G2 Mathematics K210 modelling becomes fragile when learners quietly add facts that the question never gave. A rate is assumed constant, a diagram is treated as drawn to scale, a percentage base is guessed, a whole-number condition is ignored, or a graph trend is extended far beyond the measured range. The calculation can then be internally flawless while the model itself answers a different problem.
This sixty-fourth Learner’s Guide develops an assumption audit: separate what the question states, what Mathematics allows you to derive, what the context reasonably requires, and what you have merely invented. The method is especially useful for K210 problem solving and real-world contexts, where representation and interpretation matter as much as arithmetic.
The current 2027 K210 syllabus assesses standard techniques, problem solving in varied contexts, translation between representations, formulation of problems in mathematical terms and interpretation of results in context. Paper 2 Section A ends with a real-world scenario question, and real-world contexts may integrate ideas from multiple topics. Those are official syllabus features; the assumption-audit method below is an eduKateSengkang training framework.
The four assumption states
- Stated: explicitly given by the question, diagram, table or instruction.
- Derived: follows mathematically from stated information.
- Context-required: not written as a formula, but necessary because of the real-world meaning, such as a whole number of buses.
- Invented: added by the learner without support, such as assuming a rate stays constant when the question never says or implies it.
The audit is not about avoiding all assumptions. Mathematical modelling often requires interpreting the context. The skill is knowing which assumptions are licensed and which ones silently change the problem.
1. Diagram looks proportional
For this pattern, begin from stated dimensions and angle/right-angle markings. A legitimate next step is relationships that follow from geometry. The common modelling error is treating every drawn length or angle as visually accurate. Because the arithmetic after that assumption may still be perfect, the mistake can hide until the final interpretation.
Run the audit this way: use only labelled or mathematically derivable information unless the question explicitly authorises measurement. In training, compare answers from deliberately distorted diagrams to expose visual assumption. The question is not “Did I assume anything?” but “Which assumptions are licensed by the stated mathematics and context, and which ones changed the problem without permission?”
2. Straight-looking line
For this pattern, begin from axis values and plotted relationship. A legitimate next step is gradient or equation if supported by data/model. The common modelling error is assuming an apparently straight segment is exactly linear without enough evidence. Because the arithmetic after that assumption may still be perfect, the mistake can hide until the final interpretation.
Run the audit this way: check whether the question or graph data justify a linear model. In training, use graphs that look nearly linear but have one changing gradient. The question is not “Did I assume anything?” but “Which assumptions are licensed by the stated mathematics and context, and which ones changed the problem without permission?”
3. Constant speed
For this pattern, begin from distance and time information. A legitimate next step is average speed from total distance/total time. The common modelling error is assuming instantaneous or segment speeds are constant. Because the arithmetic after that assumption may still be perfect, the mistake can hide until the final interpretation.
Run the audit this way: distinguish average rate from constant rate. In training, compare same average speed produced by different journeys. The question is not “Did I assume anything?” but “Which assumptions are licensed by the stated mathematics and context, and which ones changed the problem without permission?”
4. Constant rate
For this pattern, begin from given rate over a stated condition. A legitimate next step is amount over a matching interval when constancy is justified. The common modelling error is extending the rate to conditions where it may change. Because the arithmetic after that assumption may still be perfect, the mistake can hide until the final interpretation.
Run the audit this way: look for wording, formula or context that licences constant-rate use. In training, practise rates that change after thresholds or time bands. The question is not “Did I assume anything?” but “Which assumptions are licensed by the stated mathematics and context, and which ones changed the problem without permission?”
5. Percentage base
For this pattern, begin from original, final or reference quantity named in the problem. A legitimate next step is percentage change relative to the correct base. The common modelling error is using the final value because it is visually nearest. Because the arithmetic after that assumption may still be perfect, the mistake can hide until the final interpretation.
Run the audit this way: write the base noun before dividing. In training, audit wrong percentage answers by denominator choice. The question is not “Did I assume anything?” but “Which assumptions are licensed by the stated mathematics and context, and which ones changed the problem without permission?”
6. Reverse percentage
For this pattern, begin from sale value and remaining percentage relationship. A legitimate next step is original value from the multiplicative relationship. The common modelling error is assuming the percentage was taken from the sale value. Because the arithmetic after that assumption may still be perfect, the mistake can hide until the final interpretation.
Run the audit this way: define original as the base before calculation. In training, use paired forward/reverse questions to reveal base direction. The question is not “Did I assume anything?” but “Which assumptions are licensed by the stated mathematics and context, and which ones changed the problem without permission?”
7. Repeated percentage changes
For this pattern, begin from successive increases/decreases and their order. A legitimate next step is multiplicative factors applied to changing bases. The common modelling error is adding percentages because the signs look compatible. Because the arithmetic after that assumption may still be perfect, the mistake can hide until the final interpretation.
Run the audit this way: track the new base after each change. In training, compare additive shortcut with multiplicative model. The question is not “Did I assume anything?” but “Which assumptions are licensed by the stated mathematics and context, and which ones changed the problem without permission?”
8. Ratio part-to-part
For this pattern, begin from named quantities and order. A legitimate next step is equivalent ratios and scaled quantities. The common modelling error is treating one part as the whole. Because the arithmetic after that assumption may still be perfect, the mistake can hide until the final interpretation.
Run the audit this way: label each ratio term before simplification. In training, use ratios where part-to-part and part-to-whole give plausible but different values. The question is not “Did I assume anything?” but “Which assumptions are licensed by the stated mathematics and context, and which ones changed the problem without permission?”
9. Ratio order
For this pattern, begin from A:B stated in the text. A legitimate next step is corresponding scaled values. The common modelling error is reversing the ratio because the numbers are symmetric or familiar. Because the arithmetic after that assumption may still be perfect, the mistake can hide until the final interpretation.
Run the audit this way: write the quantity name above each term. In training, audit ratio errors for ownership rather than arithmetic. The question is not “Did I assume anything?” but “Which assumptions are licensed by the stated mathematics and context, and which ones changed the problem without permission?”
10. Map scale
For this pattern, begin from drawing-to-actual relationship and units. A legitimate next step is actual distance/area from the scale. The common modelling error is applying a length scale directly to area without squaring. Because the arithmetic after that assumption may still be perfect, the mistake can hide until the final interpretation.
Run the audit this way: identify dimensional type before scaling. In training, use matched distance and area scale questions. The question is not “Did I assume anything?” but “Which assumptions are licensed by the stated mathematics and context, and which ones changed the problem without permission?”
11. Similarity
For this pattern, begin from given or proved similarity. A legitimate next step is corresponding side factors and derived area/volume relationships. The common modelling error is assuming shapes are similar because they look alike. Because the arithmetic after that assumption may still be perfect, the mistake can hide until the final interpretation.
Run the audit this way: prove or use stated similarity before applying factors. In training, include non-similar distractor diagrams in practice. The question is not “Did I assume anything?” but “Which assumptions are licensed by the stated mathematics and context, and which ones changed the problem without permission?”
12. Right angle
For this pattern, begin from explicit right-angle mark or mathematically established perpendicularity. A legitimate next step is Pythagoras/trigonometric relations. The common modelling error is assuming an angle is 90° because it appears square. Because the arithmetic after that assumption may still be perfect, the mistake can hide until the final interpretation.
Run the audit this way: require a stated or derived right angle before right-triangle methods. In training, distort practice diagrams visually. The question is not “Did I assume anything?” but “Which assumptions are licensed by the stated mathematics and context, and which ones changed the problem without permission?”
13. Isosceles triangle
For this pattern, begin from given equal sides/angles or derived equality. A legitimate next step is related equal base angles or side conclusions. The common modelling error is assuming symmetry from appearance. Because the arithmetic after that assumption may still be perfect, the mistake can hide until the final interpretation.
Run the audit this way: identify the equality evidence first. In training, use near-isosceles drawings with no equality marks. The question is not “Did I assume anything?” but “Which assumptions are licensed by the stated mathematics and context, and which ones changed the problem without permission?”
14. Parallel lines
For this pattern, begin from parallel marking or stated condition. A legitimate next step is angle relationships from parallelism. The common modelling error is using alternate/corresponding angle rules on merely similar-looking lines. Because the arithmetic after that assumption may still be perfect, the mistake can hide until the final interpretation.
Run the audit this way: verify parallel status before angle transfer. In training, remove parallel arrows from familiar diagrams to test discipline. The question is not “Did I assume anything?” but “Which assumptions are licensed by the stated mathematics and context, and which ones changed the problem without permission?”
15. Circle centre
For this pattern, begin from explicit centre label. A legitimate next step is radii equality and central-angle relationships. The common modelling error is assuming a visually central point is the centre. Because the arithmetic after that assumption may still be perfect, the mistake can hide until the final interpretation.
Run the audit this way: use labels and circle properties, not appearance. In training, move points off-centre in training figures. The question is not “Did I assume anything?” but “Which assumptions are licensed by the stated mathematics and context, and which ones changed the problem without permission?”
16. Radius versus diameter
For this pattern, begin from given measurement and label. A legitimate next step is one from the other through factor two. The common modelling error is using a diameter as radius because the number is attached to the circle. Because the arithmetic after that assumption may still be perfect, the mistake can hide until the final interpretation.
Run the audit this way: name the quantity before formula substitution. In training, build area/circumference contrast pairs. The question is not “Did I assume anything?” but “Which assumptions are licensed by the stated mathematics and context, and which ones changed the problem without permission?”
17. Perimeter versus area
For this pattern, begin from question target and dimensions. A legitimate next step is appropriate one- or two-dimensional result. The common modelling error is assuming the familiar rectangle formula must be area. Because the arithmetic after that assumption may still be perfect, the mistake can hide until the final interpretation.
Run the audit this way: name around versus surface before choosing formula. In training, use same dimensions with different targets. The question is not “Did I assume anything?” but “Which assumptions are licensed by the stated mathematics and context, and which ones changed the problem without permission?”
18. Area versus volume
For this pattern, begin from question target and unit dimension. A legitimate next step is square/cubic units and scale factors. The common modelling error is using area reasoning in a three-dimensional problem. Because the arithmetic after that assumption may still be perfect, the mistake can hide until the final interpretation.
Run the audit this way: use units as a type check before calculation. In training, ask learners to predict final unit before formula. The question is not “Did I assume anything?” but “Which assumptions are licensed by the stated mathematics and context, and which ones changed the problem without permission?”
19. Time interval
For this pattern, begin from start/end times and clock convention. A legitimate next step is elapsed time. The common modelling error is assuming a time crosses noon/midnight incorrectly. Because the arithmetic after that assumption may still be perfect, the mistake can hide until the final interpretation.
Run the audit this way: build a timeline when day boundary matters. In training, mix 12-hour and 24-hour cases. The question is not “Did I assume anything?” but “Which assumptions are licensed by the stated mathematics and context, and which ones changed the problem without permission?”
20. Travel schedule
For this pattern, begin from departure/arrival/waiting information. A legitimate next step is journey duration or connection time. The common modelling error is assuming zero transfer time or that every service is immediately available. Because the arithmetic after that assumption may still be perfect, the mistake can hide until the final interpretation.
Run the audit this way: model each stated segment and wait separately. In training, use schedules with missed connections if timing is wrong. The question is not “Did I assume anything?” but “Which assumptions are licensed by the stated mathematics and context, and which ones changed the problem without permission?”
21. Transport capacity
For this pattern, begin from capacity per vehicle and required passengers. A legitimate next step is minimum number of vehicles. The common modelling error is rounding to nearest instead of rounding up. Because the arithmetic after that assumption may still be perfect, the mistake can hide until the final interpretation.
Run the audit this way: apply whole-number feasibility after the calculation. In training, compare 4.1, 4.5 and 4.9 vehicle cases. The question is not “Did I assume anything?” but “Which assumptions are licensed by the stated mathematics and context, and which ones changed the problem without permission?”
22. Packing capacity
For this pattern, begin from items per container or space limits. A legitimate next step is maximum complete groups. The common modelling error is rounding up when partial groups are not allowed. Because the arithmetic after that assumption may still be perfect, the mistake can hide until the final interpretation.
Run the audit this way: interpret complete groups before final integer choice. In training, pair minimum-required and maximum-fit problems. The question is not “Did I assume anything?” but “Which assumptions are licensed by the stated mathematics and context, and which ones changed the problem without permission?”
23. Money rounding
For this pattern, begin from prices, rates and transaction rules. A legitimate next step is final monetary amount. The common modelling error is rounding every intermediate value to cents even when calculation should continue more precisely. Because the arithmetic after that assumption may still be perfect, the mistake can hide until the final interpretation.
Run the audit this way: preserve working precision unless the context defines transaction-level rounding. In training, compare final-only rounding with stage-by-stage transaction rules. The question is not “Did I assume anything?” but “Which assumptions are licensed by the stated mathematics and context, and which ones changed the problem without permission?”
24. Tax or fee threshold
For this pattern, begin from stated bands or conditions. A legitimate next step is piecewise cost under those bands. The common modelling error is assuming one rate applies to the entire amount. Because the arithmetic after that assumption may still be perfect, the mistake can hide until the final interpretation.
Run the audit this way: map each threshold interval explicitly. In training, use tariff/tax examples with boundary values. The question is not “Did I assume anything?” but “Which assumptions are licensed by the stated mathematics and context, and which ones changed the problem without permission?”
25. Interest period
For this pattern, begin from rate and period definition. A legitimate next step is growth over matching periods. The common modelling error is assuming annual rate applies directly to monthly intervals without conversion or formula support. Because the arithmetic after that assumption may still be perfect, the mistake can hide until the final interpretation.
Run the audit this way: align rate period and time period before calculation. In training, use same nominal rate with different compounding periods. The question is not “Did I assume anything?” but “Which assumptions are licensed by the stated mathematics and context, and which ones changed the problem without permission?”
26. Currency exchange
For this pattern, begin from quoted direction and units. A legitimate next step is converted amount. The common modelling error is assuming rates are symmetric or reciprocal without checking quotation direction. Because the arithmetic after that assumption may still be perfect, the mistake can hide until the final interpretation.
Run the audit this way: write currency per currency explicitly. In training, reverse quotation direction in paired questions. The question is not “Did I assume anything?” but “Which assumptions are licensed by the stated mathematics and context, and which ones changed the problem without permission?”
27. Recipe scaling
For this pattern, begin from servings and ingredient quantities. A legitimate next step is scaled ingredients under proportional assumption. The common modelling error is assuming every practical quantity scales linearly if the context does not support it. Because the arithmetic after that assumption may still be perfect, the mistake can hide until the final interpretation.
Run the audit this way: use proportional scaling only where the problem defines or implies it. In training, contrast mathematical recipe scaling with packaging constraints. The question is not “Did I assume anything?” but “Which assumptions are licensed by the stated mathematics and context, and which ones changed the problem without permission?”
28. Mixture percentage
For this pattern, begin from component amount and total mixture. A legitimate next step is percentage concentration by the defined base. The common modelling error is using one component as denominator when total is required. Because the arithmetic after that assumption may still be perfect, the mistake can hide until the final interpretation.
Run the audit this way: write component/whole before calculating. In training, compare part-to-part and part-to-whole interpretations. The question is not “Did I assume anything?” but “Which assumptions are licensed by the stated mathematics and context, and which ones changed the problem without permission?”
29. Average speed
For this pattern, begin from total distance and total time. A legitimate next step is overall average speed. The common modelling error is averaging two speeds arithmetically without equal-time justification. Because the arithmetic after that assumption may still be perfect, the mistake can hide until the final interpretation.
Run the audit this way: return to total distance divided by total time. In training, use equal-distance and equal-time contrasts. The question is not “Did I assume anything?” but “Which assumptions are licensed by the stated mathematics and context, and which ones changed the problem without permission?”
30. Average rate
For this pattern, begin from total change over total interval. A legitimate next step is overall average rate. The common modelling error is averaging segment rates blindly. Because the arithmetic after that assumption may still be perfect, the mistake can hide until the final interpretation.
Run the audit this way: reconstruct total quantity and denominator. In training, test with unequal intervals. The question is not “Did I assume anything?” but “Which assumptions are licensed by the stated mathematics and context, and which ones changed the problem without permission?”
31. Mean from frequency table
For this pattern, begin from values and frequencies. A legitimate next step is weighted total divided by total frequency. The common modelling error is treating category values as equally frequent. Because the arithmetic after that assumption may still be perfect, the mistake can hide until the final interpretation.
Run the audit this way: write sum(xf)/sum(f) structure. In training, use same values with different frequencies. The question is not “Did I assume anything?” but “Which assumptions are licensed by the stated mathematics and context, and which ones changed the problem without permission?”
32. Median from grouped/list data
For this pattern, begin from ordered positions and count. A legitimate next step is middle value(s) according to data structure. The common modelling error is assuming the arithmetic mean is required because ‘average’ language feels familiar. Because the arithmetic after that assumption may still be perfect, the mistake can hide until the final interpretation.
Run the audit this way: identify statistic requested before calculation. In training, mix mean/median/mode prompts on same dataset. The question is not “Did I assume anything?” but “Which assumptions are licensed by the stated mathematics and context, and which ones changed the problem without permission?”
33. Probability denominator
For this pattern, begin from defined sample space. A legitimate next step is favourable/total when equally likely structure is appropriate. The common modelling error is using the number of favourable categories instead of outcomes. Because the arithmetic after that assumption may still be perfect, the mistake can hide until the final interpretation.
Run the audit this way: construct the actual outcome space. In training, use cases with unequal category sizes. The question is not “Did I assume anything?” but “Which assumptions are licensed by the stated mathematics and context, and which ones changed the problem without permission?”
34. Mutually exclusive events
For this pattern, begin from event definitions and overlap information. A legitimate next step is addition without overlap only when appropriate. The common modelling error is assuming named categories cannot overlap. Because the arithmetic after that assumption may still be perfect, the mistake can hide until the final interpretation.
Run the audit this way: test whether one outcome can satisfy both events. In training, use Venn-style counterexamples. The question is not “Did I assume anything?” but “Which assumptions are licensed by the stated mathematics and context, and which ones changed the problem without permission?”
35. Independent events
For this pattern, begin from explicit conditions or probability structure. A legitimate next step is multiplication where independence is justified. The common modelling error is assuming repeated events are independent automatically. Because the arithmetic after that assumption may still be perfect, the mistake can hide until the final interpretation.
Run the audit this way: ask whether one outcome changes the next probability. In training, contrast replacement/no-replacement cases. The question is not “Did I assume anything?” but “Which assumptions are licensed by the stated mathematics and context, and which ones changed the problem without permission?”
36. Complement
For this pattern, begin from event and total probability one. A legitimate next step is P(not A)=1-P(A). The common modelling error is using complement when events do not cover the full space. Because the arithmetic after that assumption may still be perfect, the mistake can hide until the final interpretation.
Run the audit this way: verify exhaustive partition. In training, include three-category spaces where a chosen pair is not complementary. The question is not “Did I assume anything?” but “Which assumptions are licensed by the stated mathematics and context, and which ones changed the problem without permission?”
37. Graph axis scale
For this pattern, begin from labels, units and tick intervals. A legitimate next step is correct coordinate/gradient interpretation. The common modelling error is assuming each grid square has value one. Because the arithmetic after that assumption may still be perfect, the mistake can hide until the final interpretation.
Run the audit this way: read one interval on each axis before data use. In training, use non-unit and non-zero-starting axes. The question is not “Did I assume anything?” but “Which assumptions are licensed by the stated mathematics and context, and which ones changed the problem without permission?”
38. Graph origin
For this pattern, begin from visible axis start and labels. A legitimate next step is intercepts relative to actual axes. The common modelling error is assuming graph begins at zero because many classroom examples do. Because the arithmetic after that assumption may still be perfect, the mistake can hide until the final interpretation.
Run the audit this way: read displayed scale before interpreting magnitude. In training, use truncated axes to test visual assumptions. The question is not “Did I assume anything?” but “Which assumptions are licensed by the stated mathematics and context, and which ones changed the problem without permission?”
39. Trend beyond data
For this pattern, begin from measured range. A legitimate next step is interpolation or model-supported prediction. The common modelling error is extending the same trend far outside the data. Because the arithmetic after that assumption may still be perfect, the mistake can hide until the final interpretation.
Run the audit this way: mark measured versus predicted range. In training, compare interpolation and extrapolation questions. The question is not “Did I assume anything?” but “Which assumptions are licensed by the stated mathematics and context, and which ones changed the problem without permission?”
40. Line of best fit
For this pattern, begin from data scatter and intended model. A legitimate next step is estimated trend relationship. The common modelling error is assuming every point lies exactly on the model. Because the arithmetic after that assumption may still be perfect, the mistake can hide until the final interpretation.
Run the audit this way: use the line as a model rather than a list of exact measured values. In training, practice with noisy but clear trends. The question is not “Did I assume anything?” but “Which assumptions are licensed by the stated mathematics and context, and which ones changed the problem without permission?”
41. Correlation
For this pattern, begin from paired data pattern. A legitimate next step is association statement. The common modelling error is claiming one variable caused the other. Because the arithmetic after that assumption may still be perfect, the mistake can hide until the final interpretation.
Run the audit this way: separate data relationship from causal explanation. In training, use confounded real-world examples. The question is not “Did I assume anything?” but “Which assumptions are licensed by the stated mathematics and context, and which ones changed the problem without permission?”
42. Function domain
For this pattern, begin from given/contextual allowed inputs. A legitimate next step is valid outputs over that domain. The common modelling error is using algebraic solutions outside the real-world or mathematical domain. Because the arithmetic after that assumption may still be perfect, the mistake can hide until the final interpretation.
Run the audit this way: test candidate values against original conditions. In training, mix positive-length and unrestricted algebra cases. The question is not “Did I assume anything?” but “Which assumptions are licensed by the stated mathematics and context, and which ones changed the problem without permission?”
43. Equation root
For this pattern, begin from derived solution candidates. A legitimate next step is valid roots after checking. The common modelling error is assuming every algebraic candidate is valid in the original equation. Because the arithmetic after that assumption may still be perfect, the mistake can hide until the final interpretation.
Run the audit this way: substitute into the original relation. In training, use transformed equations with invalid candidates. The question is not “Did I assume anything?” but “Which assumptions are licensed by the stated mathematics and context, and which ones changed the problem without permission?”
44. Negative solution
For this pattern, begin from algebraic result and context. A legitimate next step is accepted/rejected value based on domain. The common modelling error is changing a negative value to positive automatically because it ‘looks wrong’. Because the arithmetic after that assumption may still be perfect, the mistake can hide until the final interpretation.
Run the audit this way: ask whether the quantity can be directed or whether the context forbids negative values. In training, contrast displacement-like versus physical-length contexts. The question is not “Did I assume anything?” but “Which assumptions are licensed by the stated mathematics and context, and which ones changed the problem without permission?”
45. Whole-number solution
For this pattern, begin from continuous calculation and context. A legitimate next step is integer decision if objects are indivisible. The common modelling error is leaving a decimal number of people/items. Because the arithmetic after that assumption may still be perfect, the mistake can hide until the final interpretation.
Run the audit this way: interpret the object type before submission. In training, use contexts where ordinary rounding and ceiling/floor differ. The question is not “Did I assume anything?” but “Which assumptions are licensed by the stated mathematics and context, and which ones changed the problem without permission?”
46. Exact form
For this pattern, begin from question instruction and mathematical expression. A legitimate next step is symbolic answer. The common modelling error is assuming every answer should become a decimal because a calculator is available. Because the arithmetic after that assumption may still be perfect, the mistake can hide until the final interpretation.
Run the audit this way: preserve exact form when required. In training, mix π, fractions and radicals with approximate counterparts. The question is not “Did I assume anything?” but “Which assumptions are licensed by the stated mathematics and context, and which ones changed the problem without permission?”
47. Approximate form
For this pattern, begin from non-exact result and stated/default accuracy. A legitimate next step is rounded numerical answer. The common modelling error is reporting every calculator digit as meaningful precision. Because the arithmetic after that assumption may still be perfect, the mistake can hide until the final interpretation.
Run the audit this way: apply the current K210 accuracy rule only at the final stage unless specified otherwise. In training, compare display digits with justified reporting precision. The question is not “Did I assume anything?” but “Which assumptions are licensed by the stated mathematics and context, and which ones changed the problem without permission?”
48. Unit conversion
For this pattern, begin from given units and target units. A legitimate next step is converted value. The common modelling error is changing number without tracking dimension or direction. Because the arithmetic after that assumption may still be perfect, the mistake can hide until the final interpretation.
Run the audit this way: write unit factors or reason through scale explicitly. In training, use paired length/area/volume conversions. The question is not “Did I assume anything?” but “Which assumptions are licensed by the stated mathematics and context, and which ones changed the problem without permission?”
49. Compound unit
For this pattern, begin from quantities and unit relationship. A legitimate next step is rate/density/speed unit. The common modelling error is inverting numerator and denominator. Because the arithmetic after that assumption may still be perfect, the mistake can hide until the final interpretation.
Run the audit this way: say the compound unit in words before formula use. In training, contrast km/h with h/km, g/cm³ with cm³/g. The question is not “Did I assume anything?” but “Which assumptions are licensed by the stated mathematics and context, and which ones changed the problem without permission?”
50. Real-world graph
For this pattern, begin from axes, context and data. A legitimate next step is contextual interpretation. The common modelling error is treating mathematical shape as sufficient without reading what variables mean. Because the arithmetic after that assumption may still be perfect, the mistake can hide until the final interpretation.
Run the audit this way: translate each axis into a sentence before interpreting. In training, use identical graph shapes with different real-world meanings. The question is not “Did I assume anything?” but “Which assumptions are licensed by the stated mathematics and context, and which ones changed the problem without permission?”
51. Real-world optimum
For this pattern, begin from constraints and objective. A legitimate next step is best feasible option. The common modelling error is assuming mathematical maximum/minimum is feasible without contextual limits. Because the arithmetic after that assumption may still be perfect, the mistake can hide until the final interpretation.
Run the audit this way: apply constraints after finding candidate optimum. In training, use capacity/budget/time-limited examples. The question is not “Did I assume anything?” but “Which assumptions are licensed by the stated mathematics and context, and which ones changed the problem without permission?”
52. Real-world approximation
For this pattern, begin from calculated result and practical requirement. A legitimate next step is usable contextual answer. The common modelling error is assuming more decimal places always improve realism. Because the arithmetic after that assumption may still be perfect, the mistake can hide until the final interpretation.
Run the audit this way: match precision to instruction and decision context. In training, compare exact calculator output with sensible reported value. The question is not “Did I assume anything?” but “Which assumptions are licensed by the stated mathematics and context, and which ones changed the problem without permission?”
53. Real-world missing information
For this pattern, begin from available variables. A legitimate next step is what can and cannot be determined. The common modelling error is inventing a value to force a calculation. Because the arithmetic after that assumption may still be perfect, the mistake can hide until the final interpretation.
Run the audit this way: state insufficiency or express answer in terms of unknown when appropriate. In training, give problems with one intentionally missing quantity. The question is not “Did I assume anything?” but “Which assumptions are licensed by the stated mathematics and context, and which ones changed the problem without permission?”
54. Assumption from habit
For this pattern, begin from familiar textbook pattern. A legitimate next step is actual live conditions. The common modelling error is copying the method from a similar question without checking the changed condition. Because the arithmetic after that assumption may still be perfect, the mistake can hide until the final interpretation.
Run the audit this way: name one feature that must be true for the familiar method to apply. In training, use contrast pairs differing by one decisive condition. The question is not “Did I assume anything?” but “Which assumptions are licensed by the stated mathematics and context, and which ones changed the problem without permission?”
55. Assumption from diagram
For this pattern, begin from visual impression. A legitimate next step is stated mathematical facts. The common modelling error is measuring or inferring from appearance. Because the arithmetic after that assumption may still be perfect, the mistake can hide until the final interpretation.
Run the audit this way: treat diagram as representation unless measurement is authorised. In training, distort diagrams systematically in training. The question is not “Did I assume anything?” but “Which assumptions are licensed by the stated mathematics and context, and which ones changed the problem without permission?”
56. Assumption from calculator
For this pattern, begin from clean decimal output. A legitimate next step is mathematical validity. The common modelling error is believing a calculator result proves the model was correct. Because the arithmetic after that assumption may still be perfect, the mistake can hide until the final interpretation.
Run the audit this way: check representation, unit and context independently. In training, give wrong-formula calculations that still produce plausible numbers. The question is not “Did I assume anything?” but “Which assumptions are licensed by the stated mathematics and context, and which ones changed the problem without permission?”
57. Assumption from wording
For this pattern, begin from ordinary-language expectation. A legitimate next step is formal mathematical meaning. The common modelling error is interpreting words such as average, rate, minimum or proportion casually. Because the arithmetic after that assumption may still be perfect, the mistake can hide until the final interpretation.
Run the audit this way: translate the target word into its mathematical definition. In training, build vocabulary-to-operation drills. The question is not “Did I assume anything?” but “Which assumptions are licensed by the stated mathematics and context, and which ones changed the problem without permission?”
58. Assumption from previous subpart
For this pattern, begin from earlier result. A legitimate next step is new subpart target. The common modelling error is assuming every later part depends on the earlier answer. Because the arithmetic after that assumption may still be perfect, the mistake can hide until the final interpretation.
Run the audit this way: read the new question independently and identify actual dependency. In training, use multi-part questions with mixed dependent/independent later parts. The question is not “Did I assume anything?” but “Which assumptions are licensed by the stated mathematics and context, and which ones changed the problem without permission?”
59. Assumption from preferred method
For this pattern, begin from learner’s familiar technique. A legitimate next step is problem structure. The common modelling error is forcing every question into algebra, ratio or graphing because that method feels safe. Because the arithmetic after that assumption may still be perfect, the mistake can hide until the final interpretation.
Run the audit this way: choose the representation that matches the relation. In training, compare two valid methods and one forced method. The question is not “Did I assume anything?” but “Which assumptions are licensed by the stated mathematics and context, and which ones changed the problem without permission?”
60. Assumption from overconfidence
For this pattern, begin from quick recognition. A legitimate next step is verified problem type. The common modelling error is skipping constraints because the question looks familiar. Because the arithmetic after that assumption may still be perfect, the mistake can hide until the final interpretation.
Run the audit this way: name the changed condition before method execution. In training, use familiar surfaces with altered bases, units or domains. The question is not “Did I assume anything?” but “Which assumptions are licensed by the stated mathematics and context, and which ones changed the problem without permission?”
61. Assumption from anxiety
For this pattern, begin from uncertain feeling. A legitimate next step is actual evidence of model weakness. The common modelling error is adding extra conditions or re-solving without reason. Because the arithmetic after that assumption may still be perfect, the mistake can hide until the final interpretation.
Run the audit this way: identify the exact unsupported step before changing the model. In training, apply confidence-weighted checking rather than emotional rework. The question is not “Did I assume anything?” but “Which assumptions are licensed by the stated mathematics and context, and which ones changed the problem without permission?”
62. Assumption during checking
For this pattern, begin from first solution state. A legitimate next step is independent verification. The common modelling error is checking with the same hidden assumption and getting the same answer. Because the arithmetic after that assumption may still be perfect, the mistake can hide until the final interpretation.
Run the audit this way: choose a check that does not inherit the original assumption. In training, use units, estimate, boundary or substitution as independent tests. The question is not “Did I assume anything?” but “Which assumptions are licensed by the stated mathematics and context, and which ones changed the problem without permission?”
The four-column assumption audit
- Given: facts explicitly stated.
- Derived: facts proved from the given information.
- Context: real-world requirements that determine interpretation.
- Unsupported: anything added without mathematical or contextual justification.
Use all four columns in slow practice. In the examination, compress the method to a quick question whenever a modelling step feels suspicious: given, derived, context or invented?
Connect to the K210 paper
K210 assesses interpretation, problem formulation and contextual interpretation, and Paper 2 includes a real-world scenario at the end of Section A. The assumption audit is therefore most useful where the learner has to decide what Mathematics the situation permits before calculating. It also supports the Section B choice because unfamiliar wording can tempt the learner to invent conditions instead of reading the actual structure.
Links back into the learning system
Use Vol 0055 when an unsupported assumption leaves a distinctive wrong-answer signature. Use Vol 0035 for feasibility and domain constraints, the Mathematics Hub for underlying capability, the PSLE Learner’s Guide for earlier representation habits, and Examination Craft for checking and recovery under time.
Readiness criteria
- You can separate stated information from visual impression.
- You label the percentage, ratio or rate base before calculating.
- You identify when constancy, proportionality or similarity has actually been justified.
- You treat units and domains as constraints on the model.
- You do not invent missing information to force a calculation.
- You interpret real-world integer and feasibility conditions after solving.
- Your checking uses at least one method that does not inherit the original assumption.
Official-source discipline
For current K210 assessment objectives, paper structure, accuracy notes and real-world-context guidance, use the official 2027 G2 Mathematics K210 syllabus. If SEAB updates the syllabus, the current official document takes priority over this guide.
Final rule: do not let the clean calculation hide a dirty assumption
Many advanced Mathematics errors are not arithmetic errors. They begin one step earlier, when the learner quietly decides what the diagram means, which value is the base, whether a rate stays constant or whether a real-world quantity can be fractional.
Audit the model before trusting the calculation. Use what is stated, derive what Mathematics permits, interpret what the context requires and refuse to invent the rest. That is how a real-world K210 problem remains the problem on the page rather than the easier problem the learner accidentally created.
