Examiner Intelligence

Provenance: compiled 2026-09-29. Source is research/veridian/WME01-verified-facts.md in full (§1–§9, including the per-lesson audit log's own primary-source re-extractions) plus content/veridian/technique-hub/wme01/forge-exam-technique-guide.md (the fresh, full deep-read of the January 2025 mark scheme and examiner report, and of the June 2025 mark scheme, that closed gaps the facts bank itself had left open) — every quote below was already matched against a real Pearson PDF by one of those two documents before this pass; no new primary-source PDF was fetched here. This document was written only after reading the rest of this folder (codex-formula-and-method-bank.md, signal-frequency-and-boundaries.md, forge-battle-card.md, forge-forty-eight-hour-checklist.md, forge-exam-day-protocol-extended.md, forge-error-pattern-tracker.md, forge-vocabulary-precision-list.md, forge-mock-paper-protocol.md, all nine codex-deep-dive-*.md files and all eight signal-topic-brief-*.md files), the nine lesson files in content/veridian/curriculum/wme01/, and content/veridian/question-bank/wme01.ts — specifically to avoid restating any of them. forge-vocabulary-precision-list.md is the closest neighbour: it catalogues exact words and mark-scheme codes a candidate must get right. This document asks a different question — not which words a candidate must use, but which words and rhetorical moves the report itself habitually reaches for, series after series, and what that habit implies about how a real script is actually read. forge-error-pattern-tracker.md's Category 11 (method-choice efficiency) is the other near neighbour; §1.4 below is deliberately scoped away from repeating its citations and instead asks what the report's willingness to be blunt about causation, sitting next to its refusal to be numerically precise about frequency, says about the report's own voice. Direct quotations below are kept under roughly fifteen words each, with the series and question named; anything longer is paraphrased. No likelihood percentage or grade boundary is stated anywhere below — where a source itself states a number or a superlative, it is quoted as that report's own words about that exact question in that exact series, never converted into a course-wide or future-series claim. Not affiliated with or endorsed by Pearson Edexcel.

#WME01 — Examiner Intelligence

#Pearson Edexcel IAL Mechanics 1 — how the report itself talks

#0. What this is

Every other document in this folder tells you what the examiners found: a trap, a mark allocation, a word whose exact meaning costs a mark if blurred. This one asks a narrower question — across the series this course has actually read, does the report's own language repeat? Not the mechanics itself, and not the vocabulary a candidate has to get right (that is forge-vocabulary-precision-list.md's job) — the phrasing, the emphasis, and the moves a report habitually makes that a mark scheme alone would never show.

The answer, checked below, is yes, in four countable ways, each grounded in at least two independently-named series. Two further findings are real and quotable but surfaced only once in the material this pass could check — those are labelled SINGLE FINDING and left that way rather than dressed up as a trend.

Denominator, stated up front (matching this folder's own convention): the facts bank confirms 17 series with at least one genuine primary document (mark scheme and/or examiner report), and an examiner report specifically reviewed for 11 of those — Oct 2019, Jan 2020, Jan 2021, Jan 2022, Oct 2022, Jan 2023, Jun 2023, Oct 2023, Jan 2024, Jun 2024, and Jan 2025 (facts bank §5). Jan 2021's report was downloaded but never deep-mined; Jun 2023's and Jan 2024's reports were read but no quotable report-language from either surfaces anywhere in this course's research to date; Jun 2025's examiner report has not even had its existence checked (only its mark scheme has). This document draws quoted report language from 8 of those 11 reviewed series: Oct 2019, Jan 2020, Jan 2022, Oct 2022, Jan 2023, Oct 2023, Jun 2024, and Jan 2025 (the last one via the FORGE guide's own fresh full read, deeper than the facts bank's original partial mining of it). A pattern below naming four of those eight series is a real plurality of the quoted evidence, not a cherry-picked pair.

---

#1. Confirmed patterns

#1.1 A small, reused set of hedge words does almost all the frequency-reporting work

A WME01 report almost never gives a number for how many candidates did something. It reaches for the same handful of graded-but-unmeasured quantifiers, and the same words recur, series after series: "the most common mistake" (Oct 2019 Q3), "the most common error" (Oct 2022 Q3; Jan 2025 Q1(a), per FORGE), "a common error" (Jan 2023 Q1; Jan 2025 Q7(b)), "many candidates" / "many failed to" (Oct 2022 Q4; Jan 2023 Q5), "a significant number" (Jun 2024 Q7; Jan 2025 Q6(c)), "some candidates" / "some responses" (Jan 2020 Q3; Jun 2024 Q3), "a relatively small proportion" (Jan 2025 Q1(b), per the question bank's own quotation of the report), and, at the extremes, "a very small number" (Jun 2024 Q5) and "a substantial number" (Jun 2024 Q6). Eight independently-dated series, one shared lexicon of soft quantifiers standing in for a statistic no report actually gives. WMA12's own sibling document in this course found the identical habit for Pure Mathematics 2 — worth noting as a possible house style of this exam board's report-writing generally, not an artefact of which WME01 series this course happened to read, though this document only claims it for WME01's own eight series above.

#1.2 When an error is named, the report increasingly favours quoting the literal wrong value over describing the error type

Distinct from 1.1's frequency-hedging, this pattern is about how a wrong answer itself gets reported. Rather than saying "candidates made a sign error" or "candidates misread the tensions," a WME01 report frequently reproduces the actual, specific wrong number or expression a candidate wrote. Jan 2020's Q2 report doesn't describe "confusion about the tension relationship" in the abstract — it lists the literal pairs seen: "T and 20T, T and 20, or T and T." Jan 2022's Q1 report doesn't say "weight was confused with mass" and stop — it names the resulting wrong term directly: "resulting in the use of 5g." Jun 2024 supplies two more in the same series: Q2's report names the literal wrong angles substituted into the cosine rule — "30, 60, or in some cases 330 degrees" — and Q7's report names both a literal wrong vector, "an answer of 12i + 16j," and a literal wrong fraction, "24/10 rather than 10/24." Jan 2025 continues the habit on two separate questions: Q1(b)'s report is quoted (via the question bank) naming the exact wrong pair "p=19 and q=−16"; Q6's report (via FORGE) is described as recording "0.75 instead of 0.75mg." Four independently-named series — Jan 2020, Jan 2022, Jun 2024, Jan 2025 — each supplying at least one instance of this same reporting habit: don't just name the mistake's shape, quote the number.

#1.3 The report occasionally interrupts its own retrospective account to state, directly, what candidates needed to understand — using one of a small family of stock verbs

Most report prose is purely retrospective ("candidates did X"). A distinct, smaller register breaks that pattern to state a requirement directly, and it reaches for the same small verb family each time: "appreciate" or "be reminded." Oct 2019's Q6 report uses both in the same document: "Candidates need to appreciate that the question refers to 'the model'" for its modelling sub-part, and, separately, on the same question's quadratic-solving sub-part, "candidates need to be reminded to show working when solving quadratics." Oct 2022's Q1 report repeats the second construction on an unrelated topic (momentum, not modelling), stating candidates "need to be reminded that the final answer needed to be positive" — because the question asked for a speed. Jan 2023 repeats the first construction, twice, on two different questions in the same series: Q4's report states "it was necessary to appreciate that 'about to tilt' implied" that one of the two reactions had to be taken as zero, and Q5's report states "many failed to appreciate that they needed to consider two stages of the motion." Three independently-named series — Oct 2019, Oct 2022, Jan 2023 — each supplying at least one instance of the report stepping out of pure description to name, in this same small vocabulary, what a candidate was required to understand rather than merely what a candidate did.

#1.4 The report hedges on how many, but not on why — comparative and causal claims about method choice are stated flatly, not softened

Set 1.1 next to this one and the contrast is the finding: a WME01 report that will not commit to a number for how many candidates erred will commit, without any hedge at all, to a causal or comparative claim about why one method fared worse than another. Jan 2022's Q1 report states plainly that "more efficient methods... were successfully employed" for an alternative approach, with no qualifier softening the comparison. Jun 2024's Q2 report is equally direct about the opposite case — the cosine-rule route "tended to have more success" than resolve-and-Pythagoras, again unhedged. Jun 2024's Q4 report on a moments question goes further, stating a flat causal claim rather than a mere correlation: taking moments about the wrong point "inevitably led to more errors being made." Oct 2023's Q1 report makes the same kind of unhedged causal claim independently, on the same topic: "where two moments equations were used, there was more opportunity for these errors." Three independently-named series — Jan 2022, Jun 2024, Oct 2023 — each stating, without a single softening word, that one credit-worthy method produces worse outcomes than another. The report's hedging is selective: it is reserved for candidate counts, not for the report's own judgements about method and consequence.

---

#2. Single findings — real, but not (yet) a pattern

• The report explains its own marking decision, not just candidate behaviour, when a numerically "correct" answer still earns nothing. Oct 2019's Q3 report, on a car-and-trailer question, states that omitting the weight components from both particles' equations "fortuitously led to a correct value of D but this received no credit," since the terms were missing from both equations. This is the report speaking about why its own scheme withholds a mark from a right-looking number — a distinct register from every pattern in §1, and closer to the report justifying a design decision than describing what a candidate did. No second series in the material this pass could check makes this exact move (a coincidentally-correct final value explicitly denied credit, with the coincidence itself named); it may recur, but this pass found it once.
• Exactly one series in this reviewed set abandons the hedge-word register entirely for a genuine superlative statistic. Every other named finding in this document sits inside §1.1's soft-quantifier norm. January 2025's own report on that series' final question breaks from it outright, calling the question "by far the most demanding, with a modal mark of 0" (quoted via forge-exam-technique-guide.md). That is a real statistic-shaped claim (a modal mark, not a hedge word) sitting in a single named series, with no second instance of this specific register-break confirmed anywhere else in this course's WME01 research to date.

---

#3. What this implies for how you write an answer

Briefly, because the technique itself already has a home in forge-exam-technique-guide.md, forge-error-pattern-tracker.md, and forge-vocabulary-precision-list.md — this section only states what follows directly from §1, not a new technique layer:

• Per §1.2: when a report is willing to quote a specific wrong number, that number is worth treating as a named, recurring trap in its own right, not an isolated slip — the report itself is telling you this exact substitution has happened before, in these exact terms.
• Per §1.3: where a report's language shifts from "candidates did X" into "candidates need to appreciate / be reminded," that shift marks a conceptual gap the report considers separable from a computational one — read those specific sentences as the closest thing this paper's record has to the examiners naming, in their own words, the understanding (not the arithmetic) they actually wanted.
• Per §1.4: a report's silence on "how many" chose the harder method is not silence on whether it matters — the reports repeatedly state, flatly, that it does. Where two valid methods exist, don't treat the mark scheme's equal credit as evidence they are equally safe under exam conditions.

---

#4. Series ledger for this document

---

#5. Flagged gaps

• This document draws on 8 of the 11 series with an examiner report actually reviewed by this course, and 8 of the archive's full 17 series with any primary document at all. Jan 2021's report sits downloaded but unmined; Jun 2023's and Jan 2024's reports were read but yielded no quotable line in this course's existing research; a genuinely recurring register habit confined to any of those three would not appear anywhere above, and nothing here claims otherwise.
• §1.3's "appreciate / be reminded" pattern rests on three series, two of which (Oct 2019, Jan 2023) each supply two of its examples; Oct 2022 supplies one. Stated plainly here, as in §1.3 itself: a real three-series pattern, not inflated by treating two heavily-quoted series as three independent confirmations.
• §1.1's hedge-word scale is described qualitatively, not measured — no WME01 report in this pass defines "many" against "some" against "a significant number" with an actual candidate count, so the implied ranking is read from context and ordinary English, not from any source's own stated scale.
• This document does not attempt to measure how often each pattern recurs within a single report — only that each recurs across the named series above. A future pass that deep-mines Jan 2021's already-downloaded report, or finally locates and reads Jun 2025's, could either strengthen every pattern above with a ninth series or surface a second instance of either §2 finding.