Exam Technique Guide

Provenance: written 2026-09-28, grounded in research/veridian/WMA14-verified-facts.md (11 series with at least one primary document reviewed; 4 full Examiners' Reports — January 2021, October 2021, October 2022, January 2024) and cross-referenced against the 12 live lesson files under content/veridian/curriculum/wma14/, including paper-anatomy.ts, whose own verified M/A/B conventions this guide builds on rather than re-derives. No local WMA14 PDF archive exists on this machine — checked directly this session: ~/Downloads/PAPERI/IAL_Unit4_Complete_2024_2025/IAL_Materials/ has Business/ and Economics/ subfolders but no Mathematics/ one, and that archive's own Master_Mathematics_Unit4_{Past_Papers,Mark_Schemes,Examiner_Reports}.md stubs are each under 60 bytes — headers only, no content. Every citation below therefore traces through the verified-facts file's own independently-fetched primary sources (qualifications.pearson.com directly, or a third-party mirror whose cover-page metadata was independently checked against Pearson's own publication codes), not a local archive.

#WMA14 · FORGE — Exam Technique Guide

This is a technique and exam-craft layer, not a re-teach. The eleven content lessons in content/veridian/curriculum/wma14/ already teach the mathematics — proof structure, algebraic manipulation, the calculus, the vectors — in full depth, each with its own cited trap-taxonomy. This guide sits above that: how the paper is actually built, how Pearson's own M/A/B system actually pays out marks, what to check before you hand a script in, and where the errors already documented for this paper cluster once you look across all twelve lessons at once rather than one at a time.

Every specific claim below is traceable to a real series and question, cited in the format (Month Year, Qn). Where a number is a genuine estimate built from real data rather than a verbatim quote, that is stated explicitly.

---

#1. Paper Structure and a Real Timing Budget

paper-anatomy.ts already establishes WMA14's headline shape, verified against the spec (Issue 3, April 2019, p.26) and three real question papers: 75 marks, 90 minutes, every question compulsory, and — genuinely, not a simplification — the question count is not fixed. Real papers checked directly have run 9 questions (June 2022), 11 (October 2022), and 9 again (January 2026). That gives a flat average of 1.2 minutes per mark, and that page is right that the average is the wrong pacing tool: a question's own printed tariff is what matters, and this paper's own tariffs vary a lot more than "flat average" suggests.

#What a topic is actually worth, drawn from real questions

The table below is not a per-paper weighting — each row is a real question from a different series, not eleven rows from one 75-mark script. Treat it as evidence for a topic's typical scale (a short, self-contained item vs. a long, multi-technique one), not a guaranteed mark allocation for any specific sitting.

Two things this table makes concrete that the flat-average figure alone does not:

• The gap between the smallest and largest confirmed real questions is wide — a 4-mark proof item and a 16-mark combined parametric/normal/area item (Jan 2026 Q7) can sit on the same paper. A candidate who spends proof-question time on a 4-mark item as if it were worth 10 is giving away pacing; a candidate who treats a 16-mark combined question as a single unit rather than budgeting ~19 minutes across its four distinct parts is at real risk of running out of time on it alone.
• The paper's biggest confirmed real item deliberately chains spec points together — Jan 2026 Q7 needs parametric coordinates (3.1), a tangent/normal via parametric differentiation (5.1), and a parametric-curve area with a straight-edge component (6.5), inside one 16-mark question, with a rubric line banning calculator-only working printed twice. A candidate who has only drilled each spec point in isolation is not fully prepared for how this paper actually combines them at its largest scale.

Strategic reading, not a sourced claim: since question count and per-question tariff both vary, the first two minutes of reading time are worth spending on the paper's own printed tariffs before committing to a pacing plan — not recalling how a specific practice paper was shaped. This is direct inference from the verified "question count is not fixed" finding, not itself a quoted examiner instruction.

---

#2. The M/A/B Mark-Coding System — Playing It Correctly

paper-anatomy.ts already verifies the definitions (general marking guidance, stable Jan 2021 → Jun 2024): an M mark is for "knowing a method and attempting to apply it"; an A mark can only be awarded once its M mark is earned; a B mark is "unconditional... (independent of M marks)." What follows here is how to actually use that system to bank marks, drawn from real scheme lines across multiple lessons' own marked-solutions and audit logs — not a restatement of the definitions themselves.

Quote the formula first. This is not study advice — it is the scheme's own stated mechanism: "the formula should be quoted first," because a method mark can then be seen and awarded independently of whatever happens to the arithmetic afterward. Where a formula is not quoted, the same method mark can still be earned by implication from correct working, but is put at risk by any slip in that working. Worth doing on every question that opens with a named formula: the binomial series, the scalar product, $\pi\int y^2\,dx$, integration by parts, the cosine rule.

Dependent marks are common, and they are not one lump mark. Reading across this paper's own marked-solutions rather than one lesson at a time: proof by contradiction runs M1 → A1 → dM1 → A1* on its richest real anchor (Jun 2022 Q9); the position-vectors foot-of-perpendicular anchor runs M1 → M1 → A1 → dM1 (Oct 2021 Q7); the binomial anchor runs B1 → M1 → A1 → A1 (Oct 2024 Q1). A "dM1" mark is explicitly gated on the mark before it — for example, substituting a solved parameter back into a line's equation to get coordinates is its own dependent method mark, not folded into the mark for solving the parameter itself. The practical consequence: write every intermediate stage as its own visible line. A scheme built this way is rewarding the stages themselves, and an examiner working "best fit" against a script that shows only a final answer has nothing to attach the intermediate marks to.

"Show that" and "hence show" targets are graded on the path, not the destination. Any question with a printed target answer is testing whether every line is a genuine forward consequence of the one before it. Section 4 below names this pattern directly — it is common enough across this paper's proof, differential-equation, and parametric-elimination content that three separate lessons independently tag it. If your own working does not reach the printed target, the correct move is to go back and find the real error, not to write the target down anyway; examiners are specifically trained to catch a script that "adjusts" its own working to fit a given answer, and the final mark on an "ag" (answer given) line requires that every prior line was actually correct.

Exact answers stay exact until the question says otherwise. Verbatim: "marks will normally be lost if the candidate resorts to using rounded decimals" where an exact answer or surd working was required. This is a real, quotable, mark-costing habit independent of whether the underlying method was right.

A misread doesn't necessarily zero a whole question. The real scheme for Oct 2024 Q1 caps a specific misread (reading the binomial index's sign wrong) at B1M1A0A0 rather than 0/4 — the method-shaped marks survive, only the marks that depend on the misread quantity are capped. Worth knowing in the moment a slip is spotted mid-question: finishing the method correctly from the wrong premise is still worth banking, not a reason to abandon the question.

A rubric line on one question can override the paper's general calculator policy. WMA14 is calculator-allowed throughout — an ordinary scientific calculator, never a CAS — but a specific real question can still print its own, stricter ban: "Solutions relying on calculator technology are not acceptable," printed directly above Jan 2026 Q7, once, covering all four of its parts. A numerically-correct final answer with no algebraic working shown scores nothing on a question like that, regardless of the paper's general policy.

Leaving two attempts uncrossed is not automatically a penalty. The general marking guidance: if every attempt is crossed out, or none are, the examiner marks all the attempts and scores the best one; only when all but one attempt is crossed out does the mark scheme restrict itself to "the attempt which is NOT crossed out." The actionable version: never cross out a correct or stronger attempt while leaving a weaker one visible — if genuinely unsure which of two workings is better, leaving both is not a mistake, since the better one is what gets scored either way.

"It could be done in your head" is not a safe assumption. The rubric's own general policy: full credit assumes working is shown; a step that "could be done 'in your head'" is one of the few genuine exceptions, and mark schemes state explicitly where that applies. Anything else without working is a real risk, not a stylistic choice.

---

#3. Self-Mark Checklist

Work through the relevant block immediately after a practice paper or past question. Every item below is grounded either in a verified real mark-scheme requirement or a real, cited examiner-report finding — cross-referenced to the full lesson and trap-taxonomy item where one exists, not restated from it.

#Proof by contradiction (1.1)

• Stated the assumption as an explicit sentence, not just an equation — real scripts that skip this got the method mark but "full proofs with assumption, reason and conclusion were less common" (Jan 2021, Q3).
• Checked that every derived statement is actually true, not just plausible-sounding — the documented false assertion "n is an integer so n² + 1 is odd" is exactly the failure mode to check for.
• Named the intermediate result in words at each linked step ("...so p is even"), not just left in algebra (Oct 2021, Q10).
• If a fraction is assumed in lowest terms, stated that explicitly both where it's introduced and where the final contradiction needs it (Oct 2021, Q10).
• Gave the actual reason a derived equation is impossible (parity, a negative square, a used-up finite list), not just "this has no solutions" (Jan 2024, Q8).
• If exhausting a finite set of cases, checked every case, not one representative case (Jan 2026, Q8).
• Used a genuine algebraic derivation throughout, not a graph, a handful of tested values, or an unjustified calculator answer (Jan 2024, Q8).

See proof-by-contradiction.ts for the full worked technique — trap-taxonomy items assumption-and-conclusion-omitted, plausible-but-false-assertion-in-place-of-derivation, logical-step-skipped-before-the-conclusion-that-needs-it, lowest-terms-condition-not-stated, impossible-claim-asserted-without-a-reason, non-algebraic-check-substituted-for-a-required-contradiction, not-all-valid-factor-pairs-checked.

#Algebraic manipulation — partial fractions & binomial expansion (2.1, 4.1)

• Checked the degree of the numerator against the denominator before choosing a method — a top-heavy fraction needs division first.
• Used exactly the right power of a repeated linear factor in the denominator, not one power too many.
• Chose substitution over full simultaneous-equation expansion where both are viable — Pearson's own examiners record substitution as less error-prone (Jan 2024, Q2a).
• Included the $1/a$ scaling factor on every $\int \frac{1}{ax+b}\,dx$ integral — a two-series-confirmed slip (Oct 2022 Q2; Jan 2024 Q2b).
• Factored correctly to reach a clean $(1+kx)^n$ bracket before expanding.
• Kept the sign of the linear coefficient through every term of the expansion, especially with a negative bracket.
• Expanded an expression already in $x^2$ directly as one binomial in $x^2$, rather than splitting into two brackets.
• Substituted the required value into the original expression before deciding what the series approximates — check whether a reciprocal or other transform is needed, don't assume (Oct 2024, Q1b: "substitute x into the original expression").
• Gave an exact fraction/surd where asked, not a decimal.

See partial-fractions-decomposition-and-integration.ts and binomial-expansion-for-rational-n.ts — trap-taxonomy items extra-power-created-by-over-multiplying-a-repeated-factor, sign-lost-in-a-long-division-remainder, 1-over-a-scaling-factor-omitted-on-a-ln-integral, linear-coefficient-sign-lost, x-squared-term-collapsed-into-x, range-of-validity-not-derived-from-u, substituted-into-only-one-side-of-the-approximation, series-value-submitted-without-taking-the-reciprocal, decimal-given-when-exact-form-required.

#Parametric equations, differentiation, area & volume (3.1, 5.1, 6.1, 6.5)

• Eliminated the parameter through a genuine completed derivation — either substitution, or a coefficient-matching identity carried all the way through — not an assumed target form checked loosely against a couple of points.
• Matched the domain/range method to the actual parameter interval: closed-and-bounded (substitute both endpoints) vs. open/unbounded (reason about limits at the excluded/infinite ends).
• Differentiated every term with the correct rule, especially a mixed term needing product and chain rule together (e.g. $y^2x$) (Jan 2024, Q3a).
• Evaluated the gradient/tangent/normal at the point the question actually names, not a default point such as the origin (Jan 2021, Q6).
• Squared $y$ correctly for $\pi\int y^2\,dx$, including any cross term if $y$ is a compound or trig expression.
• Checked which trig identity a squared $y$ actually needs — the double-angle identity for a cross term from a sum/difference, or the separate power-reduction identity for a bare $\sin^2\theta$/$\cos^2\theta$ — these are two distinct identities, not one.
• Converted limits into the correct parameter values before integrating (solved $x=k$ for $\theta$ if needed), rather than reused the given interval unconverted.
• Used $\pi$, not $2\pi$, for a single revolution unless the question genuinely asks for $2\pi$.

See parametric-equations-conversion-and-domain-range.ts, implicit-and-parametric-differentiation-tangents-and-normals.ts, volume-of-revolution.ts, area-under-a-parametric-curve.ts — trap-taxonomy items guessed-target-form-instead-of-deriving-it, endpoint-substitution-attempted-on-an-open-or-unbounded-domain, domain-confused-with-range, partial-rearrangement-leaves-t-in-the-equation, calculus-reached-for-on-an-elimination-question, endpoints-substituted-for-the-whole-range, wrong-point-used-for-gradient-evaluation, product-rule-term-dropped-differentiating-a-mixed-xy-term, chain-rule-y-factor-dropped-differentiating-a-y-power-term, dy-dx-added-reflexively-to-a-term-with-no-y-at-all, reciprocal-error-simplifying-a-compound-parametric-fraction, wrong-constant-or-missing-pi, double-angle-identity-not-applied-when-squaring-y, power-reduction-identity-not-used-for-a-bare-sin-or-cos-squared, wrong-integration-method-chosen-for-the-resulting-fraction, compound-expression-not-squared-correctly, x-limits-used-directly-as-t-limits-without-converting, dx-dt-factor-dropped-entirely, area-formula-confused-with-the-adjacent-volume-formula.

#Integration technique & differential equations (5.2, 6.2, 6.3, 6.4)

• Quoted the integration-by-parts formula and identified $u$ and $dv/dx$ before starting.
• Recognised that a second application of parts is a dependent method mark, not independent — a slip on the first application costs both.
• On substitution, converted $dx$ into the new variable's differential correctly, not relabelled it $du$ directly.
• Converted the limits of integration into the new variable — didn't reuse the original $x$-limits.
• Checked whether the integrand is a recognisable reverse-chain-rule form before defaulting to parts or partial fractions out of habit (Oct 2022, Q5).
• Kept every constant on the correct side/denominator when separating variables — didn't "move" one incorrectly mid-rearrangement.
• Made the target variable the subject correctly at the end — no illegal reciprocal distribution across a sum ($\frac1{p+q}\neq\frac1p+\frac1q$).
• If a "show that" target is printed and the working doesn't reach it, went back to find the real error rather than writing the printed target down anyway.
• In a modelling context, checked the geometric formula actually fits the shape described, and substituted every related quantity as a function of the differentiation variable before differentiating — not held it fixed (Jan 2024, Q4).
• Gave the final answer in the units the question actually asked for, not the units the working happened to be in (Oct 2021, Q9).

See integration-by-parts-and-substitution.ts, separable-differential-equations.ts — trap-taxonomy items coefficient-and-differentiation-slips-in-a-second-parts-application, inverse-function-misapplied-to-a-substitution-term, dx-relabelled-as-du-without-being-converted, limits-of-integration-not-converted-to-the-new-variable, a-numerical-factor-lost-within-the-substituted-expression, ln-x-squared-treated-as-interchangeable-with-2-ln-x-mid-derivation, sphere-formula-used-for-a-cylinder, two-term-rate-collapsed-to-one, constant-integrated-as-a-logarithm, answer-given-in-the-wrong-time-unit, denominator-mishandled-when-separating, fudged-reverse-fit-on-a-show-that-answer, related-quantity-treated-as-fixed-instead-of-substituted.

#Vectors (7.1–7.7)

• Used the correct pair of vectors sharing the correct vertex for the specific angle named (e.g. angle ABC needs BA and BC, not AC and BC) — three-series-confirmed (Jan 2021, Oct 2022, Jan 2024).
• Checked whether the question needs the acute or obtuse angle, rather than defaulting to the acute value $\arccos$ returns.
• Confirmed the scalar (dot) product output is a single number, not confused with a vector (cross) product output.
• To show two lines are skew, shown both non-intersection and non-parallelism — not substituted a non-perpendicularity check (Jan 2021, Q8).
• To show two direction vectors aren't parallel, shown they aren't scalar multiples of each other — not just that they aren't identical (Oct 2022, Q9).
• For a foot-of-perpendicular question, formed the vector from the external point to the line's general point (e.g. AX), rather than dotted the general point directly against the direction vector (Oct 2021, Q7).
• Treated finding the foot's coordinates as its own dependent step after solving for the parameter — not folded into the same mark.
• Included the final square root when calculating a magnitude.

See vectors-scalar-product-and-skew-lines.ts, position-vectors-distance-and-foot-of-perpendicular.ts — trap-taxonomy items wrong-vectors-selected-for-the-angle-calculation, acute-angle-reported-when-obtuse-required, scalar-vs-vector-product-confusion, skew-shown-via-not-perpendicular-instead-of-not-parallel, non-parallel-shown-via-non-identical-not-non-scalar-multiple, general-point-substituted-for-ax, ab-formed-in-the-wrong-order, magnitude-missing-the-final-square-root, unit-vector-not-actually-normalised.

---

#4. Common-Error Digest

The 12 curriculum lesson files each already carry their own trap-taxonomy, built from real examiner-report and mark-scheme findings. Cross-checking every specific error cited anywhere in WMA14-verified-facts.md against all 12 files' own trap-taxonomy items found no genuinely uncovered candidate-error pattern — every documented mistake from the four full examiner reports reviewed (Jan 2021, Oct 2021, Oct 2022, Jan 2024) is already named somewhere in the topic lessons. That is itself a real, checkable finding worth stating plainly rather than manufacturing a "new" trap to satisfy a quota. What genuinely is new here, because it only becomes visible once you read across all twelve files at once:

#The "fudged-reverse-fit" pattern is the single most repeated meta-trap on this paper

Three independent lessons — spanning three different spec points — each carry their own, separately-sourced example of the same underlying failure, tagged with the shared identifier fudged-reverse-fit:

• Proof by contradiction (plausible-but-false-assertion-in-place-of-derivation): a candidate writes a sentence that reads like a derivation but states something untrue, landing on the shape of a proof without the substance (Jan 2021, Q3).
• Parametric elimination (guessed-target-form-instead-of-deriving-it): a candidate assumes the target cartesian form and checks it loosely against the parametric equations, rather than completing a genuine coefficient-matching derivation (Jan 2021, Q4).
• Separable differential equations (fudged-reverse-fit-on-a-show-that-answer): a candidate, working toward a printed target answer, "adjusts" intermediate steps following an earlier mistake to land on the given form anyway (Jan 2021, Q10).

Three different topics, three different series/questions, one shared mechanism: any question with a printed target answer is graded on whether every line is a genuine consequence of the one before it, not on whether the final line matches. This is the single highest-leverage habit to check on any "show that" or "hence show" instruction anywhere on this paper — the vectors lessons were checked for the same pattern during their own audits and no genuine on-topic instance was found, which is itself informative: this is a proof/elimination/DE-family trap on the current evidence, not a universal one.

#A second, less formally cross-referenced pattern: the wrong reference point substituted

Two lessons, on different spec points, each document a candidate substituting the wrong geometric object where a difference vector was required:

• Implicit differentiation (wrong-point-used-for-gradient-evaluation): evaluating a gradient at the origin by default, rather than the point the question actually specifies (Jan 2021, Q6).
• Position vectors (general-point-substituted-for-ax): dotting a line's general point $c+\lambda d$ directly against the direction vector, instead of first forming $AX$ (the vector from the external point to that general point) (Oct 2021, Q7).

This is this guide's own synthesis, not a shared tag already in the source files — but the underlying instinct is the same in both: reaching for a convenient, nearby object (the origin; the line's own defining point) instead of the specific object the question's geometry actually requires.

#Real, new-format items with mechanics verified but no documented error yet

Two of the richest real items found in this paper's most recent reviewed series carry a genuine, disclosed limitation: the January 2026 series has a verified mark scheme and question paper, but no examiner's report has been published for it yet (checked directly — the same URL pattern used for the mark scheme and question paper 404s for a report). These items are in the lessons' own worked/marked-solution blocks, verifying real scheme mechanics, but are deliberately not added to any trap-taxonomy, because no documented candidate misconception exists for them yet:

• Jan 2026 Q7 — area under a parametric curve, bounded partly by the curve and partly by a normal line (16 marks, the largest confirmed real question in this whole facts bank).
• Jan 2026 Q9 — parallelogram/trapezium area from vectors, using either $\sin\theta$ or the component cross-product (11 marks).

The honest technique implication: for these two specific question-shapes, there is no examiner-report-backed "here is what goes wrong" yet — treat the self-mark checklist's general discipline (dependent-mark chains, exact answers, "show that" honesty) as the available defence, not a topic-specific trap list, until a report is published and this guide can be updated against it.

#Cross-reference table

---

#5. Exam-Day Protocol

Short, and every line below is grounded in a verified real convention rather than generic exam advice.

1. Check your desk before you start. The Mathematical Formulae and Statistical Tables (Yellow) booklet is provided; bring an ordinary scientific calculator — Appendix 6 bars any calculator with symbolic algebra, differentiation, or integration built in, and bars communication with another machine.
2. Spend the first two minutes counting real questions and reading every printed tariff. Real papers have run 9 to 11 questions; recalibrate against the paper in front of you, not a remembered practice paper's shape.
3. Watch for a question-specific rubric override on top of the general calculator policy. A real WMA14 question has printed, twice, that calculator-only working earns nothing on it — even on a calculator-allowed paper.
4. Quote the formula first, every time one applies. It is the scheme's own stated safety net for the method mark, not a stylistic preference.
5. Attempt every proof-by-contradiction item, even if the topic feels unfamiliar. The general commentary records this as the topic candidates most often skip — and the assumption/setup step alone routinely banks a method mark before any of the harder logical scaffolding is needed.
6. On any "show that" or "hence show" question, never write the printed target down unless your own working genuinely reaches it. If it doesn't, go back and find the real error — this is the single most repeated documented failure across this paper's proof, parametric, and differential-equation content (Section 4 above).
7. Keep exact values through the whole solution unless a decimal is explicitly requested — converting early costs marks even when the final rounded value would have matched.
8. Write every dependent stage of a multi-mark derivation as its own visible line. This paper's real schemes routinely split a single derivation into two, three, or four dependent marks rather than one lump mark — each stage shown is a mark genuinely available to claim.
9. Sanity-check units and magnitude at the end of any modelling question (a tank, a cone, a growth model) before moving on — a numerically-correct answer in the wrong unit has cost real marks on this paper (Oct 2021, Q9).
10. Don't cross out an attempt unless you're certain you want it discarded. Leaving two genuine attempts uncrossed means the better one is scored either way; only crossing out all-but-one restricts the examiner to that single remaining attempt.

---

#6. Flagged Gaps

Stated honestly rather than papered over, matching the verified-facts file's own discipline:

• No local WMA14 PDF archive exists on this machine. Confirmed directly this session — every citation above traces through research/veridian/WMA14-verified-facts.md's own independently-verified live fetches, not a local archive copy.
• The evidence base is 11 series with at least one document, and only 4 full Examiners' Reports (Jan 2021, Oct 2021, Oct 2022, Jan 2024). Any "confirmed across N series" claim in this guide is scoped to that count — narrower than this course's Economics/Business precedent, and should not be read as an exhaustive survey of every WMA14 series ever set.
• The January 2026 series has no published examiner's report yet. The four items drawn from it in this guide (Q3 differential equation, Q7 area under a parametric curve, Q8 proof, Q9 vectors area) verify real scheme mechanics, not documented candidate errors — Section 4 above states this explicitly for the two richest of them.
• The real per-question mark tariffs in Section 1 are drawn from different series' papers, not one single sitting. They illustrate a topic's typical scale; they are not a claim that all these tariffs appear together on any one real script.
• No calculator-specific technique section is included, because no examiner report reviewed for this paper documents calculator use itself as a source of error — inventing one would have no real basis, consistent with the verified-facts file's own explicit caution on this point.
• The question-count range (9–11) rests on three confirmed data points (June 2022, October 2022, January 2026), not an exhaustive survey — treat it as the confirmed range across the series actually reviewed, not a guaranteed ceiling.

---

#Sources Used

Primary Pearson documents (all independently verified in research/veridian/WMA14-verified-facts.md, converted with pdftotext -layout and cross-checked against -raw, and, where the source was a third-party mirror rather than qualifications.pearson.com directly, checked against the PDF's own cover-page publication code):

• Specification, Issue 3, April 2019 (pp.26–29, 68–69, 85–86).
• Mathematical Formulae and Statistical Tables, Issue 2, January 2021.
• Mark schemes: January 2021, June 2021 (provisional), June 2022, January 2023, June 2023, October 2023, January 2024, June 2024, October 2024 (Q1, Q2), January 2026 (Q3, Q7, Q8, Q9).
• Examiners' Reports (full): January 2021, October 2021, October 2022, January 2024. Examiner's Report (single-question extract): October 2024.
• Question papers: June 2022, October 2022, January 2026.

Real questions cited by series and number in this guide: Jan 2021 Q3, Q4, Q6, Q8, Q9; Oct 2021 Q1, Q2, Q5(c), Q6, Q7, Q8(b), Q9, Q10; Jun 2022 Q8(b), Q9; Oct 2022 Q1, Q2, Q3(b), Q4, Q5, Q6, Q7(i)(ii), Q9, Q11(a); Jan 2024 Q1, Q2(a)(b), Q3(a), Q4, Q5(a), Q6, Q7(b), Q8; Oct 2024 Q1, Q2; Jan 2026 Q3, Q7, Q8, Q9.

Course files cross-referenced (not duplicated): all 12 files under content/veridian/curriculum/wma14/ — paper-anatomy.ts, proof-by-contradiction.ts, partial-fractions-decomposition-and-integration.ts, parametric-equations-conversion-and-domain-range.ts, binomial-expansion-for-rational-n.ts, implicit-and-parametric-differentiation-tangents-and-normals.ts, volume-of-revolution.ts, integration-by-parts-and-substitution.ts, separable-differential-equations.ts, area-under-a-parametric-curve.ts, position-vectors-distance-and-foot-of-perpendicular.ts, vectors-scalar-product-and-skew-lines.ts.

Not affiliated with or endorsed by Pearson Edexcel. Every quoted fragment above is under fifteen words with its own series/question attribution; all longer commentary is this course's own, built on — not reproduced from — the cited primary documents.