Deep Dive: Parametric Equations Conversion and Domain Range

Provenance (2026-09-29): read in full — research/veridian/WMA14-verified-facts.md §1 (spec table), §4 (mark-scheme conventions), §5.3 (the topic's own evidence, lines 420–445), §7–8 (evidence-strength and flagged-issues lists), and the "Lesson-audit log" entry for this topic (lines 1313–1449); read in full — content/veridian/curriculum/wma14/parametric-equations-conversion-and-domain-range.ts (all 1015 lines, including its own header's sourcing-discipline notes); grepped and read the relevant sections of — content/veridian/technique-hub/wma14/codex-formula-and-method-bank.md (§3, §2.3), forge-exam-technique-guide.md (§1 table, §4), and signal-frequency-and-boundaries.md (§1, including its stated 7-series denominator); confirmed content/veridian/question-bank/wma14.ts contains no entry matching this topic. All new arithmetic below was independently computed and checked with SymPy before being written in.

#CODEX Deep-Dive: Parametric Equations — Conversion to Cartesian Form, and Domain/Range

Spec anchor: WMA14 (P4) §3.1, "Parametric equations of curves and conversion between cartesian and parametric forms" — one flat spec line, no guidance text, no lettered sub-parts. This document exists to go deeper than the lesson file's own worked chains without repeating them: it lays out the complete real theory (every distinct sub-case the spec point actually contains), a short new micro-example for each one, and then a synthesis of every real exam angle this specific topic has actually been tested from. The full mark-by-mark worked treatments (five worked chains, a chain-drill, one marked solution with its own common-wrong-path, and six MCQs) already live in content/veridian/curriculum/wma14/parametric-equations-conversion-and-domain-range.ts and are not reproduced here.

#0. How thin the real evidence actually is — stated up front

This is one of WMA14's less richly evidenced topics, and this document says so rather than dressing it up. The facts bank's own §7 shortlist — "topics with enough verified material for a genuine standalone lesson" — names five topics (proof by contradiction, vectors, volume of revolution, binomial expansion, integration by parts); parametric conversion and domain/range is not one of them. signal-frequency-and-boundaries.md uses a stated denominator of 7 series with real, question-by-question mining (January 2021, October 2021, June 2022, October 2022, January 2024, October 2024, January 2026); on that denominator, this topic is confirmed in 3 of 7 — January 2021, October 2021, October 2022. Of those three, only one (January 2021, Q4) has a real, numbered question with real parametric equations and a re-extracted, verbatim mark scheme behind it. The other two — October 2021 Q5(c) and October 2022 Q1 — are attested only as examiner-report prose: a real quoted finding about what candidates did wrong, with no surviving record in this course's sources of the actual parametric equations, domain, or numbers used on those two questions. Every claim below that names a series and question number is scoped to exactly what kind of evidence backs it — a full re-extracted mark scheme, or prose-only commentary — and does not blur the two.

A fourth series, January 2026 (Q7), touches a skill adjacent to this topic — evaluating a curve's coordinates at one given parameter value — but as one small part of a 16-mark question whose main content is area under a parametric curve (spec 6.5) and parametric differentiation (spec 5.1). Per signal-frequency-and-boundaries.md's own convention, that question is counted toward the 6.5 topic's "1 of 7," not toward this topic's "3 of 7" — §7 below explains why, and does not double-count it.

#1. The core model: what "eliminating a parameter" is

A parametric curve is a pair of equations $x=f(t)$, $y=g(t)$, where one variable $t$ (the parameter) drives both $x$ and $y$ at once. Every value of $t$ in the parameter's own domain $D$ produces exactly one point $(f(t),g(t))$; letting $t$ range over $D$ traces out the curve. "Converting to cartesian form" means finding a single equation relating $x$ and $y$ directly, with $t$ eliminated entirely, that every one of those points satisfies.

There is exactly one class of method that reliably does this, and it is pure algebra, never calculus: express the shared $t$ in terms of one variable, then remove it from the description of the other. Two real, independently confirmed sub-routes exist for the "express $t$" step (§2 and §3 below), and a third, structurally different case handles a trig-parameter pair (§4). None of the three involves differentiating $x(t)$ or $y(t)$ — a real, quoted finding on this exact question type records candidates doing exactly that and scoring nothing: "scored no marks" (Oct 2022, Q1), reaching for $\frac{dy}{dx}=\frac{dy/dt}{dx/dt}$ on a question that asks for the direct relationship between $x$ and $y$, not its gradient.

#2. Sub-case: direct substitution (the default route)

Rearrange whichever of the two given equations is simpler to solve for $t$, then substitute that expression into the other equation at every occurrence of $t$. This is the default and, on the one fully re-extracted real anchor available (January 2021, Q4), the route the mark scheme's main method line credits first: "Attempts to find t in terms of x and substitutes into y" (WMA14/01, Jan 2021, Q4(a) mark scheme).

Micro-example (Veridian-original, independently checked with SymPy): $x=2t+3$, $y=t^2-4t$. Rearranging the linear equation, $t=\frac{x-3}{2}$; substituting into $y$ and simplifying gives

$$4y = x^2-14x+33, \quad\text{equivalently } y=\frac{(x-3)(x-11)}{4}.$$

A genuine, documented failure mode on exactly this step is a partial substitution — replacing $t$ in one term of the second equation while leaving it in another, so the "cartesian" equation that results still has $t$ sitting inside it: "ended up with y in terms of both x and t" because candidates "only partially rearranged" (Oct 2022, Q1). An equation is not eliminated until $t$ is gone from every term, not most of them.

#3. Sub-case: reciprocal- or awkward-in-t pairs — clearing a compound fraction

When $t$ sits inside a reciprocal (or another form that is not immediately solvable by one rearrangement step), "solve for $t$" still applies, but the substitution step now produces a fraction nested inside another fraction, which has to be cleared by multiplying every term through by the same factor before anything can be simplified. This is a genuinely different algebraic mechanic from §2's polynomial case, and the one fully re-extracted real anchor for this topic needs exactly it: $x=\frac1t+2$ (WMA14/01, January 2021, Q4) inverts to $t=\frac1{x-2}$, and substituting that into $y=\frac{1-2t}{3+t}$ produces a fraction inside a fraction that must be cleared before the numerator and denominator can be combined — worked in full, mark-by-mark, in the lesson file's own worked chain (not reproduced here).

Micro-example (Veridian-original, distinct numbers, independently checked with SymPy): $x=\frac2t-1$, $y=\frac{t+1}{t+2}$, $t>0$. Rearranging, $t=\frac{2}{x+1}$; substituting into $y$ and clearing the nested fraction gives

$$y = \frac{x+3}{2(x+2)}.$$

#4. Sub-case: assuming the target's general form and matching coefficients

A second, structurally different route exists for exactly the case §3 makes awkward: rather than inverting $x(t)$ at all, assume the cartesian relationship has some general algebraic form (justified by the structure of $x(t)$ and $y(t)$ — e.g. a rational form $\frac{ax+b}{cx+d}$ when both are themselves rational in $t$), substitute $x(t)$ into that assumed form, and set the result equal to $y(t)$ as a genuine identity in $t$ — matching corresponding coefficients (or, equivalently, cross-multiplying and comparing powers of $t$) to solve for every unknown constant. This is a real, mark-scheme-credited alternative, not a shortcut around a proper derivation: the January 2021 Q4(a) scheme's own "Alternative 1" runs exactly this way, verbatim in its opening line: "Assume g(x)=(ax+b)/(cx+d) and substitute in x=1/t+2" (WMA14/01, Jan 2021, Q4(a) mark scheme), reaching the identical $g(x)=\frac{x-4}{3x-5}$ that direct substitution also reaches.

Micro-example (Veridian-original, distinct numbers, independently checked with SymPy): $x=\frac1t$, $y=\frac{2t+3}{t+1}$, $t>0$. Assume $g(x)=\frac{ax+b}{cx+d}$; substituting $x=\frac1t$ and multiplying numerator and denominator by $t$ gives $g\left(\frac1t\right)=\frac{a+bt}{c+dt}$. Setting this equal to $y(t)=\frac{2t+3}{t+1}$ as an identity in $t$ and matching coefficients gives $a=3,b=2,c=1,d=1$ (up to a common scalar), so

$$g(x)=\frac{3x+2}{x+1},$$

independently confirmed to be identical to the direct-substitution result.

The one real, documented way this goes wrong is not the method itself. A real examiner-report finding records some candidates who "assumed the general form of the answer... but this strategy was often unsuccessful" (Jan 2021, Q4) — but that finding describes informal guessing and checking against a couple of points, not the systematic assume-and-match method above. The mark scheme for the very same question credits the systematic version in full. The trap is guessing coefficients without ever setting up and solving the matching identity — not the act of assuming a general form at all.

#5. Sub-case: trig-parameter pairs — isolate, square, and add

When $x$ and $y$ are each written in terms of $\cos t$ and $\sin t$, $t$ cannot be isolated directly — it sits inside a trig function. The identity $\sin^2\theta+\cos^2\theta\equiv1$ (a WMA12 prerequisite result, spec 6.1 on that paper) replaces direct substitution: isolate $\cos t$ and $\sin t$ separately, each with coefficient exactly $1$ on its own side, before squaring anything; only once both are fully isolated, square each and add. Squaring before isolating produces an expanded mess that never resolves to the identity's clean $1$.

Micro-example (Veridian-original, distinct numbers, independently checked with SymPy): $x=2+5\cos\theta$, $y=-1+2\sin\theta$. Isolating, $\cos\theta=\frac{x-2}5$, $\sin\theta=\frac{y+1}2$; squaring and adding and invoking the identity gives

$$\frac{(x-2)^2}{25}+\frac{(y+1)^2}{4}=1,$$

an ellipse centred at $(2,-1)$.

#6. Domain: mapping the parameter's own restriction through x(t)

Once $t$ is restricted to a domain $D$, the set of $x$-values a real point of the curve actually reaches — $\{f(t):t\in D\}$ — is the valid domain of the cartesian equation; the bare algebraic equation, considered on its own, may make sense for a wider set of $x$ than the curve genuinely produces. Two structurally different cases exist, depending on the shape of $D$.

#6.1 Closed, bounded parameter domain

When $D=[p,q]$ and $x=f(t)$ is monotonic on $D$ (guaranteed for a linear $f$ with nonzero coefficient), the two endpoints of $D$ map directly onto the two endpoints of the $x$-domain — substitute both literal endpoint values of $t$ and state the resulting interval. The direction of the mapping depends on whether $f$ is increasing or decreasing, and getting that backwards is a real, checkable slip.

Micro-example (Veridian-original, distinct numbers, independently checked with SymPy): $x=5-2t$, $t\in[0,4]$. Here $\frac{dx}{dt}=-2<0$, so $x$ is decreasing as $t$ increases: $t=0$ gives the larger $x$-value ($x=5$), and $t=4$ gives the smaller ($x=-3$). The domain is $-3\leq x\leq5$ — stated correctly only by tracking which endpoint of $t$ maps to which end of the interval, not by assuming the first-listed endpoint of $t$ automatically gives the smaller $x$-value.

#6.2 Open and/or unbounded parameter domain

When $D$ excludes an endpoint, or has no upper bound (e.g. $t>0$), there is no literal endpoint value of $t$ to substitute at all. Each excluded or infinite end has to be reasoned about as a limit: what does $x$ approach as $t$ tends to the excluded value, and what does $x$ approach (or do) as $t\to\infty$. This is exactly the reasoning the one fully re-extracted real anchor's own domain mark requires: with $t>0$ in $x=\frac1t+2$, the scheme's own accuracy line states the result directly, "k = 2 or x > 2" (WMA14/01, Jan 2021, Q4(a) mark scheme) — reached only via the limits $t\to0^+$ and $t\to\infty$, never by substituting a literal value of $t$ that does not exist.

Micro-example (Veridian-original, distinct numbers and distinct domain shape from the real anchor — bounded above by an excluded value rather than below, independently checked with SymPy): $x=3-\frac2t$, $t>0$. Here $\frac{dx}{dt}=\frac2{t^2}>0$, so $x$ is increasing throughout: as $t\to0^+$, $x\to-\infty$; as $t\to\infty$, $x\to3$ (approached, never reached, since no value of $t$ is actually infinite). Combining the two limits: $x<3$.

#7. Range: why checking two endpoints is not enough

Range is a structurally different question from domain, and this is exactly where the closed-bounded domain shortcut of §6.1 stops being safe. $y=g(t)$ is very often not monotonic over $D$ — a quadratic in $t$ has exactly one turning point — and if that point sits at some $t^\ast$ interior to $D$, the values $g$ takes near $t^\ast$ are not bounded between $g(p)$ and $g(q)$ at all. The real, quoted mechanism: some candidates' "attempts to substitute either end of the domain in the parametric equation for y" produced only one of the two available marks, "with the minimum value missing" (Oct 2021, Q5(c)) — the exact interior value the two-endpoint shortcut cannot see.

#7.1 Non-monotonic case: locate the interior turning point first

Rewrite $y$ as a function of $x$ (or work directly in $t$), find any interior turning point by completing the square or by setting $\frac{dy}{dx}=0$ (equivalently $\frac{dy}{dt}=0$), check that its $t$-value genuinely lies inside the stated domain (a turning point outside $D$ belongs to a part of the curve that was never drawn), then combine that interior value with the two endpoint values.

Micro-example (Veridian-original, distinct numbers, independently checked with SymPy): $y=x^2-2x$ for $x\in[-1,3]$ (arising, in a full question, from some parametric pair). Completing the square, $y=(x-1)^2-1$: minimum $-1$ at $x=1$, interior to $[-1,3]$. Both endpoints happen to give the same value — $y(-1)=3$ and $y(3)=3$ — which is exactly the shape of case that makes the two-endpoint shortcut most tempting and most wrong: nothing about two equal endpoint values hints that anything different happens between them. The true range is $-1\leq y\leq3$.

#7.2 Monotonic case: no turning point exists at all

When $g$, rewritten as a function of $x$, is strictly monotonic on its domain (confirmed by checking $g'(x)$ never changes sign), §7.1's method has nothing to find — searching for a turning point that does not exist leaves no method at all for the boundary values a monotonic function's range still needs. Two boundaries are found instead: evaluate the function at its own excluded endpoint directly (valid whenever it is continuous there, even though that $x$-value is itself excluded from the domain, because the value is still the genuine limit the range approaches), and find the value it approaches as $x\to\infty$ — for a rational function $\frac{ax+b}{cx+d}$, this second boundary is the ratio of leading coefficients, $\frac{a}{c}$, since the constants become negligible once $x$ is large. Both boundaries are confirmed on the one fully re-extracted real anchor: with $g(x)=\frac{x-4}{3x-5}$, $x>2$ strictly increasing throughout (no turning point exists), the scheme credits "attempting g(2) for their g or for attempting their a/their c" (WMA14/01, Jan 2021, Q4(b) mark scheme) as the two ways to reach the two boundaries, giving $-2<g(x)<\frac13$.

Micro-example (Veridian-original, distinct numbers, independently checked with SymPy): $g(x)=\frac{2x-1}{x+2}$, $x>1$. $g'(x)=\frac{5}{(x+2)^2}>0$ throughout, so $g$ is strictly increasing with no turning point anywhere. $g(1)=\frac13$ (the excluded-endpoint boundary); as $x\to\infty$, $g(x)\to\frac21=2$ (the leading-coefficient boundary). Range: $\frac13<g(x)<2$.

#8. Domain and range are never interchangeable answers

The most bluntly attested single finding for this whole topic is not an algebra error at all: candidates "confused range with domain and gave answers in terms of x rather than y" (Jan 2021, Q4). Domain is a statement about which inputs ($x$-values) are used, fixed entirely by how $t$ is restricted; range is a statement about which outputs ($y$-values) actually occur, which depends on the full behaviour of $g$ across that domain, turning points included. They answer different questions, and one is never a smaller or partial version of the other.

#9. Adjacent skill: evaluating coordinates at one given parameter value

A small, distinct skill that belongs to the same spec point (parametric curves) but is not conversion or domain/range at all: given a specific numeric value of $t$, substitute it into both $x(t)$ and $y(t)$ to find one point's coordinates. This appears as a short opening part of a much larger real question: given $x=8\cos^3\theta$, $y=6\sin^2\theta$, the point at $\theta=\frac\pi3$ has coordinates $(1,4.5)$ — verified independently ($\cos\frac\pi3=\frac12\Rightarrow x=8\times\frac18=1$; $\sin\frac\pi3=\frac{\sqrt3}2\Rightarrow y=6\times\frac34=4.5$) — worth 2 marks inside a 16-mark question whose remaining parts test a tangent/normal via parametric differentiation (spec 5.1) and area under a parametric curve (spec 6.5) (WMA14/01A, January 2026, Q7). No examiner's report exists yet for this series (sat 22 January 2026), so only the mark scheme's mechanics — not any documented candidate error — can be cited for it.

Micro-example (Veridian-original): $x=t^2+1$, $y=3t-2$; at $t=2$, the point is $(5,4)$.

#10. Wrong-tool and process traps confirmed on this exact topic

Two further real findings, both from the same question, concern process rather than any specific piece of algebra:

• Reaching for calculus on a pure substitution question. "A few candidates thought it appropriate to use calculus and scored no marks" (Oct 2022, Q1) — forming $\frac{dy}{dx}$ answers a different question (the curve's gradient) from the one asked (the direct relationship between $x$ and $y$).
• Partial rearrangement. Substituting into one occurrence of $t$ while leaving another untouched produces an expression that still contains $t$ and is therefore not yet a cartesian equation at all — "a disappointing number of responses" did exactly this (Oct 2022, Q1, paraphrased; see §2 above for the fuller quoted fragment).

#11. Synthesis — every real exam angle this topic has actually been tested from

Read across these four rows, the topic's real exam angles cluster into exactly the sub-cases worked through in §2–§9 above: which elimination route to use and how it can be guessed rather than derived; which domain method applies to which shape of parameter interval; why an interior turning point can defeat an endpoints-only range method; and the direct domain/range terminology confusion. No fifth angle, and no numeric grade-boundary or likelihood claim, is supported by anything in the reviewed sources.

#12. What the real record does not support

Stated plainly, per this course's standing sourcing discipline: there is no official per-topic mark weighting for WMA14 to check any of the above against — the spec (p.69) publishes only assessment-objective mark ranges for the whole 75-mark paper (AO1 25–30, AO2 25–30, AO3 5–10, AO4 5–10, AO5 5–10), nothing at the level of an individual spec point. No grade-boundary figure specific to this topic exists in any source reviewed for this document. "3 of 7 series" is a count against a small, explicitly stated denominator, not a probability, and is not rounded into one here. Two of this topic's three confirmed real questions (Oct 2021 Q5(c), Oct 2022 Q1) are attested only through examiner-report prose, with no surviving record of the actual parametric equations or numbers used — this document does not invent them to fill that gap, and stops exactly where the real evidence does.

#13. Where the full worked treatment lives

Every mark-by-mark worked chain, the chain-drill, the full marked solution (with its own common-wrong-path), all six MCQs, the diagram, the trap-taxonomy, and the reference card for this topic are in content/veridian/curriculum/wma14/parametric-equations-conversion-and-domain-range.ts and are not duplicated here. The method skeleton (a compressed, formula-adjacent version of §2–§9 above) is cross-linked from codex-formula-and-method-bank.md §2.3; the topic's place in the paper's real timing and mark-tariff picture is in forge-exam-technique-guide.md §1; its frequency count against the stated 7-series denominator is in signal-frequency-and-boundaries.md §1.