Pure Mathematics 4

Forty Eight Hour Checklist

Provenance: written 2026-09-29, against research/veridian/WMA14-verified-facts.md (read directly in full, all eight numbered sections plus the lesson-audit log — §1 spec contents, §2 assessment structure, §3 formula booklet vs. memorise split, §4 mark-scheme conventions, §5 per-topic findings, §6 series/ document breakdown, §7 lesson-readiness notes, §8 flagged issues), content/veridian/technique-hub/wma14/ forge-exam-technique-guide.md (read in full), content/veridian/technique-hub/wma14/codex-formula-and- method-bank.md (read in full), and content/veridian/technique-hub/wma14/signal-frequency-and-boundaries.md (read in full). The 11 lesson files in content/veridian/curriculum/wma14/ and content/veridian/question-bank/wma14.ts were checked by filename, header and trap-taxonomy locator rather than re-opened line by line, matching the sourcing approach the sibling WMA11/WMA12/WMA13/WBS paper checklists already use for this document type. No new primary-source PDF was fetched for this document — everything below is reframed from the four sources named above into a time-boxed countdown with explicit self-test items and an explicit "don't start this now" list; where a fact already has a home in the facts bank, FORGE, CODEX or SIGNAL, this document points at it by name and citation rather than re-stating it in full.

#WMA14 — 48-Hour Pre-Exam Checklist

Pearson Edexcel International A-Level, Unit P4: Pure Mathematics 4. 90 minutes, 75 marks, every question compulsory, calculator permitted throughout but barred from symbolic algebra, differentiation or integration, retrievable stored formulae, or communication with another machine (spec Appendix 6; facts bank §2). Formula booklet provided (Mathematical Formulae and Statistical Tables, Issue 2, January 2021 — "the Yellow booklet").

#What this document is, and isn't

This is not a content lesson and not a sixth copy of the exam-technique guide. It assumes you have already worked through the relevant lessons in content/veridian/curriculum/wma14/ and read (or at least skimmed) FORGE and CODEX before the last 48 hours. What this adds is a time-boxed triage: what has a documented payoff in the last two days, in what order, self-tested against this specific paper's own confirmed weak points — and just as concretely, what this paper's own real record says is a poor use of your remaining hours. Every claim below traces to the facts bank, FORGE, CODEX or SIGNAL by name and, where a specific question is cited, by series and question number. Nothing new is asserted here that a primary source doesn't establish, and nowhere below is a likelihood, a probability, or a grade-boundary figure invented — see "On grade boundaries" near the end for why that section is short.

#Two things to accept about this paper before the countdown starts

The paper's shape is not fully fixed, and marks are not spread evenly. Question count is confirmed at 9 (June 2022), 11 (October 2022) and 9 again (January 2026) — three real data points, not an exhaustive range (facts bank §2, §8). Real per-question tariffs checked directly run from 4 marks (a short proof item) to 16 marks (January 2026 Q7, a combined parametric/normal/area question — the largest confirmed single item in this paper's whole evidence base). 90 minutes ÷ 75 marks gives a flat 1.2 minutes per mark as an arithmetic planning tool, but the real spread between a 4-mark item and a 16-mark item means that figure is only a starting point — budget from each question's own printed tariff as you reach it, not from a remembered practice paper's shape (FORGE §1).
A question-specific rubric can override the paper's general calculator-allowed policy, and it has happened on this paper's own most recently confirmed series. WMA14 is calculator-allowed throughout — an ordinary scientific calculator, never a CAS (spec Appendix 6) — but January 2026 Q7 prints, twice, "Solutions relying on calculator technology are not acceptable," 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 (facts bank §5.6; FORGE §2). Read every question's own instruction line, not just the paper's cover-page rubric.

#The countdown

#T-48h to T-24h — the last full review session

Run one complete self-mark against the real M/A/B mechanics, not just the maths. Pick a past paper you've already attempted (or attempt one now) and re-mark your own work using the real mark-code rules — specifically the dependent-mark chains below and the "quote the formula first" rule — rather than only checking final answers.
Work through the cross-topic habits list and the per-topic self-test below once, in full, marking honestly which items you can answer from memory and which you can't.
If — and only if — you found a genuine, specific gap, spend the rest of this block closing that one gap, using the relevant curriculum lesson file or CODEX's matching method skeleton (§2.1–§2.10). Don't open a topic you weren't already shaky on going in — see "What not to start" below.
Don't attempt a second full past paper today if you already ran one. One honestly re-marked paper against the real mark-scheme mechanics is worth more here than two rushed ones marked only against final answers.

#T-24h to T-3h — the day before

Re-read the memorise-vs-booklet split once, then check yourself against the real list (facts bank §3; CODEX Part 1). The booklet's own P4 section prints the general binomial series for any rational $n$ , the integration-by-parts formula, and the $\csc x$ / $\sec x$ integrals — but the only item spec P4.2.3 flags as something you must know with nothing to fall back on is the component-form scalar product, $(x,y,z)\cdot(a,b,c)=xa+yb+zc$ . Everything else P4-specific is in the booklet in front of you. Check this against the real list, not a guess at what "feels" memorised.
Run the "confirmed repeat offenders" list below — the patterns with the deepest cross-series confirmation in this paper's evidence base, not a guess at what's "likely."
Confirm your calculator is on the permitted list: no symbolic algebra, no differentiation/integration facility, no stored retrievable formulae, no communication facility (spec Appendix 6). Check this tonight, not tomorrow morning.
Don't start a brand-new past paper this late in the day. A rushed, partial attempt the evening before produces more anxiety than signal; use the time on the self-test bank instead.

#T-3h to T-0 — exam morning

Ten-minute skim only, not new work: the "confirmed repeat offenders" list, plus whichever of your own T-48h self-test items you marked as shaky.
Physically check: the Yellow booklet, a permitted calculator, nothing that could look like a stored-formula device.
Remind yourself of three cross-topic rules that cost real marks regardless of topic: quote the formula/rule before you use it; keep exact fractions and surds until told a decimal is acceptable; on any "show that" question, never write the printed target down unless your own working genuinely reaches it.
Don't use this window to learn anything new. Everything in "What not to start" below applies with extra force this close to the exam.

#The last few minutes before you're told to begin

Read every question's own instruction line before starting it. January 2026 Q7 shows this paper will print a question-specific rubric override when it wants one — don't assume the cover page's general calculator policy is the only rule in play.
Count the real number of questions on the paper in front of you and note each one's printed tariff before committing to a pacing plan — this paper's own confirmed range is 9 to 11 questions, and you won't know which until you're looking at it.

#Cross-topic habits — self-test these first, they apply to every question

Real, mark-scheme-mechanical habits confirmed in the paper's own general marking guidance (stable across every series checked, January 2021 through June 2024) or across multiple independent series — not generic exam advice.

Quote the formula before you substitute into it. Verbatim from the general marking guidance: "the formula should be quoted first," because the method mark can then be awarded independently of what happens to the arithmetic afterward; where it isn't quoted, the same mark can still be earned by implication from correct working, "but may be lost if there is any mistake in the working" (facts bank §4). Worth doing on every question that opens with a named result: the binomial series, the scalar product, $\pi\int y^2\,dx$ , integration by parts, the cosine rule.
Treat a "dM1" as genuinely dependent, not a second independent mark. This paper's own richest marked-solutions run multi-stage dependent chains — proof by contradiction's M1→A1→dM1→A1* (June 2022 Q9); the foot-of-perpendicular anchor's M1→M1→A1→dM1 (October 2021 Q7); a second application of integration by parts is explicitly dependent on the first, not independent of it (June 2022 Q8(b); CODEX §2.7). A slip on the first stage of a chain like this costs every mark built on it — write each stage as its own visible line so there's something for an examiner to attach the dependent mark to.
Keep exact fractions and surds until told a decimal is acceptable. Verbatim: "marks will normally be lost if the candidate resorts to using rounded decimals" where an exact answer is required (facts bank §4). Confirmed directly costing marks on a binomial approximation question (January 2021, Q1).
On any "show that"/"hence show" question, never write the printed target down unless your own working genuinely reaches it. Three independently sourced lessons on three different spec points — proof by contradiction, parametric elimination, separable differential equations — each document a version of the same failure, tagged fudged-reverse-fit (FORGE §4). One honest caveat on this pattern's own evidence: all three of its named anchors come from the same series, January 2021 (Q3, Q4, Q10) — real evidence that the failure mode shows up three separate times within one sitting, not yet three independently-dated confirmations across different series. Treat the underlying mechanism as real and worth guarding against on every "show that" question regardless; don't overstate it as "confirmed every series."
A misread doesn't necessarily zero the whole question. The real October 2024 scheme caps a specific misread (reading a binomial index's sign wrong) at B1M1A0A0 rather than 0/6 — the method-shaped marks survive; only the marks depending on the misread quantity are lost (FORGE §2). If you spot a slip mid-question, finishing the method correctly from the wrong premise is still worth banking, not a reason to abandon the question.
Don't cross out an attempt unless you're certain you want it discarded. General marking guidance: if every attempt is crossed out, or none are, the examiner marks all of them and scores the best; only crossing out all-but-one restricts scoring to that single remaining attempt (facts bank §4). Leaving two genuine attempts uncrossed means the better one is scored either way.

#Confirmed repeat offenders — documented on two or more independent series

WMA14's evidence base is narrower than some sibling papers in this course: four full Examiners' Reports (January 2021, October 2021, October 2022, January 2024), plus mark-scheme-only verification for several further series. Most individual traps below trace to a single series' report. The items in this section are the ones that recur independently across more than one series — treat these as the highest- confidence entries on the whole list.

Wrong pair of vectors selected for a named angle, or the acute angle reported when the obtuse angle was needed. The single most robustly attested error pattern in this paper's whole evidence base — confirmed independently across three series: January 2021 Q2(a) ("many candidates gave the acute angle 67.35° as their final answer" when angle ABC, not BAC, was asked for), October 2022 Q3(b), and January 2024 Q6(c) ("use of incorrect direction vectors was fairly common, often using position vectors especially that of the point of intersection"). Self-test: for angle $ABC$ , are the two vectors you dot both starting from $B$ — $\overrightarrow{BA}$ and $\overrightarrow{BC}$ — never a position vector or a vector from the wrong endpoint?
The $\frac1a$ scaling factor dropped on a $\int\frac1{ax+b}\,dx$ integral. Confirmed independently a year apart: October 2022 Q2 ("a common error was to omit the ½ in $\int\frac1{2x-1}dx$ ") and January 2024 Q2(b) ("a number of students failed to divide the second logarithm by two"). Self-test: does the $\frac1a$ factor appear on the line where you first write the $\ln$ result, not as an afterthought?
A binomial expansion in $x^2$ split into two needless $(1\pm kx)^{n/2}$ brackets. Flagged as a live source of confusion in October 2021 Q4, and explicitly confirmed improved a year later — Pearson's own words, October 2022 Q4: "the idea of expanding a binomial expansion in $x^2$ did not cause the same issues as last year." Read this pairing as good news as much as a warning: the real record says this specific confusion is genuinely teachable and fixable with one clean self-check, not a deep-set issue worth heavy revision time. Self-test: for $(1-4x^2)^n$ , do you substitute $X=x^2$ directly into the general series, rather than factoring into two separate brackets?
Sign lost on the linear coefficient inside a binomial bracket. Confirmed independently three years apart: January 2021 Q1 ("a common incorrect 'x' term used was $5x/4$ ") and January 2024 Q1 ("errors... usually the result of using 'x' or '4x' in the place of ' $-4x$ '"). Self-test: after factoring $(ax+b)^n$ , does the sign of $\frac{a}{b}$ survive every subsequent term, not just the first?

A related, higher-level finding that is not a single repeated error but a persistently weak topic: proof by contradiction is the topic the examiner reports most consistently flag as weak across three independent series (January 2021: "candidates often omit questions on this topic or struggle to adopt a suitable strategy"; October 2021; January 2024: "many then simply commented that the equation formed had no solutions, giving no explanation as to why"). The pattern each time is the same: the algebra is usually fine, the logical scaffolding is what's missing — the explicit assumption sentence, the named reason a derived result is impossible, the closing sentence tying it back to the claim (facts bank §5.1, §7). Don't skip a proof-by-contradiction question because it "feels unfamiliar" — the setup step alone routinely banks a method mark before the harder logical writing is even needed.

#Per-topic self-test — one pass through every spec area

Short, specific self-checks, each tied to a real citation already established in the facts bank, FORGE or CODEX. If any item below is a genuine "no," go to the named lesson file to fix it — not this document.

Proof by contradiction (1.1) (proof-by-contradiction.ts): Have you stated the assumption as an explicit sentence, checked every derived statement is actually true (not just plausible-sounding — "n is an integer so n² + 1 is odd" is the documented false-but-plausible line to watch for, January 2021 Q3), and named the specific impossibility rather than just "this has no solutions" (January 2024 Q8)? If a question turns into checking a finite list of factor pairs, have you checked all of them, not one representative case (January 2026 Q8's own dM1 requires both pairs checked)?
Partial fractions (2.1) (partial-fractions-decomposition-and-integration.ts): Have you checked numerator degree against denominator degree before choosing a method — a top-heavy fraction needs division first? Have you matched the decomposition form to the real factor structure, including both terms for a repeated linear factor? Are you choosing substitution over full simultaneous equations where both are viable — Pearson's own examiners record substitution as less error-prone (January 2024, Q2a)?
Parametric conversion & domain/range (3.1) (parametric-equations-conversion-and-domain-range.ts): Have you completed a genuine elimination or coefficient-matching derivation, not assumed a target form and checked it against a couple of points (January 2021 Q4)? Does your domain/range method match the actual parameter interval — both endpoints for a closed interval, limiting behaviour for an open or unbounded one — and have you checked for an interior turning point before assuming the range sits at the two endpoints (October 2021 Q5(c): "the minimum value missing" because it sat inside the interval, not at an edge)?
Binomial expansion (4.1) (binomial-expansion-for-rational-n.ts): Sign and $x^2$ -collapse habits above. Have you substituted the required value into the original algebraic expression first, to confirm exactly what the series approximates — check whether a reciprocal is needed before assuming it matches directly (October 2024 Q1(b): the series approximated $6^{-1/3}$ ; the target was $\sqrt[3]6$ , its reciprocal, and submitting the raw series value cost two of the question's six marks)?
Implicit & parametric differentiation, tangents/normals (5.1) (implicit-and-parametric-differentiation-tangents-and-normals.ts): Are you evaluating a gradient at the point the question actually names — never the origin by default ("a surprisingly large number of candidates simply took P to be the origin," January 2021 Q6)? On a mixed term like $y^2x$ , are you applying the product rule and the chain rule together, the single hardest mechanical step documented for this topic (January 2024 Q3a)? Are you checking that an $x$ -only term never picks up a $dy/dx$ factor (October 2021 Q1: " $4x^2+8\to8x+8\,dy/dx$ ")?
Connected rates of change (5.2) (separable-differential-equations.ts): Have you substituted every related geometric quantity as a function of the differentiation variable before differentiating, rather than holding one fixed (January 2024 Q4: "many students treating $l$ ... as a constant")? Does the formula you've quoted actually fit the solid described — a confirmed real error applies the sphere formula to a cylinder (October 2021 Q9)?
Volume of revolution (6.1) (volume-of-revolution.ts): Have you quoted $V=\pi\int y^2\,dx$ first, squared $y$ fully including any cross term, and picked the right identity — power-reduction for a bare $\sin^2\theta$ / $\cos^2\theta$ , double-angle for a cross term from a squared sum (January 2021 Q9)? Before defaulting to parts or substitution, have you checked whether the integrand is a recognisable reverse-chain-rule form (October 2022 Q5: "wasted much time pursuing incorrect methods")? Is it $\pi$ , not $2\pi$ , unless the solid genuinely needs the doubled constant?
Integration by parts & substitution (6.2) (integration-by-parts-and-substitution.ts): Have you chosen $u$ as the factor that simplifies on differentiation? On a substitution, have you converted the entire integrand and the limits into the new variable — not relabelled $dx$ as $du$ directly (October 2021 Q6), and not reused the original $x$ -limits with the new integrand (October 2022 Q7(i))?
Integration using partial fractions (6.3) (partial-fractions-decomposition-and-integration.ts): $\frac1a$ -scaling habit above, on every term.
Separable differential equations (6.4) (separable-differential-equations.ts): Have you kept every constant on the correct side when separating (October 2021 Q2), and checked your final rearrangement for an illegal reciprocal-of-a-sum step ( $\frac1{p+q}\neq\frac1p+\frac1q$ , same question)? If a "show that" target doesn't match your working, have you gone back to find the real error rather than writing the target down anyway (January 2021 Q10)? Is your final answer in the units the question actually asked for (October 2021 Q9: seconds given instead of minutes)?
Area under a parametric curve (6.5) (area-under-a-parametric-curve.ts): Have you converted the $x$ -limits into $t$ -limits before integrating? If the bounded region includes a straight edge as well as the curve (a normal, a chord, an axis segment), have you split it into an elementary straight-edge component plus a parametric-integral component, rather than forcing the whole region into one integral (January 2026 Q7's own real structure)?
Position vectors, distance, foot of perpendicular (7.4–7.5) (position-vectors-distance-and-foot-of-perpendicular.ts): For a foot-of-perpendicular question, have you formed $\overrightarrow{AX}=X-\mathbf a$ from the line's general point, rather than dotted the general point $\mathbf c+\lambda\mathbf d$ directly against the direction vector (October 2021 Q7: "most incorrect responses attempted to set $(4-4\lambda,\ldots)\cdot(-4,-3,5)=0$ rather than $AX\cdot(-4,-3,5)=0$ ")? Have you treated finding the foot's coordinates as its own dependent step after solving for the parameter, not folded into the same mark?
Scalar product, angle, skew/parallel lines (7.6–7.7) (vectors-scalar-product-and-skew-lines.ts): Direction-vector habit above. To show two lines are skew, have you shown both non-intersection and non-parallelism — not substituted a non-perpendicularity check (January 2021 Q8)? To show two direction vectors aren't parallel, have you shown they aren't scalar multiples of each other, not just that they aren't identical (October 2022 Q9)?

#What NOT to start learning fresh in the last 48 hours

Each item below is something this course's own WMA14 research has explicitly flagged as excluded from the spec, unresolved even by primary-source checking, or genuinely thin on evidence — not a guess at what's unimportant.

Volume of revolution about the $y$ -axis, $\pi\int x^2\,dy$ . The spec's own guidance states plainly: " $\pi\int y^2\,dx$ is required, but not $\pi\int x^2\,dy$ " (facts bank §1). Don't spend revision time building or practising this variant — it is explicitly outside what this unit examines.
Partial fractions with a quadratic factor, $(x^2+a)$ , in the denominator. Spec guidance excludes this explicitly: "Quadratic factors in the denominator... are not required" (facts bank §1; CODEX Part 1, §2). If a practice resource includes this form for P4, it is testing content from a different unit.
A "deeper" proof of why $\pi\int y^2\,dx$ must be memorised versus looked up. Its status is genuinely unresolved by the spec and booklet text — it sits in neither the spec's own "must memorise" list (which is narrowly the dot-product formula) nor the booklet's P4 section, and no examiner report reviewed penalises not knowing the formula itself, only misapplying it after writing it down (facts bank §3, §8). Know it as a formula you'll need on the page; don't spend time resolving an ambiguity this course's own research couldn't resolve either.
Building the vector cross-product route for parallelogram area from scratch, if it isn't already familiar. The spec's own guidance names only the $\sin\theta$ route via the dot-product-derived angle; the cross-product alternative's real-exam credit rests on one 2026 mark scheme's own "Alt" note, not on spec or booklet text (CODEX §7, Flagged Gaps). If $\sin\theta=\dfrac{|\mathbf a\times\mathbf b|}{|\mathbf a||\mathbf b|}$ isn't a route you already use comfortably, the $\sin\theta$ -from-the-angle method alone is the spec-traceable one and is enough.
Treating any January 2026 item (Q3, Q7, Q8 or Q9) as if it carries a documented "common wrong answer." No examiner's report has been published for that series yet — checked directly, the same URL pattern that supplied the mark scheme and question paper 404s for a report (facts bank §5.6, §8; FORGE §4). These four items are genuinely useful for practising correct method flow against real, verified mark-scheme mechanics — but there is no confirmed candidate-misconception list for them yet. Don't build revision time around a guessed trap for these specific questions.
A "calculator technique" sub-skill. No examiner report reviewed for this paper discusses calculator use itself as a source of error — AO5 marks in the schemes reviewed are awarded through ordinary accuracy-to-3-significant-figures and "awrt" tolerances, not through a distinct calculator-specific question type (facts bank §8; FORGE §6). Inventing a calculator-technique revision block has no real basis in this paper's own examined content.
A brand-new full past paper less than roughly a day out, if you've already run one this week. A rushed, partially completed paper the evening before leaves no time to properly re-mark it against the real M/A/B mechanics that matter more here than the raw score — the self-test bank above is a better use of that window.

#On grade boundaries: there is nothing to calibrate against yet

Unlike some sibling papers in this course, no real Pearson grade-boundary figure for WMA14 has been verified anywhere in this course's research. WMA14-verified-facts.md was built from the specification, the formula booklet, mark schemes and examiner's reports — none of which publish grade boundaries, since those are released separately on Pearson's own results-day boundary documents. A direct search of the facts bank for boundary-related content found only mathematical domain/range boundaries inside worked questions, not grade-boundary marks (SIGNAL §2). Stating a number here without a primary Pearson grade-boundaries document behind it would break this course's own no-fabrication rule, so this checklist does not give you a target score, an A* mark, or anything resembling one. If you want a real figure, Pearson's own results-day boundary documents (searchable via qualifications.pearson.com, by series and paper code WMA14/01) are the only source that could genuinely supply one, and fetching and verifying those is flagged as open future work in this course's own research, not something already done.

#Flagged gaps in this checklist

The evidence base behind this checklist is genuinely narrower than some sibling papers'. Four full Examiners' Reports (January 2021, October 2021, October 2022, January 2024) underlie most of the traps above; several further series (June 2021, January 2023, June 2023, October 2023, June 2024) contributed only shared general-marking-guidance boilerplate, no topic-specific citations (SIGNAL §1). Most individual traps above therefore trace to a single series; only four items are independently confirmed across two or more series, and only one — the direction-vector confusion in angle-finding — across three. This checklist names that distinction explicitly rather than presenting every trap as equally well-attested.
No official per-topic mark weighting exists for this paper, only the assessment-objective ranges out of 75 (AO1 25–30, AO2 25–30, AO3 5–10, AO4 5–10, AO5 5–10 — facts bank §2). This checklist accordingly gives no "spend X minutes on topic Y" split; SIGNAL's own frequency table (§1) gives a real question's typical scale when tested, not a full-paper allocation, and this document inherits that limit rather than guessing a split.
Area under a parametric curve (6.5) and position vectors/foot-of-perpendicular (7.4–7.5) are each evidenced by only one series in this course's review (SIGNAL §1 flags this directly as a sourcing artefact, not evidence of rarity — both are compulsory, spec-listed skills). Don't read their thinner presence in the "confirmed repeat offenders" section above as license to deprioritise either topic.
No grade-boundary figure exists to calibrate against — see the section above; this is a genuine, disclosed gap in this course's WMA14 research, not an oversight in this specific document.
This document does not address exam-day wellness or logistics beyond equipment checks (venue, arrival timing, general test anxiety) — no document in this course's WMA14 research addresses that territory, and nothing here should be read as covering it.

#Sources used

research/veridian/WMA14-verified-facts.md (§1 spec contents; §2 assessment structure and calculator policy; §3 formula-booklet/memorise split; §4 mark-scheme conventions; §5.1–5.7 per-topic findings; §6 series/document breakdown; §8 flagged issues). content/veridian/technique-hub/wma14/forge-exam- technique-guide.md (§1 timing table; §2 mark-coding mechanics; §3 self-mark checklist; §4 common-error digest and the fudged-reverse-fit cross-topic finding; §5 exam-day protocol; §6 flagged gaps). content/veridian/technique-hub/wma14/codex-formula-and-method-bank.md (Part 1 GIVEN/MEMORISE masterlist; Part 2 method skeletons §2.0–§2.10; Flagged Gaps). content/veridian/technique-hub/wma14/ signal-frequency-and-boundaries.md (§1 topic-frequency table and its stated 7-series denominator; §2 grade-boundary gap; §3 flagged gaps). No new primary Pearson document was fetched for this file; no grade-boundary figure, likelihood percentage, or per-topic mark weighting is stated anywhere above beyond what those four sources already establish.

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.