful representation.
5. Compare0.58 and0.6. Choose a useful representation or renaming method.
6. List all two-digit numbers greater than30 from digits1,2,3,4 without repetition. Choose an organising tool.
7. Figure n uses3n+1 matchsticks. Find Figure100 and explain why a direct rule is preferable to drawing.
8. Estimate498×21 before calculating exactly.

Questions 9–16: Tools and checks

9. An object begins at2.3 cm and ends at8.9 cm on a ruler. Find its length.
10. A right angle is drawn imperfectly and appears89°. The diagram states it is a right angle. Which information controls the mathematical answer?
11. Translate point(2,3) four right and two up.
12. Use fraction strips or an equivalent diagram to explain3/4+1/8.
13. A table shows taxi cost $4+$2 per km. Explain why the table is more informative than assuming direct proportion.
14. A graph rises5 vertical squares, each representing20 units. Find the numerical change.
15. Calculate24×13 one way and verify it with a structurally different method.
16. For157÷12=13 remainder1, state the strongest inverse check.

Questions 17–24: Strategy evaluation

17. Why might mental calculation be a poor choice for a long decimal subtraction even if it is possible?
18. Why might a bar model be a poor choice for an exhaustive counting problem?
19. A calculator returns96.426 for48.7×19.8. Use estimation to evaluate the result.
20. Give one case where a bound can answer a question before exact calculation.
21. A learner uses random trial to find all factor pairs of72. Propose a better method.
22. Compare written multiplication and25×36=100×36÷4. Which is more efficient here, and why?
23. Give one example where switching strategies is sensible after a first attempt.
24. Create a Primary Mathematics problem and name a solving tool, an explaining tool and a checking tool that need not be the same.

24. Worked solutions

1. 585. Move1 from386 to199:200+385=585. 2. 1200. Double25 to50 and halve48 to24; repeat to100×12. 3. Written subtraction is reasonable because several place-value columns and regroupings must be tracked; exact answer1905. 4. Bar model: two equal B parts plus18=74; B=28,A=46.

5. Rename0.6 as0.60 or use a number line;0.60>0.58. 6. Organise by tens digit3 then4:31,32,34,41,42,43. 7. 301; the direct rule avoids constructing99 previous figures. 8. Estimate about10,000; exact498×21=10,458.

9. 6.6 cm. 10. The stated right-angle condition controls:90°. A sketch not drawn exactly to scale does not override it. 11. (6,5). 12. Convert3/4 to6/8, then6/8+1/8=7/8.

13. The fixed$4 fee is visible at zero distance, showing total cost is not directly proportional to distance. 14. 100 units. 15. Example exact calculation312; area decomposition240+72=312. 16. Check12×13+1=157 and remainder1<12.

17. Mental work may overload working memory and make place-value errors harder to inspect. 18. A bar model is designed for quantitative relationships, not necessarily to guarantee no cases are omitted; an organised list/table may fit better. 19. Expected product is near50×20=1000, so96.426 is implausibly small. 20. Example: four 6 m gaps cover at most24 m, so a25 m path with gaps at most6 m needs at least5 gaps.

21. List factor pairs in increasing smaller-factor order:1×72,2×36,3×24,4×18,6×12,8×9, then stop before reversal. 22. The quarter-of-100 route is especially efficient:25×36=3600÷4=900. Written multiplication remains valid. 23. Example: abandon random rectangle counting and switch to boundary-line pairs when omissions appear. 24. Answers vary; each tool choice should be justified by its job.

25. Strategy laboratory: one problem, four legal routes

Problem: Find 3/8 of 240.

Unit fraction: 240÷8=30;3×30=90.

Percentage connection: 3/8=37.5%; 37.5% of240=90, but this is less efficient unless that conversion is already fluent.

Bar model: divide a 240-unit bar into8 equal parts of30 and select3.

Factor simplification: 240×3÷8=(240÷8)×3=90.

All routes are valid. The unit-fraction and factor-simplification routes are shortest for these numbers. The bar model is strongest if the learner does not yet understand why division by8 occurs. Strategy choice depends on the learning problem, not only the final arithmetic.

26. Parent and tutor guide

When a learner uses a slow method successfully, do not replace it immediately. First ask the learner to explain why it works. Then compare a second method on time, transparency and error risk.

If a learner chooses an advanced-looking tool poorly, return to the job. A calculator cannot decide which number is the percentage base. A bar model cannot repair an incorrect unit conversion by itself. A protractor cannot prove an angle property that follows from stated geometry.

Build a small repertoire, then build selection. Mastery is not owning twenty strategies. It is recognising a few powerful ones and knowing when each earns its place.

Delayed return

Three days later, present four unlabeled problems: one friendly arithmetic calculation, one relationship-heavy word problem, one exhaustive search and one measurement task. Ask the learner to choose a tool before solving and name one alternative tool for checking.

27. Mastery receipt

• I choose a method from the mathematical job rather than habit.
• I distinguish solving, explaining and checking tools.
• I use mental methods when number structure is friendly.
• I use written methods when bookkeeping matters.
• I choose diagrams and tables when relationships need to be made visible.
• I know the limitations of measurement and digital tools.
• I switch strategies when the first route becomes inefficient.
• I can justify why one method is more suitable than another for a particular problem.

Sources and scope

Illustrative Mathematics — Use Appropriate Tools Strategically describes mathematically proficient learners as considering available tools, their usefulness and their limitations rather than using tools automatically.

NSW Department of Education — Mathematics K–6 Effective Teaching Approaches describes Working Mathematically through interconnected communicating, understanding/fluency, reasoning and problem-solving processes.

For Singapore subject scope and permitted methods in a particular school context, use the MOE Primary Mathematics Syllabus P1–P6, updated October 2025 and current school instructions. This article does not override assessment-specific calculator or tool rules.

Continue the capability batch

• Working Systematically, Organised Search and Recording
• Mathematical Communication, Precision and Multiple Forms
• Mathematical Connections, Transfer and the Same Structure
• BTT Primary Mathematics Learning Hub

The Quiet Return

A powerful mathematician does not merely know methods. The mathematician sees the problem, chooses the method that earns its place, checks what the method cannot see, and moves on when a better route becomes available.