Deep Dive: Binomial Expansion for Rational N

Provenance — 2026-09-29: grounded in research/veridian/WMA14-verified-facts.md §1 (spec text), §3 (formula-booklet contents), §4 (mark-scheme conventions), §5.4 (the topic's own verified errors), §6–8 (sourcing/coverage caveats); cross-checked against the live lesson content/veridian/curriculum/wma14/binomial-expansion-for-rational-n.ts, the spec-4.1 block inside partial-fractions-decomposition-and-integration.ts, and this paper's own codex-formula-and-method-bank.md §4/§2.4, forge-exam-technique-guide.md §1–3, and signal-frequency-and-boundaries.md §1 — to confirm every section below adds material not already published in those five places rather than restating them. No new primary Pearson document was fetched this session; every citation traces to sourcing already independently verified in the facts bank.

#CODEX Deep-Dive · Binomial Expansion for Rational n (WMA14, spec 4.1)

This is a single-topic reference, not a re-teach. The full worked chains, the method-comparison, the chain-drill, both marked-solutions and the MCQ set already live in binomial-expansion-for-rational-n.ts — none of that is reproduced here. What follows is (1) the complete theoretical model behind the one formula this spec point tests, including the one piece of that model no other WMA14 document states — where the coefficient formula itself actually comes from; (2) the three genuinely distinct shapes spec 4.1 requires, illustrated with short original micro-examples different from every number already used elsewhere in this course; (3) a notation hazard worth naming exactly once; (4) the real mark-scheme architecture on this topic's one fully-verified past-paper anchor; and (5) a citation-by-citation synthesis of every real exam angle this topic has actually been tested from, read across all five series the facts bank has evidence for at once, rather than one series at a time.

#1. Spec anchor

Spec section 4 ("Binomial expansion") has exactly one numbered point, 4.1, and no lettered sub-parts — the smallest single-point content area in the whole P4 unit. Its guidance names two required skills: deriving the expansion of $(ax+b)^n$ directly, valid "for |x| < b/a" (Pearson's own spec wording, quoted exactly), and separately, expanding a rational function after first decomposing it into partial fractions. The general series itself — $(1+x)^n = 1+nx+\frac{n(n-1)}{2!}x^2+\cdots$, valid for $|x|<1,\ n\in\mathbb R$ — is printed in the exam formula booklet's own P4 section and is not something a candidate has to memorise from scratch.

#2. The formal model

#2.1 The general series

$$(1+x)^n = 1 + nx + \frac{n(n-1)}{2!}x^2 + \frac{n(n-1)(n-2)}{3!}x^3 + \cdots + \frac{n(n-1)\cdots(n-r+1)}{r!}x^r + \cdots, \qquad |x|<1,\ n\in\mathbb{R}.$$

#2.2 Where that coefficient actually comes from

No WMA14 document in this course currently derives this formula — every one of them states it as booklet-given and moves straight to application, which is the correct exam-facing priority, since the spec never asks a candidate to derive it. But a genuine reference document on this topic is incomplete without stating where it comes from, so here it is, clearly flagged as standard calculus background rather than spec- or exam-board-sourced content: build $f(x)=(1+x)^n$'s own Maclaurin series — its Taylor series about $x=0$ — by repeated differentiation. $f(0)=1$. $f'(x)=n(1+x)^{n-1}$, so $f'(0)=n$. $f''(x)=n(n-1)(1+x)^{n-2}$, so $f''(0)=n(n-1)$. Each further differentiation brings down one more factor and drops the power by one, so the $r$-th derivative at $0$ is $f^{(r)}(0)=n(n-1)(n-2)\cdots(n-r+1)$ — exactly $r$ factors, counting down from $n$. Maclaurin's theorem says the coefficient of $x^r$ in the series is $f^{(r)}(0)/r!$, which is precisely the $\frac{n(n-1)\cdots(n-r+1)}{r!}$ the booklet prints. The formula is not an arbitrary pattern to memorise; it is what any smooth function's own derivatives at a single point produce, applied to this one function.

#2.3 Why the series is infinite unless n is a non-negative integer

Briefly, since the lesson's own mechanism block covers this in full: the coefficient's running product $n(n-1)\cdots(n-r+1)$ contains a factor that equals exactly zero once $r$ climbs past $n$ — but only when $n$ itself is a non-negative integer that the count-down actually lands on. For $n$ a genuine fraction or a negative number, the count-down $n,n-1,n-2,\ldots$ never touches zero, no factor in the product is ever zero, and no term of the series is ever exactly zero — hence "any rational $n$" genuinely produces an infinite series where the P2 finite case does not. (Full treatment, including the P2 contrast: binomial-expansion-for-rational-n.ts, "Why the series never terminates.")

#2.4 The convergence boundary, and what the spec does and doesn't ask for

$|x|<1$ is the condition under which the infinite sum above actually settles on a single finite value equal to $(1+x)^n$, rather than diverging. The spec does not require a candidate to justify this boundary — it is given, exactly like the formula itself — but the reason it sits at exactly $|x|=1$ is worth stating once for completeness: it is the boundary at which $x^r$ stops shrinking to zero as $r\to\infty$ and starts growing instead, which is what a rigorous convergence argument (the ratio test, in full generality) would formalise. This is genuinely beyond what WMA14 assesses; it is included here only so this reference states the complete real model, not because a candidate needs it.

#3. The three shapes spec 4.1 actually requires

Spec 4.1's own guidance (§1 above) names two skills; the existing lesson content splits the first of those into two recognisably different bracket-shapes. Together that is three genuinely distinct forms a WMA14 candidate must recognise on sight. Each is illustrated below with a short original micro-example — new numbers, not reused from the lesson, its MCQs, or the partial-fractions lesson's own worked chain — solving no more than two or three terms, since the point here is recognition of the shape, not a full past-paper-length derivation (those already exist, cited below each one).

#3.1 Direct: $(a+bx)^n$

Factor the constant out of the whole bracket so it reads $k(1+u)^n$, then apply §2.1 term by term, scaling every term (not just the first) by $k$ at the end.

Micro-example. $(4+2x)^{-\frac12} = 4^{-\frac12}\left(1+\frac12x\right)^{-\frac12} = \frac12\left[1+\left(-\frac12\right)\left(\frac12x\right)+\frac{\left(-\frac12\right)\left(-\frac32\right)}{2}\left(\frac12x\right)^2+\cdots\right] = \frac12-\frac18x+\frac{3}{64}x^2-\cdots$, valid for $\left|\frac12x\right|<1$, i.e. $|x|<2$. (Independently cross-checked by differentiating $(4+2x)^{-\frac12}$ directly and evaluating at $x=0$: $f(0)=\frac12$, $f'(0)=-\frac18$, $f''(0)/2!=\frac{3}{64}$ — all three match.)

Full worked chains of this exact shape, with real examiner-report-grounded error commentary: binomial-expansion-for-rational-n.ts's $(9-6x)^{-\frac12}$ chain and its two marked-solution blocks.

#3.2 A single power already inside the bracket (binomial in X)

Where the bracket is already a binomial in some power of $x$ (commonly $x^2$), the whole of that power substitutes directly for $u$ — the general formula has no opinion about what algebraic shape $u$ is, only that $|u|<1$.

Micro-example. $(1+2x^2)^{-\frac12} = 1+\left(-\frac12\right)\left(2x^2\right)+\cdots = 1-x^2+\cdots$, valid for $\left|2x^2\right|<1$, i.e. $|x|<\frac{1}{\sqrt2}$. (Cross-checked: $f(x)=(1+2x^2)^{-\frac12}$ gives $f(0)=1$, $f'(0)=0$ — correctly no $x^1$ term, since $u$ is even-powered — and $f''(0)/2!=-1$, matching the $-x^2$ coefficient.) The documented failure mode on this exact shape is treating the power inside the bracket as though it weren't there — collapsing $4x^2$ to $4x$ — a genuinely fixable error: one examiner report records it happening, a later one on a structurally identical question records it not recurring (Oct 2021, Q4; Oct 2022, Q4 — full quotes in §6 below).

Full method-comparison, including the harder-but-valid difference-of-squares alternative for this exact shape: binomial-expansion-for-rational-n.ts's $(1-9x^2)^{\frac13}$ block; MCQ3 in the same lesson drills the identical recognition point with $(1-5x^2)^{\frac14}$.

#3.3 A rational function, decomposed into partial fractions first

Spec 4.1's own guidance names this as a required skill, separate from 3.1–3.2 above: decompose the rational function into partial fractions first (spec 2.1's technique), then expand each resulting simple fraction as its own series, each brought into $k(1+u)^n$ shape independently. One point genuinely new at this stage, absent when a single bracket is expanded alone: the result is a sum of separate series, and the sum is only valid where every one of them individually converges — the combined range is the most restrictive of the individual ranges, not their average or their union.

Micro-example (deliberately simpler than the course's own full worked chain, to isolate just this one point): given the already-decomposed sum $\dfrac{1}{1-x}+\dfrac{1}{1+2x}$, expand each term — $(1-x)^{-1}=1+x+x^2+\cdots$ (valid $|x|<1$) and $(1+2x)^{-1}=1-2x+4x^2-\cdots$ (valid $|2x|<1$, i.e. $|x|<\frac12$) — and add: $2-x+5x^2+\cdots$, valid for the intersection of the two ranges, $|x|<\frac12$, not $|x|<1$.

Full worked chain, with a genuine three-linear-factor-and-a-repeated-factor decomposition (including where each fraction's own range comes from) and the same intersection reasoning applied at full scale: partial-fractions-decomposition-and-integration.ts's spec-4.1 block, expanding $\dfrac{2}{3x+1}+\dfrac{3}{x-2}-\dfrac{1}{(x-2)^2}$. That lesson's own header is explicit that the "combined range is the intersection" reasoning is course-derived mathematical reasoning, not a sourced examiner-report finding — no citation in the facts bank documents a candidate being tested specifically on this point — and this document repeats that same honest labelling rather than upgrading the claim's evidence tier.

#4. A notation hazard worth naming once

Three equivalent presentations of the same one formula use different letters for the same roles, and pattern-matching the wrong pair of letters against the wrong range formula is a real, avoidable error:

All three describe the identical fact; only which letter plays which role changes. Reading a bound off a memorised shape without checking which convention the question in front of you is actually using is a real way to invert or misapply the ratio — worth a five-second check before writing a final range down, exactly the derive-it-from-$u$-directly discipline the lesson's own trap-taxonomy item range-of-validity-not-derived-from-u is built around.

#5. Mark-scheme architecture, on the one real fully-verified past-paper anchor

WMA14/01 October 2024 Q1 (Publications Code WMA14_01_2410_MS/ER) is this topic's only item in the facts bank with a complete, independently re-verified real mark scheme: $f(x)=(8-3x)^{-\frac13}$, expand to the $x^3$ term (4 marks), then use $x=\frac23$ to estimate $\sqrt[3]{6}$ (2 marks). The scheme's own mark-code chain is B1 → M1 → A1 → A1 for part (a), then M1 → A1 for part (b) — six marks, none of them a single lump sum:

• B1: $8^{-\frac13}$ actually evaluated to $\frac12$ (not left symbolic, not decimalised).
• M1: the general expansion attempted to an unsimplified 3rd/4th term, correct structure for $n=-\frac13$.
• A1: any two of the three non-constant terms correctly simplified — genuine partial credit, independent of the third.
• A1: the complete four-term expansion, $\frac12+\frac{1}{16}x+\frac{1}{64}x^2+\frac{7}{1536}x^3$.
• M1 (part b): substitutes $x=\frac23$ and attempts the reciprocal of the resulting fraction — the reciprocal step is its own separately-earnable method mark, not automatic.
• A1 (part b): $\frac{5184}{2851}$ (equivalently $1\frac{2333}{2851}$).

Two mark-scheme mechanics on this exact anchor are easy to miss reading the lesson's marked-solution alone, but are real, quoted scheme text worth stating plainly here: both final A marks carry an explicit isw ("ignore subsequent working following a correct answer") instruction — once the correct four-term expansion or the correct $\frac{5184}{2851}$ is written down, further wrong simplification underneath it does not retract the mark. And the scheme carries a named misread cap: reading the index's sign as $+\frac13$ instead of $-\frac13$ scores a maximum of B1M1A0A0 for part (a) and M0A0 for part (b) — the method-shaped marks survive a sign misread; only the marks that depend on the misread quantity's value are lost. This is the single clearest real illustration, anywhere in this topic's evidence, of the general Pure-maths marking principle that a misread caps a question rather than zeroing it outright.

#6. Every real exam angle this topic has actually been tested from

Reading the facts bank's five sourced series for this one spec point together, rather than one at a time, the documented errors sort into five distinct mechanisms — not five degrees of the same mistake:

1. A sign lost on the linear coefficient inside $u$, surviving invisibly through the even-power terms while flipping every odd-power term — confirmed directly: candidates used "'x' or '4x' in the place of '−4x'" (Jan 2024, Q1).
2. A constant not correctly carried across the outer scaling factor, producing a specific wrong numeral rather than a vague slip — confirmed as a named wrong value, "5x/4" (Jan 2021, Q1), on a $\left(\frac14-5x\right)^{\frac12}$-shaped expansion.
3. A binomial-in-X collapsed into a binomial in x — confirmed directly, candidates "confused by the 4x² term" (Oct 2021, Q4) — paired with the rarer, explicitly year-on-year-documented finding that this exact confusion measurably improved on a later series: it "did not cause the same issues as last year" (Oct 2022, Q4). This is one of the few places in the whole facts bank where Pearson's own report states a specific misconception got fixed, not just occurred.
4. A range of validity read off the original coefficients by pattern rather than derived from $u$ — confirmed as three separate named wrong answers on the same real question, including a bound used the wrong way up and a modulus dropped or doubled with a stray "±" (Oct 2022, Q4).
5. An approximation question answered from the series alone, never reconnected to the original expression — the single most severe documented outcome in this topic's evidence, "scores of 0 marks were very common" (Oct 2021, Q4) — and its more subtle sibling, where a candidate does reconnect to the original expression but stops one step early: the series approximated "the reciprocal of what they were asked to find" (Oct 2024, examiner's report), so the raw series value gets submitted for the wrong number entirely.

A sixth, lower-stakes but still real and separately confirmed angle sits alongside the five mechanisms above rather than inside them: submitting a rounded decimal where an exact fraction or surd was required — confirmed directly, a minority of candidates "gave a decimal approximation for their answer" (Jan 2021, Q1) — the same general Pure-mathematics exact-answers principle documented across this whole qualification, not unique to this topic, but real and recurring here specifically.

Two things worth stating plainly about this list as a whole, on the same "seen in N of M series checked" discipline this course holds to everywhere else: this topic has a real citation in 5 of the 7 WMA14 series this course has question-level findings for (January 2021, October 2021, October 2022, January 2024, October 2024 — absent from the question-level record for June 2022 and January 2026, per signal-frequency-and-boundaries.md §1's own recount), and its one fully-verified real anchor (§5 above) carries 6 marks (4+2) — real, single-question figures, not a paper-wide weighting, since this spec publishes no such thing.

#7. Coverage, honestly

The five mechanisms in §6 are the complete real error record this course has verified for spec 4.1 — not a sample of a larger documented set. The facts bank itself flags this topic's evidence base as narrower than several sibling WMA14 topics: four full examiner reports total (January 2021, October 2021, October 2022, January 2024) plus one further series added later once its own report was published (October 2024), against 11 series with at least one primary document existing for the whole paper. Series named in the facts bank as not yet reviewed at all, or reviewed only for shared mark-scheme boilerplate rather than question-by-question content, include October 2020 (a dead URL on fetch, never resolved), January 2022, June 2022 (its examiner report specifically — the question paper and mark scheme for that series were checked, but that series contributed no question-level binomial citation either way), June 2023, October 2023, and January 2026 (question paper and mark scheme only; no examiner's report yet published for that series at the time the facts bank was compiled). None of those absences is evidence the topic was missing from those papers — spec 4.1 is compulsory content, tested somewhere on almost every real WMA14 paper reviewed — only that this course has not yet independently verified what, if anything, went wrong on those specific series' binomial questions.

#8. Where the real material stops

Spec 4.1 is one numbered point with no lettered sub-structure, and this course's verified evidence for it is five series deep. Every genuinely distinct mechanism (formula, non-termination, convergence, the three tested bracket-shapes, the notation hazard, one fully-worked mark-scheme anchor, and every independently sourced error pattern) that the primary sources and this course's own existing WMA14 material support is stated above. There is no further real, citable material to add without either reproducing a worked exemplar that already exists elsewhere in this course or inventing a mechanism, error, or likelihood figure the sources do not contain — this document accordingly ends here rather than padding further.

#Sources

research/veridian/WMA14-verified-facts.md §1 (spec 4.1 text and formula-booklet contents), §3 (booklet P4 section), §4 (M/A/B and isw/cao conventions), §5.4 (all five series' documented errors), §6–8 (sourcing scope and flagged gaps). content/veridian/curriculum/wma14/binomial-expansion-for-rational-n.ts (cross-checked to confirm no duplication of its worked chains, method-comparison, chain-drill, marked-solutions or MCQs). content/veridian/curriculum/wma14/partial-fractions-decomposition-and-integration.ts, spec-4.1 block (cross-checked for the same reason). content/veridian/technique-hub/wma14/codex-formula-and-method-bank.md §4 and §2.4, forge-exam-technique-guide.md §1–3, signal-frequency-and-boundaries.md §1 (cross-checked to confirm this document's frequency and mark-tariff figures match rather than contradict theirs). The Maclaurin-series derivation in §2.2 and the convergence note in §2.4 are standard calculus, presented as background mathematical reasoning rather than spec- or exam-board-sourced content, and are flagged as such in the text itself.

Not affiliated with or endorsed by Pearson Edexcel. Every quotation above is reproduced under the 15-word short-quotation convention this course holds to throughout, with attribution; everything longer than that has been paraphrased.