Every one of the 90 marks was coded to a topic, set against the forecast and the revision priorities published on this site, and scored honestly, including the misses.
Equations and inequalities carried the most marks (11), then the real-world problem (10) and vectors (8). The paper opened with a graph and a numbers-and-sequences question, ended with the real-world problem as usual, and put its 14-mark algebra question third.
Table, plot and turning point of y = 4/x − 2/x² + 4 (two separate branches); sketch of y = −(x − 2)(x + 4) with its intercepts.
Functions & graphs 7Powers and squares (3ᵃ + b = 223), cube-root bounds, a 1 s.f. estimate with over/under reason; nth term of a linear sequence, a “why one more than a multiple of 3” explanation, a geometric sequence’s sixth term and a threshold term.
Sequences & patterns 7Numbers & indices 4An inequality with a fraction; expansion with negative indices; a fractional equation that becomes a quadratic (formula, 2 d.p., 5 marks); eliminating a parameter between two formulae.
Equations & inequalities 7Algebra & formulae 5Numbers & indices 2Volume of a prism with a composite cross-section and the time to fill it; successive percentage changes in circumference leading to an area change; volume and curved-area ratios of equal-height cylinders.
Mensuration 5Speed, rates & travel graphs 3Percentages 3An accurate construction of a quadrilateral; an AAS congruence proof with reasons; an angle proof using parallel lines, congruent triangles and an isosceles triangle.
Congruence & similarity 3Construction 2Angles & polygons 2A pie chart with a difference; a combined mean; a profit matrix (P × (D − C)) and what its total means; repeated spins and “at least one of each”.
Statistics 6Matrices 3Probability 2Area ½ab sin C forms a quadratic; solve by factorisation; reject a root in context; cosine rule and the obtuse angle.
Pythagoras, trig & bearings 7Equations & inequalities 4Position vectors and the magnitude of a fraction of a vector; AB in terms of a and b; collinearity of three points and a section ratio.
Vectors 8Unit cost share, a population extrapolated from a line graph, a cost model with service capacity, and a loan-interest claim to be judged.
Real-world problem 10Every mark counted once under its main topic (the final real-world question is kept whole, as in the earlier analysis). The first table shows the paper question by question; the second shows Paper 2 over the last ten years with the 2026 column in gold.
Marks: 01–34–67–910–1415+ Gold outline = 2026
| Topic | Q1 | Q2 | Q3 | Q4 | Q5 | Q6 | Q7 | Q8 | Q9 | Total |
|---|---|---|---|---|---|---|---|---|---|---|
| Equations & inequalities | 7 | 4 | 11 | |||||||
| Real-world problem | 10 | 10 | ||||||||
| Vectors | 8 | 8 | ||||||||
| Functions & graphs | 7 | 7 | ||||||||
| Sequences & patterns | 7 | 7 | ||||||||
| Pythagoras, trig & bearings | 7 | 7 | ||||||||
| Numbers & indices | 4 | 2 | 6 | |||||||
| Statistics | 6 | 6 | ||||||||
| Algebra & formulae | 5 | 5 | ||||||||
| Mensuration | 5 | 5 | ||||||||
| Matrices | 3 | 3 | ||||||||
| Percentages | 3 | 3 | ||||||||
| Congruence & similarity | 3 | 3 | ||||||||
| Speed, rates & travel graphs | 3 | 3 | ||||||||
| Construction | 2 | 2 | ||||||||
| Angles & polygons | 2 | 2 | ||||||||
| Probability | 2 | 2 | ||||||||
| Question total | 7 | 11 | 14 | 11 | 7 | 11 | 11 | 8 | 10 | 90 |
0 marks in Paper 2: Ratio & proportion, Financial arithmetic, Sets, Circle properties, Coordinate geometry.
| Topic | 2016 | 2017 | 2018 | 2019 | 2020 | 2021 | 2022 | 2023 | 2024 | 2025 | 2026 | 2023–25 avg | My forecast | Usual range | 2026 vs range |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Real-world problem | 9 | 9 | 10 | 9 | 9 | 10 | 10 | 10 | 11 | 9 | 10 | 10.0 | 10.0 | 9–10 | inside |
| Mensuration | 6 | 10 | 5 | 11 | 12 | 11 | 10 | 13 | 7 | 6 | 5 | 8.7 | 8.9 | 6–11 | below |
| Equations & inequalities | 12 | 13 | 11 | 4 | 7 | 6 | 5 | 8 | 12 | 10 | 11 | 10.0 | 8.8 | 6–12 | inside |
| Functions & graphs | 12 | 8 | 11 | 12 | 5 | 9 | 11 | 11 | 7 | 7 | 9.7 | 8.8 | 7–11 | inside | |
| Pythagoras, trig & bearings | 8 | 11 | 14 | 11 | 5 | 6 | 9 | 11 | 7 | 6.7 | 7.1 | 4–11 | inside | ||
| Coordinate geometry | 8 | 5 | 10 | 5 | 7 | 9 | 6 | 7.3 | 6.2 | 0–8 | 0 (range includes 0) | ||||
| Statistics | 4 | 8 | 5 | 7 | 5 | 5 | 7 | 4 | 5 | 5 | 6 | 4.7 | 5.4 | 5–7 | inside |
| Probability | 5 | 4 | 4 | 4 | 7 | 4 | 6 | 5 | 4 | 6 | 2 | 5.0 | 5.1 | 4–6 | below |
| Algebra & formulae | 5 | 4 | 4 | 5 | 4 | 6 | 3 | 4 | 4 | 5 | 5 | 4.3 | 4.4 | 4–5 | inside |
| Sequences & patterns | 8 | 10 | 8 | 7 | 10 | 7 | 3.3 | 3.3 | 0–8 | inside | |||||
| Vectors | 4 | 3 | 8 | 3 | 11 | 4 | 4 | 8 | 1.3 | 3.0 | 0–5 | above | |||
| Congruence & similarity | 5 | 4 | 4 | 5 | 9 | 3 | 2 | 3 | 1.7 | 2.9 | 0–5 | inside | |||
| Percentages | 3 | 5 | 2 | 4 | 3 | 5 | 4 | 2 | 1 | 3 | 2.3 | 2.4 | 2–4 | inside | |
| Matrices | 4 | 4 | 5 | 5 | 3 | 3 | 2.7 | 2.3 | 0–5 | inside | |||||
| Ratio & proportion | 2 | 3 | 2 | 6 | 3.7 | 2.2 | 0–2 | 0 (range includes 0) | |||||||
| Financial arithmetic | 11 | 2 | 3 | 10 | 0.0 | 2.0 | 0–4 | 0 (range includes 0) | |||||||
| Numbers & indices | 2 | 1 | 2 | 1 | 1 | 1 | 4 | 3 | 6 | 2.7 | 2.0 | 1–2 | above | ||
| Circle properties | 5 | 2 | 5 | 2 | 4 | 1.3 | 1.7 | 0–4 | 0 (range includes 0) | ||||||
| Angles & polygons | 4 | 6 | 2 | 2.0 | 1.5 | 0–1 | above | ||||||||
| Sets | 2 | 4 | 1.3 | 1.1 | 0–0 | 0 (range includes 0) | |||||||||
| Speed, rates & travel graphs | 2 | 2 | 3 | 2 | 2 | 3 | 1.3 | 1.0 | 0–2 | above | |||||
| Construction | 2 | 0.0 | 0.0 | 0–0 | above |
“Usual range” is the middle 80% of 2016–25 Paper 2s. 2026 is inside it for 15 topics; above: Numbers & indices, Speed, rates & travel graphs, Angles & polygons, Construction, Vectors; below: Mensuration, Probability. 2020 and 2021 were full papers in 2020–21 and are scaled the same way.
Checked part by part against the 4052 content list.
Short answer: the big picture matched and the details were uneven. All five of the “revise first” topics from the guides appeared, and the paper’s shape (a graph question, an equation-modelling question, matrices, trigonometry, a real-world finish) is what both the forecast and the specimen papers described. Two confident calls were wrong, and I made one of them against the pattern in my own data.
| Topic | My verdict | P2 marks | In recent P2s | Result |
|---|---|---|---|---|
| Real-world problem | CERTAIN | 10 | 10/10 | ✔ appeared |
| Equations & inequalities | VERY LIKELY | 11 | 10/10 | ✔ appeared |
| Functions & graphs | VERY LIKELY | 7 | 9/10 | ✔ appeared |
| Pythagoras, trig & bearings | VERY LIKELY | 7 | 8/10 | ✔ appeared |
| Statistics | VERY LIKELY | 6 | 10/10 | ✔ appeared |
| Mensuration | VERY LIKELY | 5 | 10/10 | ✔ appeared |
| Matrices | VERY LIKELY | 3 | 5/10 | ✔ appeared |
| Probability | VERY LIKELY | 2 | 10/10 | ✔ appeared |
| Vectors | LIKELY | 8 | 7/10 | ✔ appeared |
| Sequences & patterns | LIKELY | 7 | 5/10 | ✔ appeared |
| Algebra & formulae | LIKELY | 5 | 10/10 | ✔ appeared |
| Percentages | LIKELY | 3 | 9/10 | ✔ appeared |
| Coordinate geometry | LIKELY | 0 | 7/10 | ✘ 0 marks |
| Numbers & indices | POSSIBLE | 6 | 8/10 | not scored |
| Speed, rates & travel graphs | POSSIBLE | 3 | 5/10 | not scored |
| Congruence & similarity | POSSIBLE | 3 | 7/10 | not scored |
| Angles & polygons | POSSIBLE | 2 | 2/10 | not scored |
| Ratio & proportion | POSSIBLE | 0 | 4/10 | not scored |
| Financial arithmetic | POSSIBLE | 0 | 4/10 | not scored |
| Circle properties | POSSIBLE | 0 | 5/10 | not scored |
| Construction | UNLIKELY | 2 | 0/10 | ✘ 2 marks |
| Sets | UNLIKELY | 0 | 2/10 | ✔ absent |
Scoring. “Certain”, “very likely” and “likely” count as a call that the topic appears; “unlikely” as a call that it does not; “possible” is not scored. That is 15 calls: 13 right, 2 wrong. Be fair to the base rate: 9 of the 13 “appears” calls were topics that show up in at least 8 of the last 10 Paper 2s anyway. The calls that actually took a risk were matrices, sequences, vectors, coordinate geometry, sets and construction: 4 right, 2 wrong.
| Claude specimen Paper 2 (blueprint) | 78% | |
| Plain average of 2016–25 Paper 2s | 77% | |
| My recency-weighted forecast | 76% | |
| Plain average of 2023–25 | 74% | |
| ChatGPT specimen Paper 2 (blueprint) | 71% | |
| Just repeat the 2025 paper | 68% |
Share of the 90 marks that sat in the topics a forecast expected (Σ min(actual, expected) ÷ 90). Differences of a few points are noise: the simple ten-year average did as well as my weighted forecast, and my specimen paper’s blueprint scored highest by a point or two. None of these measures says a particular question was predicted. ChatGPT’s blueprint splits marks evenly between the topics named in each question, which lowers its score a little.
| What I told students to expect | Result | What the paper did |
|---|---|---|
| Graph question (almost certain) Complete a table, plot a curve (a sum of power terms), draw a tangent, read an intersection with a line, link to an equation. | ≈ Topic and curve type right, two skills missing | Q1 was exactly a sum of power terms with a table and a plot (7 marks). There was no tangent and no line to intersect: the other two marks were a turning point and a sketch. |
| Equation-modelling question Form a quadratic from a situation, solve with the formula or by completing the square, reject a root. | ✔ Matched | Q7 formed a quadratic from an area, solved it and rejected a root with a reason; Q3(c) needed the quadratic formula (5 marks). Equations were the biggest P2 topic (11). |
| Matrices question A matrix product in a context, with the meaning of an entry. | ✔ Matched | Q6(b): a profit matrix P × (D − C), then “what does the sum of the elements represent?” |
| Non-right-angled trigonometry Sine and cosine rule, ½ab sin C, bearings, a 3-D or elevation part. | ≈ Mostly matched | Q7 used ½ab sin C, the cosine rule and the obtuse-angle idea (7 marks). No bearings and no 3-D. |
| Statistics with probability Cumulative frequency, standard deviation and a comparison; probability from the same data or a tree. | ≈ Topic right, skills different | Q6 had statistics (pie chart, combined mean) and probability (repeated spins), but not cumulative frequency, standard deviation or a tree diagram. |
| Vector or coordinate geometry Position vectors, ratios, collinearity, often with a coordinate part. | ≈ Vectors yes, coordinates no | Q8 was a full vector question: position vectors, a ratio on a line, collinearity (8 marks, above its usual range). No coordinate geometry anywhere in P2. |
| Mensuration with solids Composite solids, spheres, pyramids, segments, radians, similar solids. | ≈ Partly matched | Q4: a prism with a composite cross-section, and cylinders compared by ratio. No sphere, pyramid, segment, radians or similar-solid scaling. |
| Sequences or an algebraic proof nth term, a term question and a proof. | ✔ Matched | Q2(d): nth term, “explain why one more than a multiple of 3”, a geometric sixth term and a threshold term (7 marks). |
| Real-world problem (certain) Money, rates, a model or graph, ending in a justified recommendation; interest likely here. | ✔ Matched | Q9: costs, a graph reading, capacity limits and compound interest, ending with a verdict on a claim (10 marks). |
| Source | What it said | What happened |
|---|---|---|
| ChatGPT’s textbook guide First priority | Extended graphs, quadratic modelling, matrices, non-right-angle trigonometry, the final real-world problem | 5 of 5 appeared |
| ChatGPT’s guide Next priority | Coordinate geometry, vectors, cumulative frequency / SD / data displays, composite solids | 3 of 4: vectors yes, data displays and prisms partly (no cumulative frequency or SD), coordinate geometry no |
| ChatGPT’s guide Short check | Sequences; construction, polygons and map scales | Both appeared (sequences 7 marks; construction 2). It caught construction; I had rated it unlikely. |
| Tutor’s topic list 14 topics in bold | Vectors, matrices, CF curve, angle properties, circle properties, graphical solutions, mensuration, probability, trigonometry, patterns, quadratic word problems, arithmetic (tax, GST, hire purchase), scales and maps, real-world problems | 10 of 14 topics appeared. Absent: CF curve, circle properties, tax/GST-type arithmetic, scales and maps. Graphs appeared as plotting and a sketch, not graphical solutions. Coordinate geometry was left off the list, correctly. |
Non-bold items on the tutor’s list: construction appeared; misleading graphs, speed–time graphs and sector/segment did not.
| Topic | Result | Who was right |
|---|---|---|
| Coordinate geometry | 0 marks in P2 (5 in P1 Q14). The six-year run of P2s with coordinate geometry (2020–25) ended. | The tutor’s list. I rated it “likely” and ChatGPT’s guide put it under “Next”; both were wrong this year. My own table already showed that in 2016–25, when P1 had it (3 years) P2 had it in only 1; with 2026 that is 1 of 4. I weighted the 2023–25 average above that pattern, and it was the wrong call. |
| Construction | 2 marks, in P2 | ChatGPT’s guide and the tutor’s list. I said “unlikely” (none in P2 for ten years). Unlikely is not impossible: construction was in P1 in about half of recent years, so it can turn up in P2 too, and ten years of zeros in P2 was thin evidence. |
| Circle properties | 0 marks | Nobody. I said “possible”, the tutor’s list had it bold and ChatGPT’s guide put circle proofs under “maintain”; each was a reasonable hedge and none was borne out. |
| Matrices | 3 marks, in P2 only | My “very likely”: matrices were in exactly one paper every year of 2016–25, and again in 2026 (P1 none, P2 3). That is now 11 years out of 11. |
| Vectors | 8 marks | All sources had it high; the size (a whole question after two quiet years) beat every forecast. |
Each part was given one main topic. ChatGPT’s published ledger for the same paper agrees with mine on 84 of 90 marks. The only differences are two Question 4 parts:
Other close calls, and what they would change:
None of these recodings changes a conclusion above.
| Part | Marks | Topic | Content |
|---|---|---|---|
| 1a(i) | 1 | Functions & graphs | Complete a table of values for y = 4/x − 2/x² + 4 |
| 1a(ii) | 3 | Functions & graphs | Plot and draw a sum-of-powers curve (two separate branches) |
| 1a(iii) | 1 | Functions & graphs | Turning point read from the graph |
| 1b | 2 | Functions & graphs | Sketch a quadratic given in factorised form, with intercepts |
| 2a | 1 | Numbers & indices | 3ᵃ + b = 223: powers and square numbers |
| 2b | 1 | Numbers & indices | Smallest integer whose cube root rounds to 5 |
| 2c(i) | 1 | Numbers & indices | Estimate by rounding each number to 1 s.f. |
| 2c(ii) | 1 | Numbers & indices | Over- or under-estimate, with a reason |
| 2d(i) | 2 | Sequences & patterns | nth term of a linear sequence |
| 2d(ii) | 1 | Sequences & patterns | Explain why each term is one more than a multiple of 3 |
| 2d(iii)(a) | 2 | Sequences & patterns | Sixth term of a geometric sequence |
| 2d(iii)(b) | 2 | Sequences & patterns | Smallest term of S above a given value |
| 3a | 2 | Equations & inequalities | Linear inequality containing a fraction |
| 3b | 2 | Numbers & indices | Expand with negative indices into the form a + bxⁿ |
| 3c | 5 | Equations & inequalities | Fractional equation reducing to a quadratic; formula, 2 d.p. |
| 3d(i) | 1 | Algebra & formulae | Make t the subject |
| 3d(ii) | 4 | Algebra & formulae | Eliminate t: y in terms of x as a single fraction |
| 4a(i) | 3 | Mensuration | Volume of a prism with a composite cross-section |
| 4a(ii) | 3 | Speed, rates & travel graphs | Time to fill at a constant flow rate |
| 4b(i) | 3 | Percentages | Successive percentage changes in circumference; % change in area |
| 4b(ii)(a) | 1 | Mensuration | Ratio of volumes of two cylinders of equal height |
| 4b(ii)(b) | 1 | Mensuration | Ratio of curved surface areas |
| 5a | 2 | Construction | Accurate construction of a quadrilateral from sides and angles |
| 5b(i) | 3 | Congruence & similarity | Prove two triangles congruent (AAS), reason for each statement |
| 5b(ii) | 2 | Angles & polygons | Angle proof using parallel lines, congruent triangles and an isosceles triangle |
| 6a(i) | 3 | Statistics | Pie chart: a frequency from angles and a difference |
| 6a(ii) | 3 | Statistics | Combined mean from a pie chart |
| 6b(i) | 2 | Matrices | Matrix product P × (D − C) |
| 6b(ii) | 1 | Matrices | Interpret the sum of the elements |
| 6c(i) | 1 | Probability | Probability of five blues in a row |
| 6c(ii) | 1 | Probability | 'At least one of each' by the complement |
| 7a | 2 | Pythagoras, trig & bearings | Area ½ab sin C leading to a quadratic |
| 7b | 3 | Equations & inequalities | Solve the quadratic by factorisation |
| 7c | 1 | Equations & inequalities | Reject a root in context |
| 7d | 5 | Pythagoras, trig & bearings | Cosine rule, then the obtuse angle |
| 8a(i) | 1 | Vectors | Vector between two points from position vectors |
| 8a(ii) | 2 | Vectors | Magnitude of a fraction of a vector |
| 8b(i) | 1 | Vectors | AB in terms of a and b |
| 8b(ii) | 3 | Vectors | Show three points are collinear |
| 8b(iii) | 1 | Vectors | Ratio along a line from the collinearity result |
| 9a | 1 | Real-world problem | Share of a unit cost |
| 9b | 2 | Real-world problem | Estimate a future population from a line graph |
| 9c | 7 | Real-world problem | Cost model, loan interest and a justified decision |