Mathematics

QCAA Specialist Mathematics IA1 PSMT Guide: How to Aim for 20/20

A current QCE Specialist Mathematics IA1 PSMT guide to formulation, complex procedures, technology, verification and evaluation, with full-mark insights.

By Sylligence · Published 2026-07-17 · Updated 2026-07-17 · 9 min read

A top-band Specialist Mathematics PSMT uses sophisticated mathematics as a connected solution, not as decoration. It justifies the assumptions that make the context modelable, uses technology efficiently, verifies results during development, and evaluates the solution against those assumptions, observations and task requirements. A large matrix or complicated screenshot is not enough by itself.

This guide combines the current QCAA syllabus and subject report with anonymised insights from two privately supplied Specialist Mathematics PSMTs that the student confirmed received full marks.

Specialist Mathematics IA1 at a glance

The Specialist Mathematics 2025 v1.4 syllabus applies to students completing the course in 2026 or beyond.

| Requirement | Current QCAA condition | | --- | --- | | Instrument | IA1 Problem-solving and modelling task | | Weighting | 20% | | Response | Individual written response | | Length | Up to 10 A4 pages and 2000 words | | Class time | 3 hours, plus students' own time out of class | | Subject matter | At least one topic from Unit 3 or Unit 4 | | Technology | Required beyond simple computation or word processing | | Appendixes | May include raw data and repeated calculations but are not marked |

| Criterion | Marks | Top-band evidence | | --- | ---: | --- | | Formulate | 4 | Justified important assumptions, observations and mathematical translation | | Solve | 7 | Accurate mathematics, efficient technology and a complete solution | | Evaluate | 5 | Verified results, justified reasonableness, strengths and limitations | | Communicate | 4 | Correct mathematical language, standalone organisation and mathematically justified decisions |

What does a full-mark reasoning chain look like?

Use this as a diagnostic, not a required set of headings:

problem -> success measure -> important assumptions/observations -> mathematical representation -> related procedures -> verification -> contextual test -> refinement -> final judgment

The two supplied full-mark responses modelled sporting competitions with dominance or supremacy matrices. Both started with a baseline, tested predictions against later results, diagnosed what the baseline ignored, changed the mathematics and measured whether prediction performance improved.

The transferable lesson is model refinement based on evidence. You do not need a sports context or a dominance matrix.

How do you write a strong Formulate section?

Define success before judging the model

If a model will be called accurate, fair, efficient or reasonable, define how that word will be tested.

One supplied response set a prediction benchmark and required distinct rankings so each future match produced an unambiguous prediction. Another compared prediction accuracy on rounds excluded from model construction.

Strong success measures are:

• relevant to the actual problem
• measurable or clearly testable
• established before the final result is known
• used later in evaluation

Avoid changing the benchmark after seeing the model's performance.

Justify assumptions through their mathematical effect

For example, assuming team strength remains constant is not merely a statement about sport. It allows early-round results to be used for later predictions. It also creates a limitation: injury, selection and form changes may weaken later predictions.

That one assumption should therefore appear across the response:

1. Formulate: justify why it makes modelling possible.
2. Solve: use historical data under that assumption.
3. Evaluate: explain when the assumption makes the prediction less reasonable.

This is the kind of criterion connection that makes a PSMT coherent.

What counts as complex procedure in Specialist Mathematics?

The current QCAA subject report describes complex procedures as multiple related mathematical steps. A procedure is stronger when each step depends on the previous step and advances the solution.

In a matrix context, that might involve:

1. constructing a matrix from contextual data
2. applying justified matrix operations
3. forming a ranking or model output
4. resolving a limitation such as tied rankings
5. testing predictions on unseen outcomes
6. modifying inputs or weights for a mathematically justified reason
7. comparing the revised output with the baseline

The complexity lies in the connected reasoning. It does not come from displaying a 16-by-16 matrix with no explanation.

In other tasks, complex procedures may involve vectors, proof, complex numbers, calculus, differential equations or statistical inference. Use the mathematics the problem genuinely needs.

How should technology be used?

Technology is efficient when it makes the solution faster, easier or more effective while the mathematical logic remains inspectable.

For a matrix model, a spreadsheet can efficiently perform multiplication, calculate row sums, update weights and test predictions. The report should still explain:

• what each matrix entry represents
• why the operation is valid
• why any weight or scale was selected
• how the output becomes a contextual ranking or decision
• which result was independently checked

Do not fill the main report with unreadable raw matrices if a well-labelled excerpt, formula and output table communicate the reasoning. Put repetitive data in an appendix, remembering that the appendix itself is not marked.

How do you refine a model properly?

Refinement should repair a demonstrated limitation.

One full-mark response found that a binary win-loss model treated a narrow victory like a dominant victory. It then introduced scaled winning-margin information. The scale was justified relative to typical match scoring so one extreme game did not overwhelm the base value of a win. Prediction accuracy increased from roughly 54% to 63%.

The other response found that outcome-only rankings left important performance differences hidden. A revised model used the proportion of goals scored, changed rankings for explainable reasons and improved prediction accuracy from about 72% to 83%.

The general refinement pattern is:

The current model cannot represent [specific feature], which affects [result or decision]. Modify [mathematical component] using [justification], then compare [same performance measure] before and after.

A revision is not automatically better because it is more complicated. Measure the change.

How do you verify and evaluate the solution?

Verify results during development

Verification could include:

• checking a sample matrix operation manually
• reproducing an output with a second method
• confirming a ranking from the underlying data
• testing predictions on data not used to build the model
• checking special or boundary cases
• confirming that weights or probabilities behave as intended

Out-of-sample testing is particularly useful for predictive models because it checks performance beyond the data used to construct them.

Evaluate through the assumptions and observations

The top Evaluate band requires more than an accuracy percentage. Ask:

• Does the result meet the success measure defined in Formulate?
• Which earlier assumption makes the result useful?
• Which earlier assumption limits where it can be used?
• What does the model capture that a simpler method misses?
• What important variable or mechanism remains absent?
• Does the model create artefacts, such as self-dominance or inflated ranking values?
• Is the improvement large enough to justify the extra complexity?

Justify strengths and limitations

| Weak statement | Stronger direction | | --- | --- | | The spreadsheet is adaptable | Explain which inputs can change and whether the mathematical structure remains valid | | The model ignores injuries | Explain how changing team strength undermines predictions built from early data | | The model is accurate | State the test data and benchmark, then interpret the achieved rate | | Higher-order matrices are better | Show which tie or schedule-strength problem they resolve and whether prediction improves |

A warning about dominance-matrix tasks

The 2025 Specialist Mathematics subject report specifically says that if dominance matrices are used, the task must allow students to go beyond simple win-loss data. That advice is directed at assessment design, but it also reveals a useful quality check for a student response: merely tallying wins is unlikely to show the full depth expected in an authentic Specialist modelling solution.

Possible extensions beyond simple win-loss data include:

• indirect dominance through common opponents
• justified weighting of different orders
• margin or proportion data
• time-weighted performance
• uncertainty or probability
• changing conditions or model comparison

These are examples, not a checklist. Use only what fits the task.

Specialist PSMT checklist

Formulate

• [ ] Success or reasonableness is defined before the final result.
• [ ] Assumptions and observations are necessary for mathematisation.
• [ ] Each important assumption or observation has a justification and modelling effect.
• [ ] Variables, representations, constraints and objectives are clear.

Solve

• [ ] The solution contains multiple related mathematical steps.
• [ ] Specialist mathematics drives the solution rather than decorating it.
• [ ] Technology is efficient and its role is explained.
• [ ] The complete problem is answered and outputs are interpreted.

Evaluate

• [ ] Results are verified during development.
• [ ] The model is tested against the pre-declared success measures.
• [ ] Reasonableness returns to assumptions and observations.
• [ ] Strengths and limitations explain a real mathematical or contextual effect.
• [ ] Refinements repair an evidenced limitation and are re-tested.

Communicate

• [ ] Mathematical notation, language, conventions and representations are correct.
• [ ] The response can be read without the task sheet.
• [ ] Large outputs are summarised without hiding the mathematical logic.
• [ ] Decisions are justified with mathematics.

Common questions

Do I need two models in a Specialist PSMT?

No fixed number is required. A baseline and refinement can be powerful when the second model repairs a measured weakness, but one complete, well-verified solution may be enough for another task.

Is a large matrix automatically a complex procedure?

No. Complexity comes from multiple related mathematical steps and justified decisions, not the dimensions of the output.

Can Excel do most of the calculation?

Yes, when it is the efficient tool for the problem. Your report still needs to show the representation, operations, reasoning, checks and interpretation.

Do I have to use dominance matrices?

No. Specialist IA1 may use subject matter from Unit 3 or Unit 4. The school task determines the context and relevant mathematics.

Can a full-mark response contain a typo or awkward sentence?

Yes. QCAA judgments use a best-fit approach. Surface errors should still be corrected, but they do not automatically erase criterion-wide top-band evidence.

Continue your review

• Read the Mathematical Methods PSMT guide for the shared 2025 Mathematics criteria.
• Review QCAA ISMG marking criteria explained.
• Use Sylligence assignment feedback and select Specialist Mathematics IA1 PSMT.

Sources and methodology

This guide was checked against the Specialist Mathematics 2025 v1.4 syllabus, the Specialist Mathematics 2025 subject report and the QCAA Specialist Mathematics subject page.

Sylligence also reviewed two privately supplied full-mark Specialist Mathematics PSMTs. Both used sporting prediction contexts and matrix models, but their wording, identity and full responses are not reproduced. The numerical changes above are brief, anonymised illustrations of modelling decisions, not task answers to copy or guarantees of a result.