Pure Mathematics 2
Question Selector Guide
Provenance: written 2026-09-29, against research/veridian/WMA12-verified-facts.md (read in full, §§1–9 — specifically §2 for the spec's own verbatim worked-example question phrasing, §3 for the rubric and "students must answer all questions" wording, and §6/§7 for the per-topic real mark-range citations reused below), content/veridian/curriculum/wma12/paper-anatomy.ts (read in full, for the two independently-verified per-question mark sequences it carries), and four existing technique-hub documents for this paper — forge-exam-technique-guide.md, forge-exam-day-protocol-extended.md, signal-frequency-and-boundaries.md and forge-battle-card.md (all read in full) — plus codex-formula-and-method-bank.md (headers checked and its Hence-rule section read, to confirm it does not already cover question-selection logic). content/veridian/question-bank/wma12.ts was checked directly and holds no selector-guide content, so there is no duplication risk there. No new primary-source PDFs were opened for this pass: every specific factual claim below already exists, with its own citation, in one of the documents above, or is the spec's own verbatim wording as transcribed in the facts bank. What is new here is the synthesis — a topic-recognition table and a Selector Matrix that combine those existing real findings for a purpose none of the four sibling documents sets out to do: deciding, in the first seconds of reading a question, whether to attempt it now.
WMA12 — Question Selector Guide
Pure Mathematics 2 (Pearson Edexcel IAL) — what to read first, and what to do with it
What this document is, and isn't
This is not a fifth copy of the paper's basic shape, timing budget, or running-order advice — those already exist, fully cited, and are referenced by name below rather than restated:
content/veridian/curriculum/wma12/paper-anatomy.ts— the standing 90-minute/75-mark/M-A-B reference.forge-exam-technique-guide.mdPart 1 — the real timing budget and the three fully mark-mapped series.forge-exam-day-protocol-extended.mdParts 1–4 — the first-pass recommendation, the question-order table, the timing-checkpoint table, and the running-behind protocol.signal-frequency-and-boundaries.mdPart 1 — the per-topic real mark ranges this document's Selector Matrix reuses rather than re-derives.
What this document adds, genuinely new to this technique hub:
- Part 1 — a table for recognising a question's spec topic from its opening wording, before you have read the whole thing, built from the spec's own verbatim worked-example phrasing (facts bank §2) plus the real, named phrasing this course's own research has already confirmed on specific past questions.
- Part 2 — a Selector Matrix: each topic's real confirmed mark range set directly next to a real, individually-cited reason to attempt it now or handle it with care, rather than defaulting to "attempt in printed order."
- Part 3 — a triage procedure that assembles Parts 1 and 2 with the already-cited running-order and running-behind findings into an actual sequence of decisions, not a repeat of any single one of them.
- Part 4 — two cautions specific to selection logic that don't appear elsewhere in this technique hub: that question position is not a reliable stand-in for topic recognition, and that a question's own internal parts are not independently selectable the way whole questions are.
Every specific claim is traceable to a real source, cited inline. Where the evidence runs out, it is flagged as a gap, not filled in.
Part 1 — Spot the topic before you commit time to it
Confirmed, from the facts bank (§3): "Students must answer all questions" — there is no choice structure on this paper, so nothing here is about whether to attempt a question. It is about recognising, from the first line, roughly what kind of question you're looking at and what it typically costs, so the decisions in Parts 2–3 can be made in seconds rather than after a full read-through.
The cues below come from two places, kept distinct: the spec's own verbatim worked-example phrasing for that content area (facts bank §2, quoted under 15 words with attribution), and phrasing this course's own research has independently confirmed on a specific named real question.
| Spec topic (lesson file) | Recognise it by | Source |
|---|---|---|
Proof (proof.ts) | An explicit instruction to prove, disprove, or find a counter-example, often naming the method. The spec's own exhaustion example ends: "Prove that their sum is divisible by 2" (facts bank §2, spec 1.2). Its own counter-example example is built around the claim "n² − n + 1 is a prime number for all values of n" (facts bank §2, spec 1.3) — a claim you're asked to find one failing case for, not prove in general. | Spec P2.3, transcribed facts bank §2 |
Factor & remainder theorems (factor-and-remainder-theorems.ts) | "Show that (ax − b) is a factor of f(x)," "find the remainder when f(x) is divided by...," or a multi-part question ending "hence factorise f(x) fully." The word "hence" is itself a recognition cue and a mark-determining one — see Part 2. | Spec 2.1; real anchor Oct 2023 Q4 (facts bank, codex-formula-and-method-bank.md) |
The circle (the-circle.ts) | "The circle C has equation...," "find the equation of the circle with centre... passing through...," "show that PT is a tangent," or two points described as the endpoints of a diameter. | Spec 3.1; real anchors Jan 2021 Q9, Jun 2025 Q2 (signal-frequency-and-boundaries.md) |
Arithmetic sequences/series (sequences-and-arithmetic-series.ts) | "An arithmetic sequence/series," a stated nth-term formula, "the sum of the first n terms," or a recurrence of the form uₙ₊₁ = f(uₙ). | Spec 4.1–4.3 |
Geometric series (geometric-series.ts) | "A geometric sequence/series," "the sum to infinity," or a stated convergence condition on r. Often appears folded into a multi-part question rather than as its own standalone item — see Part 2. | Spec 4.4; signal-frequency-and-boundaries.md Part 1 |
Binomial expansion (binomial-expansion.ts) | "Expand (a + bx)ⁿ," "find the coefficient of xᵏ," or "in ascending powers of x." | Spec 4.5 |
Exponentials & logarithms (logarithms-and-exponentials.ts) | "Solve aˣ = b," any logₐ notation, "using the laws of logarithms," or "y = aˣ." | Spec 5.1–5.3 |
Trigonometric equations (trigonometric-equations.ts) | "Solve the equation ... in the interval ...," with the interval itself (degrees vs. radians) worth reading before touching a calculator's angle mode. | Spec 6.2; real anchors Oct 2022 Q5, Jun 2025 Q8(i) on the degrees-vs-radians point (forge-exam-technique-guide.md) |
Differentiation — stationary points/curve sketching (stationary-points-and-curve-sketching.ts) | "Find the coordinates of the stationary point(s)," "determine the nature of," or "show that the curve has no turning points." | Spec 7.1 |
Definite integrals & areas (definite-integrals-and-areas.ts) | "Find the area of the region R bounded by...," or a definite integral written out with limits already given. | Spec 8.1–8.2 |
Trapezium rule (the-trapezium-rule.ts) | "Use the trapezium rule with n = ... strips/ordinates," usually alongside a printed table of x- and y-values. | Spec 8.3 |
Caveat, stated honestly: these cues are not exhaustive, and a real question can mix topics inside one multi-part item — paper-anatomy.ts itself documents January 2024 Q5(i), which combines a periodic sequence with a sum-to-infinity calculation in the same question. Recognise the dominant topic for the triage decisions below, not a guaranteed single label.
Part 2 — The Selector Matrix: real mark range, next to a real reason to move now or take care
Mark ranges below are reused directly from signal-frequency-and-boundaries.md Part 1 — a floor on what has been individually confirmed by named citation across the 20-series archive, not a claim about every series' actual weighting.
| Topic | Real confirmed mark range | Real reason to move now, or to slow down |
|---|---|---|
| Proof | 2–4 marks, 6 named series | Confirmed as this paper's cheapest topic by mark count and, per forge-exam-day-protocol-extended.md Part 2, a genuine exception to attempting in printed order on a marks-per-minute basis — worth taking as soon as it's recognised, wherever it sits. |
| Factor & remainder theorems | 3–9 marks, ≥4 series | Care. The word "hence" is mark-determining, not stylistic: reaching a correct factorisation by a route other than the one the question names (e.g. solving directly on a calculator) is a confirmed zero on that part, not partial credit (Oct 2023 Q4(c)(ii), facts bank). Read the instruction before deciding this is a "quick" one. |
| The circle | 7–10 marks, ≥5 series | The highest confirmed floor of any single-lesson topic here, and a real, specific speed tool exists: the Jan 2021 Q9 examiner report calls a sketch "advisable for candidates" — a cheap, fast first action worth taking even if the algebra is parked for later. |
| Arithmetic sequences/series | ~3 marks for the standard sum-formula proof (1 fully tallied instance) | Both series' sum-formula proofs (pairwise-reversal for arithmetic) are fixed, rehearsed sequences of steps once recognised — a fast, low-variance target if the specific proof is familiar. |
| Geometric series | Usually embedded in a multi-part question, not isolated | Care. If an earlier part of the same question re-expresses the first term, confirm which value is the current first term before quoting the sum-to-infinity formula — Jan 2024 Q5(i)'s confirmed real slip was using the original first term instead of the re-expressed one. |
| Binomial expansion | 3–7 marks, 2 series / 3 papers | Two formulas share one booklet page; P2 needs the combinatorial (a+b)ⁿ form, not the general real-n series. A one-second check of which is needed prevents starting on the wrong one. |
| Exponentials & logarithms | 3–10 marks, 5 series | Care. Only the change-of-base formula is printed in the booklet — the addition, subtraction and power laws are not. If a log-law step doesn't come immediately from memory, that's a real signal this is a "slow down" question, not a "quick" one. |
| Trigonometric equations | 3–8 marks, 4 series | Care. A confirmed real trap (Oct 2023 Q3) is dividing out a common trig factor instead of factorising, which silently deletes an entire solution branch — the same structural error as dividing a quadratic through by x. Checking the interval's units (degrees vs. radians) before picking up the calculator has separately confirmed real cost when skipped. |
| Differentiation — stationary points | 4–11 marks, 4 series | The widest confirmed range of any topic here. The real dM1 substitution mechanism (forge-exam-technique-guide.md Part 2) means a partially-worked attempt can bank real marks even unfinished — a reasonable "start it, don't necessarily finish it under pressure" candidate. |
| Definite integrals & areas | 5–6 marks, 3 series | Where more than one valid method exists, Pearson's own examiners have confirmed one route is genuinely faster than another under exam conditions (Jun 2025 Q3: combining the two equations before integrating once, rather than integrating each separately) — recognising that choice point is itself a real time-saver, not just elegance. |
| Trapezium rule | 5–6 marks, 6 series (most-confirmed topic by series count) | A mechanically fixed procedure — strip width from the number of intervals, not the number of data points; double only the interior ordinates. Fast and low-risk once actively checked against the real, recurring point-count/gap-count miscount (Oct 2020, facts bank) rather than assumed away. |
Part 3 — A time-pressure triage procedure
Because there is no question choice on this paper (Part 1), triage here is entirely about when to attempt something and how much to commit before moving on — never about skipping. The steps below combine Parts 1–2 with findings already established elsewhere in this technique hub, cited rather than re-derived.
- Fast first pass. As
forge-exam-day-protocol-extended.mdPart 1 already recommends: turn every page once, note each question's printed marks, and — using Part 1's recognition cues above — its likely topic, before committing to Question 1. - Locate the heaviest question early. Confirmed across all three fully mark-mapped real series (
forge-exam-day-protocol-extended.mdPart 2): the single heaviest question sat in the paper's back half every time — Q9 in two series, Q10 in one, never the opener. Knowing roughly where and how large that question is before you reach it, rather than discovering it with ten minutes left, is the direct, practical use of that finding for selection purposes. - Bank the fast, well-recognised topics first if you're behind. Using the Selector Matrix, Proof and a recognised standard sum-formula proof are this paper's cheapest, most mechanical real marks. The paper has no rule against answering out of printed order — only against skipping a question entirely (facts bank §3) — so taking these out of sequence when short on time costs nothing structurally.
- On any question flagged "show all stages of your working" or "show detailed reasoning," resist the shortcut instinct hardest exactly when time pressure is highest.
forge-exam-day-protocol-extended.mdPart 4 already establishes that this trap becomes more tempting, not less relevant, once behind schedule, and that it is a confirmed zero rather than partial credit on two separate series. Worth naming precisely for selection purposes: October 2023 is the one series where this bold instruction and a real, named credited-shortcut trap are independently confirmed sitting on the same question — Q4, the factor/remainder "hence factorise" item (facts bank; codex-formula-and-method-bank.md). That is not a coincidence to read past; it is Pearson flagging, on the page, the exact question where a rushed candidate is most likely to reach for a calculator instead of the required method. (October 2023's Q3, the trigonometric-equation question, carries the same bold instruction but its own confirmed error — dividing out a common factor — is an algebraic slip, not a calculator-bypass one; the two questions share the instruction without sharing the same failure mode, and this document does not claim otherwise.) - Where a real speed comparison between valid methods exists, use it. Currently confirmed for one topic only: area-between-a-curve-and-a-line questions, where combining the two equations before integrating once is the examiner-confirmed faster route (Jun 2025 Q3, Part 2 above). No equivalent named speed comparison has been found for any other topic in this pass's search — this document does not invent one.
- On a stalled question, leave a partial, uncrossed-out attempt and move on. Already covered in full by
forge-exam-day-protocol-extended.mdPart 4, cited rather than repeated here: an attempt that is never crossed out is still markable exactly as it stands.
Part 4 — Two cautions specific to selection
- Position is not a reliable stand-in for either mark value or topic. Confirmed across three fully-mapped real series (
forge-exam-day-protocol-extended.mdPart 2) and directly relevant here: a candidate triaging by question number alone, rather than by the printed marks and Part 1's recognition cues, will misjudge which questions are actually worth the most time. - Within one question, later parts are not independently selectable the way whole questions are. A part that says "using your answer to part (a)" or "hence" is not a free-standing sub-question to attempt out of order — recomputing it from scratch instead of reusing the earlier part has a confirmed dual cost of marks and time on a real paper: the Jun 2025 Q5(b) examiner report states directly that redoing the earlier working "would also have been detrimental to their timing" (already cited in
forge-exam-day-protocol-extended.mdPart 4). The selection logic in Part 3 operates at the whole-question level; it does not license skipping around inside a single question's own dependent parts.
Flagged gaps — what this document does not claim
- No Pearson-published question-selection or question-order recommendation exists in any source reviewed for this course. This was established directly in
forge-exam-day-protocol-extended.mdPart 2 and is not re-verified independently in this pass, but nothing found here contradicts it either. - The Selector Matrix's mark ranges are floors, not weightings. They reuse
signal-frequency-and-boundaries.mdPart 1 exactly — a count of what has been individually confirmed by named citation, not a claim about every series' true topic-by-topic mark distribution. A given topic's real value on the paper in front of a candidate on the day could sit outside the observed range. - The "heaviest question sits in the back half" finding (Part 3, Step 2) rests on three fully-mapped series only — January 2021, January 2024, June 2025 — not a wider sample, per
forge-exam-day-protocol-extended.md's own stated scope. - Beyond the single June 2025 Q3 finding, no further examiner-report comment making an explicit, named speed comparison between two valid methods was found for any other topic in this pass's search of the facts bank and the existing technique-hub documents. This document does not manufacture a second one to make Part 2 or Part 3 look more complete than the real evidence supports.
- The specific topics of the questions carrying "show detailed reasoning" wording in January 2021 (Q10) and October 2024 (Q9) are not established anywhere in this course's research to date —
paper-anatomy.tsand the facts bank cite these by series and question number only, not by topic, and this document does not guess at them. - No likelihood, probability, or grade-boundary figure appears anywhere above. Every count is a "seen in N of M series checked" citation, reused from an already-verified source, per this course's standing discipline.
Sources used
research/veridian/WMA12-verified-facts.md— read in full, §§1–9; source of the spec's verbatim worked-example phrasing (§2), the rubric wording (§3), and the per-topic real findings reused throughout (§6–§7).content/veridian/curriculum/wma12/paper-anatomy.ts— read in full; source of the two independently-verified per-question mark sequences (Jan 2021, Jan 2024) and the M/A/B mechanism this document assumes rather than re-derives.content/veridian/technique-hub/wma12/forge-exam-technique-guide.md— read in full; source of the third fully-mapped series (Jun 2025) and the trap-taxonomy locator this document's Part 1/2 cross-references rather than duplicates.content/veridian/technique-hub/wma12/forge-exam-day-protocol-extended.md— read in full; source of the question-order table, the timing-checkpoint table, and the running-behind protocol this document cites throughout rather than restates.content/veridian/technique-hub/wma12/signal-frequency-and-boundaries.md— read in full; source of every real mark-range figure in the Selector Matrix (Part 2).content/veridian/technique-hub/wma12/forge-battle-card.md— read in full, to confirm its own condensed danger-warnings list does not already cover question-selection logic (it doesn't; it's a last-look formula/danger digest).content/veridian/technique-hub/wma12/codex-formula-and-method-bank.md— Hence-rule section read directly, to confirm the Oct 2023 Q4 citation used in Part 3, Step 4.content/veridian/question-bank/wma12.ts— checked directly; contains two real past-question analyses, neither overlapping this document's subject.