From an educational practitioner's perspective, effective competition preparation is not about the total number of problems solved, but about understanding the nature of question types and establishing consistent response habits. This article is based on the official SASMO (Singapore & Asian Schools Math Olympiad) competition format — 6 grade groups, 25 questions (Section A: 15 questions + Section B: 10 questions) — and outlines the functional positioning of each question type and the logic behind training them. The goal is to help students move beyond the "finish and move on" mindset and adopt a "targeted practice by type" preparation model. All analysis strictly corresponds to the exam structure published on the official website, with no unverified question-type descriptions or difficulty classifications introduced.
I. Restoring the Question-Type Structure
The official SASMO competition clearly adopts a two-section structure: Section A contains 15 questions, and Section B contains 10 questions, for a total of 25 questions. This structure has remained consistent since the competition's inception, with no official data indicating any adjustments.
Section A questions focus on fundamental concept understanding and direct application, commonly including number reasoning, shape counting, simple logical judgment, and arithmetic transformations. Section B questions emphasise multi-step deduction, condition integration, and strategic choice, often incorporating composite elements such as geometric relationships, sequence patterns, and case-based classification. The two sections together cover the main line of mathematical ability development from Grades 2 to 10, but their functional roles are clearly distinct.
Question-Type Function Comparison
| Section | Number of Questions | Core Assessment Focus |
|---|---|---|
| Section A | 15 | Knowledge accuracy and response stability |
| Section B | 10 | Thinking flexibility and problem-transformation ability |
Key Conclusion: Section A is a baseline ability test; Section B is a test of thinking progression. The two cannot be trained in isolation, but they require differentiated practice pacing and feedback priorities.
II. Training Logic by Question Type
The official competition does not publish specific question-type names or classifications, but past papers consistently reveal a stable distribution of ability dimensions. Accordingly, questions can be divided into three categories based on problem-solving behaviour characteristics, each with corresponding training objectives:
Computational Questions
These involve numerical operations, unit conversions, and formula substitution as the main operations. They commonly appear in the first 10 questions of Section A and in the middle of Section B. The training focus is not on increasing calculation speed, but on establishing a "verification trigger mechanism" — such that when a result is non-integer, outside the range of common sense, or significantly inconsistent with the magnitude of the answer options, a review process is automatically initiated.
Reasoning Questions
These involve sequence patterns, logical elimination, and possibility enumeration. They frequently appear in the last 5 questions of Section A and the first 6 questions of Section B. The training approach is not to exhaust all possible paths, but to solidify a "minimal validation sample selection awareness" — for example, when testing a sequence pattern, prioritise substituting the first term, middle term, and last term rather than verifying every term individually.
Modelling Questions
These require transforming textual descriptions into mathematical structures (such as equations, inequalities, set diagrams, coordinate representations). They are concentrated in the last 4 questions of Section B. The core of training is "translation accuracy from language to symbols." It is recommended to use the "three-line note-taking method": the first line copies the key sentence from the problem statement, the second line writes the corresponding mathematical expression, and the third line annotates the domain constraints of the variables. This approach significantly reduces marks lost due to comprehension.
In summary: Computational questions emphasise closed-loop verification, reasoning questions emphasise sampling strategies, and modelling questions emphasise translation accuracy — the three question types require different training methods, but all point to the same goal: reducing non-knowledge-based errors.
III. Principles for Using Past Papers
The official SASMO competition does not provide difficulty classifications or knowledge-point weighting explanations. Therefore, the value of past papers lies not in question prediction, but in exposing individual thinking blind spots. The following three principles must be followed when using them:
First Principle
Complete the full set of 25 questions within the time limit, but only mark Section A (15 questions); do not score Section B for the time being — only mark the points where thinking got stuck. This quickly locates gaps in foundational stability.
Second Principle
For questions in Section B that were not solved, trace back to questions in Section A that test the same type of ability, and work backwards to identify the ability. For example, if a geometry question in Section B is blocked, review how you performed on the foundational questions in Section A that involve angle relationships and symmetry properties.
Third Principle
Each set of past papers should be used in at least two rounds: the first round simulates actual exam conditions, and the second round redoes the questions grouped by type — re-practice all computational questions together, then all reasoning questions together, to reinforce the muscle memory of responding to each question type.
Key Conclusion: Past papers are not the end goal; they are diagnostic tools. Repeated use is more effective than blind practice with new questions.
IV. Preparation Phase Task Breakdown
| Phase | Core Tasks | Question-Type Focus |
|---|---|---|
| Knowledge Mapping Phase (Refer to historical pacing) |
Complete the mapping between in-school knowledge framework and SASMO ability requirements | Overview of all question types; identify your own strength question types |
| Targeted Breakthrough Phase (Refer to historical pacing) |
Concentrated training on weaker question types and error cause categorisation | Grouped training by computational / reasoning / modelling categories |
| Mock Exam Adaptation Phase (Refer to historical pacing) |
Full-paper timed training and question-type score-rate statistics | Section A: ensure accuracy; Section B: ensure completeness of reasoning |
In summary: The phase does not depend on specific dates, but on the state of ability achievement. When performance on a particular question type remains stable across two consecutive sets of past papers, it is time to move on to the next phase.
The question-type design of the SASMO competition serves the purpose of layered ability identification. Its true value lies in using a standardised structure to reveal the authentic ability map. By returning to the official competition format itself and organising training around question types, one can avoid superficial learning and truly achieve "practise one question, master one category."
