Topic Brief: Hidden Quadratic Substitution

Provenance: compiled 2026-09-29, updated same day. Source is research/veridian/WMA11-verified-facts.md §4.1 ("Indices / surds, especially 'hidden quadratic' substitution questions") and its §8 lesson-audit log for hidden-quadratic-substitution.ts, cross-checked against that lesson's citations and against the single-line entry in signal-frequency-and-boundaries.md ("Hidden quadratic via substitution | 1.1/1.5/1.6 | 2 of 5 | Jan 2023 Q5, Jan 2025 Q4(b)"). This is the per-topic drill-down that single line deliberately doesn't do. No fresh PDF extraction was done directly for this document; the update pass instead cross-checked against this same directory's codex-deep-dive-hidden-quadratic-substitution.md, which did independently re-extract Oct 2019 Q8's full mark scheme from the primary PDF (WMA11_01_1910_MS/ER) and found one further, adjacent sighting (Oct 2022 Q4) — both folded in below with that attribution. Not affiliated with or endorsed by Pearson Edexcel.

#SIGNAL Topic Master Brief — Hidden Quadratic Substitution (spec 1.1 / 1.5 / 1.6)

#Frequency: seen in 2 of 5 series checked

The facts bank fully mined five WMA11 examiner reports question-by-question: Jan 2019, Oct 2020, Oct 2021, Jan 2023, Jan 2025. This specific skill — an equation that isn't visibly quadratic until a substitution (most often $p = a^x$) is made — is confirmed in 2 of those 5: Jan 2023 Q5 and Jan 2025 Q4(b). The other three fully-mined series show no confirmed instance of it: Jan 2019's own indices question (Q2) was checked and found to be a plain laws-of-indices rearrangement, not a hidden quadratic, and is correctly excluded here rather than padding the count. Two further series (Oct 2022, Jan 2024) were confirmed as genuine mark schemes but never mined question-by-question, so this topic's presence or absence in them simply isn't known — a gap, not a claim of absence.

#A third sighting, found a different way

A separate, wider search of a local past-paper archive — run for a different lesson's audit (surds-manipulation-and-exact-answers.ts, grepping for "surd"/"rationalis"/"denominator"/"conjugate", not for this topic) — turned up Oct 2019 Q8: $x - 6x^{1/2} + 4 = 0$, a hidden quadratic in $\sqrt{x}$ rather than in $a^x$, spec 1.5/1.6. This is a genuine, verified sighting, but it's kept separate from the "2 of 5" above on principle: it wasn't caught by a systematic per-topic search of that series, so its absence from every other unsearched series in that archive can't be read as confirmed absence the way the five fully-mined reports can. The honest total is "at least 3 confirmed sightings across the material checked so far," not a clean "3 of N."

#Mark-tariff pattern

Two sightings now have a fully-verified tariff. Jan 2023 Q5 (exponential sub-case): 6 marks total, split 3+3 across two parts — "show that [the equation] can be written as [a quadratic in p]" (B1 M1 A1*), then "hence solve" (M1 A1 A1). Oct 2019 Q8 (root sub-case): 5 marks in one part — M1 A1 for solving the quadratic in $u=\sqrt x$ to both roots, then M1 A1 A1 for squaring back to $x$ and stating both answers fully simplified — per this technique-hub's own CODEX deep-dive on this topic, which re-extracted it directly from the primary mark scheme (WMA11_01_1910_MS) after this bank's own text had flagged that tariff as uncaptured. Jan 2025 Q4(b) is confirmed as a "(b)" part of the same two-part "show, then hence" shape as Jan 2023 Q5, carrying the same show-working instruction, but its own mark-by-mark breakdown still wasn't captured in this research pass — only its two named traps were. A targeted (not systematic) search of Oct 2022, a series never mined for this topic, found one further, adjacent sighting (Q4(b), a fractional-exponent substitution into an already-solved cubic, 6 marks split 4+2) that is real but excluded from this brief's frequency count, since the equation it substitutes into is cubic, not quadratic. Across every real tariff seen — two counted, one adjacent — no two split marks the same way; the one safe generalisation is that converting back to the original variable always carries its own separately-creditable mark(s), never folded silently into the mark for solving the quadratic in the substituted letter.

#The recurring angle — and a genuine split into two flavours

The three sightings split into two distinct families, not one. Exponential substitution (Jan 2023 Q5, Jan 2025 Q4(b)): an equation mixing terms like $a^x$, $a^{x+k}$, $a^{x-k}$, solved via $p=a^x$. Both share the same trap, worded almost identically in both reports: candidates solve the quadratic in $p$ correctly and stop, never converting back to $x$. Jan 2023 Q5 additionally documents a second trap unique to this flavour — misreading an additive exponent shift as an operation on $p$: "$3^{x+2}=p^2$" in place of the correct $9p$, and $3^{x-1}=p^{-1}$ in place of $p/3$. Root substitution (Oct 2019 Q8): an equation linear in $x$ and $\sqrt{x}$, confirmed to share the "stops at the substituted variable" trap with the exponential flavour — the report separately records candidates who "stopped at finding the roots $3\pm\sqrt5$, thinking this was the answer" — and to add two traps of its own the exponential flavour has no equivalent for: squaring term-by-term, and a final-answer surd-simplification requirement (both detailed in the trap set below). These are genuinely different failure profiles; treating all three sightings as one undifferentiated pattern would overstate what's confirmed.

Worth naming so it isn't confused with this topic: Jan 2019 Q9 also requires spotting a "hidden" quadratic (there, inside an equation needing a fraction cleared before a discriminant question can even start) — but that's discriminant-with-a-parameter.ts's own topic, already covered by that lesson's SIGNAL brief, and isn't counted toward this document's frequency figures.

#Confirmed trap set

• Stops at the substituted variable (Jan 2023 Q5, Jan 2025 Q4(b), and — per the CODEX deep-dive's own re-extraction — Oct 2019 Q8, whose report separately records candidates who "stopped at finding the roots $3\pm\sqrt5$, thinking this was the answer") — the single most-confirmed failure on this topic, now seen across both flavours, not just the exponential one.
• Exponent shift misread as an operation on p (Jan 2023 Q5): $a^{x+k}$ treated as if it squared/powered $p$, rather than $p \cdot a^k$; $a^{x-k}$ treated as $p^{-k}$-like, rather than $p / a^k$.
• Calculator-only working, no visible method (Jan 2025 Q4(b), and the Jan 2023 Q5 mark scheme's own explicit rule: "answers just written down scores M0") — both series carry a printed "show all stages of your working" instruction attached to the question.
• Squares (or otherwise clears) the equation term-by-term instead of as one whole expression (Oct 2019 Q8, per the CODEX deep-dive's re-extraction) — the examiner report's own words: a "frequent mistake" that "immediately led to the loss of all marks." Root-substitution-specific; there is nothing to square in the exponential flavour.
• Final answer left unsimplified in surd form (Oct 2019 Q8) — distinct, root-substitution-specific; not yet confirmed on either exponential-substitution sighting.

#If this topic appears: what to expect

Expect a two-part structure — "show the equation can be written as [a quadratic]," then "hence solve" — most confirmed as an $a^x$ substitution with mixed exponent shifts, worth 6 marks split 3+3 on the most-verified series (Jan 2023 Q5). The single biggest, most-repeatable mark-loss risk is finishing at the quadratic-in-$p$ stage rather than converting back to $x$ — confirmed on both flavours of this topic, not just the exponential one; the second is misapplying exponent laws to the shifted terms while setting up the substitution. A less-frequently-sighted but now fully mark-scheme-verified alternative substitutes a root ($\sqrt{x}$) instead of a power (5 marks, Oct 2019 Q8), where the additional, distinguishing risks are squaring the equation as anything other than one whole expression, and leaving the final surd unsimplified. Either way, showing algebraic working — not a calculator-only answer — is explicitly required and explicitly penalised when absent.