Deep Dive: Hidden Quadratic Substitution

#CODEX Deep-Dive: Hidden Quadratic Substitution (WMA11, spec 1.1 / 1.5 / 1.6)

Provenance (2026-09-29). Built from a direct read of research/veridian/WMA11-verified-facts.md (§1 spec table of contents, verified against the source spec PDF; §4.1 "Indices/surds, especially 'hidden quadratic' substitution questions"; the hidden-quadratic-substitution.ts audit entry; §7 series-coverage honesty notes) and content/veridian/technique-hub/wma11/signal-topic-brief-hidden-quadratic-substitution.md in full; the existing lesson content/veridian/curriculum/wma11/hidden-quadratic-substitution.ts in full (read to avoid duplicating its worked exemplars); codex-formula-and-method-bank.md §2.1.2 (checked for overlap — one line, no duplication); content/veridian/question-bank/wma11.ts (grepped — no item on this topic) and content/veridian/specCoverage.ts (spec 1.1's guidance text). Two primary-source PDF sets were freshly opened and independently re-extracted for this document (pdftotext -layout), each identified from its own header/Publications Code/QP Log Number before use, not trusted from a summary: October 2019 (Mathematics_Unit 1_{QP,MS,ER}_October_2019.pdf, local archive IAL_Paper1_Complete; identity confirmed — QP Log Number P58436A, MS WMA11_01_1910_MS, ER WMA11_01_1910_ER) and October 2022 (Mathematics_Unit 1_{QP,MS,ER}_October_2022.pdf, same archive; identity confirmed — QP Log Number P72137A, MS WMA11_01_MS_2210, ER WMA11_01_ER_2210). Not affiliated with or endorsed by Pearson Edexcel.

#0. What this document adds, and what it deliberately does not repeat

The existing lesson already teaches one real flavour of this topic in full depth: substitution from an exponential equation ($u = a^x$), with a complete, mark-scheme-verified worked solution (Jan 2023 Q5), two original worked chains, a chain-drill, four MCQs, and six named, cited traps. None of that is reproduced here.

This document is additive in four specific ways:

1. A general mechanism stated once, covering both directions this topic actually takes — the lesson's own mechanism proof covers only the exponential direction ($a^{x+k}=a^k u$, $a^{2x}=u^2$). §1 below states the underlying test in a form that also covers the other real sub-case, which needs the substituted quantity to be squared in the opposite direction.
2. A second, genuinely distinct sub-case the lesson does not teach at all: root substitution ($u = \sqrt{x}$), anchored on a real WMA11 question — Oct 2019 Q8 — whose full mark scheme this document re-extracted directly from the primary PDF rather than relying on the facts bank's own partial citation of it (which the facts bank itself flagged as incomplete: "Oct 2019 Q8's tariff is likewise uncaptured beyond one A1 requiring the final surd answer simplified," per the SIGNAL brief).
3. A boundary case, precisely flagged as adjacent rather than folded in uncritically: a real Oct 2022 question (Q4) found by this pass's own targeted search of a series the facts bank had confirmed genuine but never mined question-by-question. It shares this topic's exact "solve → apply a domain restriction → convert back" discipline on a fractional exponent, but its base equation is a cubic, not a quadratic — the distinction matters, and is argued for explicitly rather than glossed over.
4. A single honest synthesis of every real exam angle this topic has actually been tested from, with a frequency accounting that keeps systematically-mined series separate from series found by targeted search, per this course's own evidentiary convention.

Where the real material runs out, each section says so. No likelihood percentage or grade boundary is stated anywhere below — only "seen in N of M series checked," and which M that is.

#1. The general test: when does an equation collapse to a quadratic in one new letter?

Spec 1.1 states the laws of indices "for all rational exponents" and names one further fact as expected knowledge: "the equivalence of $a^{m/n}$ and $(\sqrt[n]{a})^m$" (spec 1.1, transcribed in specCoverage.ts from the source spec PDF). That one clause is the reason this topic has more than one real shape. A rational exponent can be read two ways — as a power applied to a base ($a^x$, the base is fixed, the exponent varies) or as a root taken of a variable ($x^{1/2}=\sqrt x$, the "base" is the variable itself) — and WMA11 substitution questions are built from both readings.

The single test that covers both: an equation hides a quadratic in a new variable $u$ exactly when every term, after choosing the right "atomic" building block $b$, is either a constant, a constant multiple of $b$, or a constant multiple of $b^2$ — and at least one term genuinely needs the $b^2$ slot, or there would be nothing quadratic to solve. Set $u=b$ and the equation is, by construction, an ordinary three-term quadratic in $u$.

• Exponential sub-case: $b = a^x$. A term $a^{x+k}=a^k\cdot a^x=a^k u$ (linear in $u$); a term $a^{2x}=(a^x)^2=u^2$ (quadratic in $u$). Here $u$ is the low power and squaring the exponent's $x$-coefficient produces the $u^2$ slot. This is the lesson's own mechanism, proved in full there.
• Root sub-case: $b=x^{1/2}=\sqrt x$. Here the relationship runs the other way: $\sqrt x$ is $u$ (the irregular-looking term, degree 1 in $u$), and the plain variable $x$ itself is the one that supplies the $u^2$ slot, because $x=(\sqrt x)^2=u^2$. The "awkward" object becomes $u$; the "ordinary-looking" $x$ becomes $u^2$ — the exact mirror image of the exponential case, where the "ordinary-looking" $a^x$ becomes $u$ and the less familiar $a^{2x}$ becomes $u^2$.

Both are the same one-line test — do the exponents present, relative to some atomic unit, sit at $\{0, 1, 2\}$? — applied to spec 1.1's two readings of a rational exponent. This unifying statement is not written anywhere in the facts bank or the lesson in this form; it is offered here as a direct, checkable consequence of the laws already cited, not as a new fact requiring its own exam citation.

#2. Sub-case A: exponential substitution, $u=a^x$ — summary and one fresh micro-example

Fully taught in the lesson (hidden-quadratic-substitution.ts): recognise the same base raised to $x$ plus-or-times a constant, appearing more than once; substitute $u=a^x$; solve the resulting quadratic by any of the three standard routes; reject any negative value of $u$ outright, since $a^x>0$ for every real $x$ when $a>0$; convert the surviving root(s) back to $x$ by matching indices (no logarithm is available or needed at P1, since every real WMA11 sighting of this sub-case is built so the surviving root is an exact power of the base). The lesson's own real anchor, Jan 2023 Q5 ($3\times9^x+3^{x+2}=1+3^{x-1}$, reducing to $9p^2+26p-3=0$), carries a full six-mark, two-part mark scheme (B1 M1 A1* then M1 A1 A1) already transcribed there in full, and is not repeated here.

One fresh micro-example, on numbers not used anywhere in the lesson, to keep this sub-case self-contained on this page: solve $7^{2x}-6\times7^x-7=0$. With $u=7^x$: $u^2-6u-7=(u-7)(u+1)=0$, so $u=7$ or $u=-1$. Reject $u=-1$ ($7^x>0$ always). $7^x=7=7^1\Rightarrow x=1$. One line, one rejection, one conversion — the whole discipline in miniature.

#3. Sub-case B: root substitution, $u=\sqrt x$ — a real anchor the lesson does not cover

#3.1 The real question

Oct 2019 Q8 (5 marks): "Solve, using algebra, the equation $x-6x^{1/2}+4=0$", with the instruction to fully simplify the answer "in the form $a+b\sqrt c$" (verified directly from the QP; quoted phrase under 15 words). This is spec 1.5/1.6 exactly as the exponential sub-case is, but the substitution runs in the direction §1 describes: $u=x^{1/2}$, so $x=u^2$, turning the equation into $u^2-6u+4=0$.

#3.2 The real mark scheme, paraphrased (re-extracted from WMA11_01_1910_MS)

• M1: attempts to solve the resulting quadratic in the new letter (completing the square or the formula) with genuine working shown — the mark scheme explicitly accepts, as a minimum, identifying $u=x^{1/2}$ and writing the quadratic in $u$ before any solution appears.
• A1: both values of $u$ found, $u=3\pm\sqrt5$ (need not yet be simplified).
• M1: squares a value of the form $p\pm q\sqrt r$ back into $x$, with at least two of the four expansion terms correct.
• A1, A1: each of $x=14+6\sqrt5$ and $x=14-6\sqrt5$, stated fully simplified — the mark scheme is explicit that this final pair must be simplified, not left as $14\pm\sqrt{180}$.
• A genuine mark-scheme-credited alternative route exists and was equally reproduced in the primary source: isolate the root first ($x+4=6\sqrt x$), square both sides directly ($({x+4})^2=36x$), expand to the same quadratic in $x$ ($x^2-28x+16=0$), and complete the square the same way — an implicit-substitution route that never writes $u=\sqrt x$ down explicitly but is accepted equally, provided the squaring is done to the whole expression, not term-by-term (the mark scheme's real, quoted caveat: "candidates who square term by term will score no marks").

#3.3 What makes this sub-case's discipline genuinely different, not just a relabelled copy

Both real roots of $u$ here are positive ($3+\sqrt5\approx5.24$ and $3-\sqrt5\approx0.76$), so — unlike almost every worked exponential-substitution example — no root is rejected on this particular question. That is not a weaker version of the exponential sub-case's discipline; it is the same discipline (test every value of $u$ against what $u$ can actually be) landing on a different outcome, exactly as the lesson's own chain-drill warns for its own sub-case: the check is the same one line whether or not a root actually fails it. A root substitution absolutely can produce a value of $u$ that must be rejected — since $u=\sqrt x$ requires $u\geq0$ — it simply doesn't happen to on this specific real anchor. A short original example showing the rejection case: solve $x-\sqrt x-2=0$. With $u=\sqrt x$: $u^2-u-2=(u-2)(u+1)=0$, so $u=2$ or $u=-1$; reject $u=-1$ since $\sqrt x\geq0$; $\sqrt x=2\Rightarrow x=4$ (checks: $4-2-2=0$).

The real, examiner-confirmed traps on this sub-case are correspondingly different from the exponential sub-case's own two headline traps:

• Squaring term-by-term instead of forming and solving one clean quadratic in $u$ — confirmed as "a frequent mistake" that "immediately led to the loss of all marks" (WMA11_01_1910_ER; both quoted phrases under 15 words), a failure mode with no analogue in the exponential sub-case, where there is nothing to square in the first place.
• Stopping at $u$ — the same-named trap as the exponential sub-case, confirmed independently on this structurally different question: "many of these stopped at finding the roots $3\pm\sqrt5$, thinking this was the answer" (WMA11_01_1910_ER). Here the missing step is squaring back to $x$, not matching an index — a different piece of algebra, but the identical discipline failure.
• Final surd left unsimplified — $14\pm\sqrt{180}$ instead of $14\pm6\sqrt5$ — a sub-case-specific trap absent from the exponential family, since that family's answers are exact integers or simple fractions, never surds requiring simplification.
• No initial working shown, jumping straight to the roots of $x^{1/2}$: the mark scheme's own special-case note confirms this costs exactly the first two marks even when the final answer is fully correct (M0A0M1A1A1 is the maximum reachable), and the examiner report separately notes working "was in evidence in most cases, but some students started their answer with the roots for $x^{1/2}$ with no evidence" (quoted phrase under 15 words).

The series-level statistic is worth stating precisely rather than rounded into a vaguer claim: this was "one of the most poorly answered question[s] on the paper," with almost half of responses blank or scoring no marks and only a third reaching full marks (WMA11_01_1910_ER) — a real, sourced difficulty figure for this one series' sighting of this one question, not a claim about the sub-case's difficulty in general.

#4. A boundary case: fractional-exponent conversion following an already-solved cubic (Oct 2022 Q4)

This one is flagged deliberately as adjacent, not folded into the sub-case list above as a third hidden quadratic, because it genuinely is not one — naming that distinction precisely is more useful than a false count.

The real question (6 marks total, found by a targeted search of Oct 2022's question paper for "solve"/"hence" during this pass — not a systematic mining of the whole series). Part (a): given $f(x)=x^2(2x+1)-15x$, "solve $f(x)=0$" (4 marks) — this expands to the cubic $2x^3+x^2-15x=x(2x^2+x-15)=x(2x-5)(x+3)$, giving $x=0,\ \tfrac52,\ -3$ (M1 for reaching the factored quadratic form; dM1 for a valid method solving that quadratic; B1 for two of the three roots; A1 for all three — re-extracted directly from WMA11_01_MS_2210). This is a cubic solved by taking out a common factor $x$ and then solving the remaining quadratic — a real, verified instance of exactly the "form the quadratic before applying a quadratic technique" discipline this topic's own trap-taxonomy already names by analogy to a different paper (Jan 2019 Q9, cited in the lesson's own quadratic-not-yet-formed... trap) — but the overall equation being solved is cubic, not quadratic, so it does not itself belong in the sub-case list above.

Part (b) is where the genuine connection to this topic lives: "Hence solve" $y^{4/3}(2y^{2/3}+1)-15y^{2/3}=0$, given $y>0$ (2 marks). This expression is exactly $f(u)$ with $u=y^{2/3}$ substituted in — so part (a)'s three roots transfer directly: $u=0,\ \tfrac52,\ -3$. Since $u=y^{2/3}=(y^{1/3})^2\geq0$ for any real $y$, the negative root ($u=-3$) is impossible on sign grounds alone (M1 in the real scheme is for reaching $y^{2/3}=\tfrac52\Rightarrow y=(\tfrac52)^{3/2}$); and $u=0$ gives $y=0$, excluded by the question's own stated $y>0$ — leaving exactly one valid value, $y=\left(\tfrac52\right)^{3/2}=\dfrac{5\sqrt{10}}{4}$ (independently re-derived: $(\tfrac52)^{3/2}=\tfrac52\sqrt{\tfrac52}=\tfrac{5\sqrt5}{2\sqrt2}=\tfrac{5\sqrt{10}}{4}$, matching the real mark scheme's A1cso exactly).

This is a fractional-exponent substitution ($u=y^{2/3}$, not $a^x$ or $\sqrt x$) doing precisely this topic's central job — collapse an unfamiliar expression to a single new letter, solve, then filter surviving values by what that letter can actually equal — on a rational exponent other than $1$ or $\tfrac12$, which is itself a real, if narrow, confirmation that spec 1.1's "for all rational exponents" phrasing is not decorative: the technique genuinely is not restricted to the two exponent values the two sub-cases above happen to use. The real, confirmed trap on this part is a domain-restriction failure of exactly the family already documented elsewhere in this course on a different, unrelated question (hidden-higher-degree-equations-from-differentiation.ts's domain-restriction-overrides-the-default-both-signs-rule, sourced there to Jan 2025 Q5(c)): candidates who correctly reach the valid $u$-value but then keep an excluded solution anyway. Here the examiner report confirms the second most common error was retaining $y=0$ and/or the (impossible) negative-root value "even though the question stated that $y>0$" (WMA11_01_ER_2210, quoted phrase under 15 words) — a fresh, independently-verified third sighting of that same cross-topic trap family, on a genuinely different question and paper series than either of its two previously-documented sightings.

#5. What this topic is not — three real, checkable disambiguations

Getting the boundary of "hidden quadratic substitution" right matters more than padding its list of sightings, so three adjacent real WMA11 items are named here explicitly rather than left to be confused with it:

1. Jan 2019 Q9 (cited in this topic's own trap-taxonomy by analogy) requires forming a quadratic in $x$ by clearing a fraction before a discriminant question can even begin — a real "hidden quadratic," but its own topic is discriminant-with-a-parameter.ts, not substitution into a new letter at all: no $u$ is ever introduced, and the quadratic that needs forming is in the same variable, $x$, throughout.
2. Oct 2021 Q2 and Jan 2025 Q5(c) (the anchors for hidden-higher-degree-equations-from-differentiation.ts) produce a quartic in $x$ from a differentiation step ($15x^4+12x^2-1=2$, or similar), solved via $u=x^2$. This is structurally the closest real relative to this topic's own mechanism — same $u=(\text{variable})^2$ substitution shape as this document's §1 root-direction case — but it arises from calculus, not from an indices/exponential or root equation handed to you directly, and it is already a separately-taught lesson with its own full mark-scheme audit; it is not re-taught here.
3. A genuine non-example, checked and correctly excluded by the SIGNAL brief: Jan 2019's own indices question (Q2) was checked directly and found to be a plain laws-of-indices rearrangement with no hidden quadratic at all — confirming that not every WMA11 question mentioning indices belongs to this topic, and that the "2 of 5" frequency figure below is a genuine count, not one padded by a nearby-but-different question.

#6. Synthesis: every real exam angle this topic has actually been tested from

Keeping the systematically-mined evidence separate from evidence found by targeted search, per this course's own convention:

• Exponential substitution, systematically confirmed: seen in 2 of 5 WMA11 series the facts bank fully mined question-by-question for this specific skill (Jan 2019, Oct 2020, Oct 2021, Jan 2023, Jan 2025) — Jan 2023 Q5 and Jan 2025 Q4(b), both carrying the identical "stops at the substituted variable" trap, confirmed near-verbatim in both reports.
• Root substitution, found by targeted rather than systematic search: one confirmed sighting, Oct 2019 Q8, first located by a different lesson's audit search and re-verified in full from the primary PDF by this document. Because it was not found by a systematic per-series search, its absence from any other, unsearched series cannot be read as confirmed absence — the honest total is "at least one confirmed sighting," not "1 of N."
• Fractional-exponent conversion after an already-solved higher-degree equation, found by this document's own targeted search of one previously-unmined series: one sighting, Oct 2022 Q4(b) — genuinely adjacent (§4), not counted toward the two totals above.
• Two further series the facts bank confirms as genuine matching Pearson documents but not (or, for Oct 2022, only partially, via this pass's own single-question search) mined question-by-question for this skill: Oct 2022 (this document's own search covered one question, not the full paper) and Jan 2024 (not available in the local archive used for this pass, and not fetched live for this document — a real, stated gap, not a claim of absence).

Mark-tariff patterns actually observed (three real data points, not a general rule): Jan 2023 Q5 splits 6 marks as 3+3 across a "show" part and a "hence solve" part; Oct 2019 Q8 is a single 5-mark part (M1 A1, then M1 A1 A1 — solve the quadratic in $u$, then convert back); Oct 2022 Q4 splits 6 marks as 4+2 across "solve the cubic" and "hence, using the substitution, solve for $y$." No two of the three real tariffs actually seen split marks the same way — the one safe generalisation is that the "convert back" step reliably carries its own, separately-creditable mark(s) in every real sighting checked, never folded silently into the mark for solving the quadratic in $u$.

#7. Consolidated trap taxonomy across every real sub-case

#8. Sources

• research/veridian/WMA11-verified-facts.md — §1 (spec 1.1 verbatim text); §4.1 (Jan 2023 Q5, Jan 2025 Q4, Jan 2019 Q2/Q9 citations); the hidden-quadratic-substitution.ts lesson-audit entry (Jan 2023 Q5's full mark-scheme provenance); §7 series-coverage notes.
• content/veridian/technique-hub/wma11/signal-topic-brief-hidden-quadratic-substitution.md — the "2 of 5 series" frequency accounting and the two-family split, read in full and extended rather than restated.
• content/veridian/curriculum/wma11/hidden-quadratic-substitution.ts — the existing lesson, read in full; its exponential-substitution mechanism, marked-solution and six traps are cited by reference, not reproduced.
• content/veridian/technique-hub/wma11/codex-formula-and-method-bank.md §2.1.2 — checked for overlap (one summary line, no duplication).
• content/veridian/question-bank/wma11.ts, content/veridian/specCoverage.ts — checked; no question-bank item and one spec-guidance line (1.1's $a^{m/n}$ equivalence) used directly.
• October 2019, Pearson Edexcel WMA11/01 — QP, MS, ER, re-extracted fresh from the local archive (pdftotext -layout), identity confirmed from each document's own header (QP Log Number P58436A; MS WMA11_01_1910_MS; ER WMA11_01_1910_ER). Q8 in full: question text, both real mark-scheme routes (main and alternative), the special-case note, and the examiner report's descriptive paragraph.
• October 2022, Pearson Edexcel WMA11/01 — QP, MS, ER, re-extracted fresh from the local archive, identity confirmed from each document's own header (QP Log Number P72137A; MS WMA11_01_MS_2210; ER WMA11_01_ER_2210). Q4 in full: question text (parts a and b), the complete mark scheme for both parts, and the examiner report's descriptive paragraph.
• §1's unifying "exponents at $\{0,1,2\}$ relative to an atomic unit" statement, and all arithmetic re-derivations in §3–§4, are original work for this document, independently checkable, not sourced to any document.