Pure Mathematics 3

Deep Dive: Rational Expressions and Algebraic Division

Provenance: written 2026-09-29. Primary source, read and grepped directly for this topic: research/veridian/WMA13-verified-facts.md (§1 spec table, §4 mark-scheme conventions, §5 "1. Algebra and functions" — the "1.1 Algebraic division / rational expressions" entries and the corrected Jan 2025 Q4(a) paragraph). Cross-checked against what already exists, specifically to keep this document additive rather than repetitive: content/veridian/curriculum/wma13/rational-expressions-and-algebraic-division.ts (full file read, both header comments, every block, all five trap-taxonomy entries and all fourteen flashcards); content/veridian/technique-hub/wma13/forge-exam-technique-guide.md, codex-formula-and-method-bank.md and signal-frequency-and-boundaries.md (grepped for every existing 1.1 citation); content/veridian/question-bank/wma13.ts (grepped — contains no spec-1.1 items; its real coverage is spec 4.4 only). One real anchor cited by forge-exam-technique-guide.md (October 2024 Q9(c)) was independently re-pulled from the primary PDFs at /Users/neosteinhoff/Documents/SOL/Papers/IAL_Paper3_Complete/IAL_Materials/Mathematics/October 2024/ (Mathematics_Unit 3_QP/MS/ER_October_2024.pdf, Publications Code WMA13_01_2410_MS) and re-verified by hand and with sympy before being written in below — this turned up a real correction to how that anchor's own divisor has been described elsewhere in this course's material (§3 below).

#CODEX Deep-Dive: Rational Expressions and Algebraic Division (WMA13, spec 1.1)

#Scope, and how this relates to what already exists

Spec 1.1 is already well covered pedagogically in this course: the lesson rational-expressions-and-algebraic-division.ts derives the factor-vs-term cancelling rule from first principles, derives the remainder theorem from the division algorithm identity, and works three full division exemplars (a worked-chain, a chain-drill, and a marked-solution with a common-wrong-path) — all three using a quadratic divisor whose remainder shares a real factor with it. The technique-hub already carries a compressed method skeleton for this ("Pattern A") and for the one other confirmed real pattern where the divisor is an irreducible quadratic ("Pattern B") in codex-formula-and-method-bank.md §2.1, and forge-exam-technique-guide.md gives the cross-cutting mark-tariff and mark-coding context.

This document does not re-derive or re-quote any of that. Its job is to state the complete real model for this topic in one place, organised by the sub-cases the spec's own guidance actually creates; to add the one genuinely new fact this pass found (a real, confirmed third sub-case, and a correction to how it has been described elsewhere in this course); and to say plainly where the real evidence runs out, per this course's own standing rule against padding.

#1. What spec 1.1 actually covers

Verbatim, spec Issue 3 (April 2019), Unit P3, p.20: "Simplification of rational expressions including factorising and cancelling, and algebraic division." The spec's own guidance line adds one hard constraint that governs every sub-case below: "denominators of rational expressions will be linear or quadratic." That single sentence is what makes the sub-case taxonomy in §2 exhaustive rather than illustrative — a divisor on this paper can only ever be linear or quadratic, so there are only three structurally distinct division shapes it can produce, and the archive's real anchors turn out to instantiate all three.

#2. Simplifying a rational expression — factor, not term

Fully derived in the lesson (the $\frac{Ak}{Bk}\to\frac AB$ vs. $\frac{A+k}{B+k}$ contrast) and not repeated here. The rule in one line: a rational expression $\frac{P(x)}{Q(x)}$ only simplifies once $P$ and $Q$ have been fully factorised, and only a factor that multiplies the whole of each side may then be cancelled — never a term that merely looks the same inside an unfactorised sum. A short illustrative micro-example, distinct from every number already used in the lesson or its MCQs, to keep this reference self-contained without duplicating them: $\dfrac{x^2-4}{x^2-2x}=\dfrac{(x-2)(x+2)}{x(x-2)}=\dfrac{x+2}{x}$ , valid for $x\neq0$ and $x\neq2$ — the second exclusion is invisible in the simplified form and must be read off the original, unfactorised denominator, exactly as the lesson's own mechanism block establishes.

No real past-paper anchor was located anywhere in the archive testing this skill on its own, as a standalone multi-mark question separate from a division question — see §5.

#3. Algebraic division — the three real sub-cases the spec's own divisor constraint creates

The organising fact, already derived in the lesson from the division algorithm identity $f(x)\equiv D(x)Q(x)+R(x)$ : the division is not finished until $\deg R<\deg D$ . Because $D(x)$ is restricted to linear or quadratic, this single stopping condition cashes out into three genuinely different situations, not one:

3.1 Linear divisor — the remainder must collapse all the way to a bare constant. If $\deg D=1$ , then $\deg R<1$ forces $R$ to have no $x$ term at all. This is the strictest of the three cases, and it is the one this pass found newly confirmed, real evidence for.

Real anchor, independently re-verified against the primary PDFs: October 2024 Q9(c). The curve $f(x)=\dfrac{6x^2+4x-2}{2x+1}$ , $x>-\tfrac12$ , is to be written as $Ax+B+\dfrac{D}{2x+1}$ (3 marks). Dividing by hand: $6x^2\div2x=3x$ ; $3x(2x+1)=6x^2+3x$ ; subtracting leaves $x-2$ — still degree 1, the same degree as the divisor, so the division is not yet finished. Continuing: $x\div2x=\tfrac12$ ; $\tfrac12(2x+1)=x+\tfrac12$ ; subtracting leaves $-\tfrac52$ , now degree 0 and genuinely final. So $A=3$ , $B=\tfrac12$ , $D=-\tfrac52$ — confirmed independently with sympy and matching the real mark scheme's own B1 note exactly.

The real examiner report names two distinct errors on this part, in increasing order of severity: a milder one, misreading the true remainder as $-\tfrac32$ ; and a more serious one, stopping after the first division step and reporting "terminate the division early leaving (x − 2) as the remainder" (paraphrased above the ~15-word quotation limit this document holds itself to; the quoted fragment is verbatim). The second error is worth stating precisely because of why it is more serious than a numeric slip: $x-2$ is not merely wrong, it is structurally impossible as a final remainder against a linear divisor, since its degree ( $1$ ) is not strictly less than the divisor's ( $1$ ) — the division algorithm's own stopping condition was never met.

A correction, found in the course of verifying this anchor. forge-exam-technique-guide.md §5.1(f) already documents this same October 2024 Q9(c) finding, but describes the stopping-condition failure as happening "while the remainder still contains an x term" after dividing "by a quadratic divisor." The real divisor in this question is $2x+1$ — linear, not quadratic. The mechanism that guide names (division is not finished until the remainder's degree drops below the divisor's) is correct and is exactly the mechanism restated above; only the divisor's degree was mis-stated. This document does not edit that file — it is out of this task's scope — but the correct attribution belongs on record here, independently re-derived from the primary question paper, mark scheme and examiner report rather than taken on trust from the existing guide.

This also makes October 2024 Q9(c) the only located real anchor in the archive where the remainder must reduce to a bare constant — every other real division anchor located for this topic (below) uses a quadratic divisor, where a genuinely final remainder is still allowed to be linear.

3.2 Quadratic divisor, remainder shares a real factor — the "cancel and simplify" pattern. If $\deg D=2$ and the divisor factorises over the reals, a division that stops correctly (remainder linear, degree $1<2$ ) is not yet in its most reduced form: the remainder and the divisor may still share a common factor, and finishing the question requires factorising both and cancelling it. Real anchor: October 2020 Q9(a), $\frac{x^4-x^3-10x^2+3x-9}{x^2-x-12}\equiv x^2+P+\frac{Q}{x-4}$ , "show $Q=5$ " — already fully worked, three times over (worked-chain, chain-drill, marked-solution), in the lesson itself; not reproduced here. The real examiner report confirms candidates who divided correctly then stopped one line early lost the mark specifically for not finishing the cancellation, even with the arithmetic otherwise flawless — the exact mechanism already named in the lesson's own cancel-and-simplify-step-abandoned trap.

3.3 Quadratic divisor, irreducible — nothing to cancel; the constant is verified, not cancelled. If the divisor's discriminant is negative it has no real linear factor at all, so §3.2's route is structurally unavailable regardless of how the division was carried out. Real anchor: January 2025 Q4(a), $\frac{4x^3+2x^2+3x+8}{x^2+4}\equiv Ax+B+\frac{Cx+D}{x^2+4}$ , "show $D=0$ " — $x^2+4$ has discriminant $-16$ , no real factor. Already fully documented in the lesson's given-value-not-verified-by-substitution trap and in the technique-hub's Pattern B; not reproduced here. The real mark scheme awards this as an independent "shown" mark, earned by setting up the coefficient-matching identity and substituting to prove the stated value directly — not by any cancellation, because there is none available.

A synthesis genuinely new to this course's WMA13 material, tying 3.1–3.3 into one procedure. These three sub-cases are not three unrelated question types to memorise separately; they follow directly, and in order, from two checks anyone can make before dividing:

Is the divisor linear or quadratic? If linear, §3.1 applies: the remainder must end as a bare constant, and a lower-degree-but-nonzero-degree result is not yet finished, whatever the arithmetic looks like.
If quadratic, does it factorise over the reals (discriminant $\geqslant0$ )? If yes, a correctly finished division may still admit further cancellation (§3.2) — check by factorising the remainder too. If no (discriminant $<0$ ), no amount of factorising will ever produce a cancellation, and a "show that [constant] = value" instruction has to be met by coefficient-matching or substitution instead (§3.3).

No real anchor exists in the archive for a fourth combination (e.g. a linear divisor paired with a "show a given constant with no cancellation" instruction) — the spec's own two-degree constraint on $D(x)$ means there isn't a fourth structural case to look for, only these three.

#4. Combining rational expressions over a common denominator

A related but distinct skill from division: adding or subtracting two rational expressions by finding a common denominator first. The verified-facts archive confirms this specific step — not the cancel-after-dividing step of §3.2, but the general discipline of showing the combined-fraction working rather than jumping to the simplified result — as a common source of lost marks across at least three series (Oct 2020, Jan 2024, Jan 2025), independently of which of §3's three mechanisms the underlying question actually used to reach its final constant. No further per-series question detail for this specific finding is given in the archive beyond that three-series count, so none is invented here; it is already built into the lesson as the combined-fraction-not-fully-justified trap and is not repeated.

#5. What was not found

No standalone real past-paper question testing "simplify a rational expression and state its domain" in isolation from a division question (§2).
No real anchor for a linear divisor paired with an irreducible-quadratic-style "nothing to cancel" instruction — structurally impossible given the spec's own denominator constraint (§3.3).
No further detail beyond a three-series count for the Jan 2024 instance of the combining-fractions pattern in §4 — the archive names the series but not the question number or wording.
No likelihood, probability or grade-boundary claim is made anywhere above — every count is "seen in N of the series actually checked," per this course's own standing rule.

#Sources

research/veridian/WMA13-verified-facts.md — §1 (spec text and guidance), §5 "1.1 Algebraic division / rational expressions" (Oct 2020 Q9(a), the corrected Jan 2025 Q4(a) paragraph, and the three-series combining-fractions finding).
content/veridian/curriculum/wma13/rational-expressions-and-algebraic-division.ts — checked in full to keep this document additive; its cancelling derivation, division-algorithm mechanism, and three worked exemplars (Oct 2020 Q9(a)-pattern, plus two VERIDIAN-original divisions of the same quadratic-divisor shape) are cited, not reproduced.
content/veridian/technique-hub/wma13/codex-formula-and-method-bank.md (§2.1, Patterns A and B), forge-exam-technique-guide.md (§5.1(f) and its mark-tariff table), signal-frequency-and-boundaries.md (1.1 frequency count) — checked for existing coverage and cross-referenced rather than repeated.
Primary PDFs, independently downloaded and read for this pass: Mathematics_Unit 3_QP_October_2024.pdf, Mathematics_Unit 3_MS_October_2024.pdf, Mathematics_Unit 3_ER_October_2024.pdf (Publications Code WMA13_01_2410_MS), at /Users/neosteinhoff/Documents/SOL/Papers/IAL_Paper3_Complete/IAL_Materials/Mathematics/October 2024/. The October 2024 Q9(c) division was independently re-verified by hand and with sympy (sp.div, sp.apart) before being written in above.
Not affiliated with or endorsed by Pearson Edexcel. Every specific number, fraction under ~15 words, and mark-scheme mechanism attributed above to a real question was checked against a primary source named in this section, not carried over on trust from a secondary summary.