Mock Paper Protocol

Provenance: written 2026-09-29, against research/veridian/WMA14-verified-facts.md (read in full — §2 assessment structure, §3 formula-booklet contents, §4 mark-scheme conventions, §6 series-review breakdown, §8 flagged issues) and the standing WMA14 technique-hub documents already in this folder — forge-exam-technique-guide.md, signal-frequency-and-boundaries.md, forge-battle-card.md, codex-formula-and-method-bank.md, forge-exam-day-protocol-extended.md and forge-forty-eight-hour-checklist.md specifically (each read in full so this document does not re-derive what they already cover), plus content/veridian/question-bank/wma14.ts (237 lines, 2 real questions, both October 2024), also read in full. This is a shared, actively-written folder — the technique-hub directory grew from 25 to 33 pre-existing files during this session as concurrent batch sessions added their own WMA14 documents, so the exposure-audit grep below was re-run against the folder's state at the end of this pass, not its state at the start; §2's series table and this document's own core finding (the January 2025 series) are unaffected by that growth. A direct grep -irE series-mention audit ran against content/veridian/curriculum/wma14/*.ts (12 files) and content/veridian/technique-hub/wma14/*.md (33 files, every file that predates this one, counted at time of writing) for §3's exposure table — methodology stated there, fully reproducible, though a rerun later in this batch will likely return different counts as more sibling documents land. This pass also independently located and verified a genuine WMA14 series not named anywhere in WMA14-verified-facts.md's own series list, nor in any sibling technique-hub document checked above: January 2025 (WMA14/01, Publications Code WMA14_01_2501_MS, Question Paper Log Number P76196A), a real Pearson question paper and mark scheme carrying the PMT (Physics & Maths Tutor) watermark, held locally at /Users/neosteinhoff/Documents/SOL/Papers/IAL_Paper4_Complete/IAL_Materials/Mathematics/January_2025/ (the raw PDFs — the cover page was opened and visually checked directly this session: genuine Pearson branding, the real barcode, "Tuesday 21 January 2025", "©2025 Pearson Education Ltd.") and as a text conversion at /Users/neosteinhoff/Documents/SOL/Papers/IAL_Paper4_Complete/Master_Consolidated_Docs_U4/Mathematics_Unit4_Master_{QP,MS}.md, both read in full. This location is outside this repository — it is this machine's personal exam-archive folder, not research/veridian/ — so §2 below flags it as a gap in that other file rather than silently folding it in. No claim below about this series is estimated; every mark tariff is read directly off the real question paper's own printed (n) marks and cross-checked against the mark scheme's per-part totals, which sum to the paper's own stated 75.

#WMA14 (Pure Mathematics 4) — Mock Paper Protocol

#What this document is, and isn't

This is the protocol for running a real, timed WMA14 mock and self-marking it honestly — sourcing, sequencing, timing, and post-sitting audit. It does not repeat material that already exists elsewhere in this technique hub:

• forge-exam-technique-guide.md §1 already gives the real per-question tariff table and explains why a flat per-mark average is the wrong pacing tool for this paper. §2 already defines the M/A/B system, dM1/ddM1, and the "quote the formula first" rule with verbatim sourcing. This document adds the order of operations for applying those definitions to your own script (§6 below), not a restatement of them.
• forge-battle-card.md §4 already carries a ranked top-5 danger list, built by counting how many independent topics/series each error pattern is confirmed across. §7 below uses that ranking directly rather than inventing a new one.
• signal-frequency-and-boundaries.md §2 already confirms, and explains why, no real WMA14 grade boundary exists anywhere in this course's research. Calibrate a mock score against that document's own honesty on this point — never against a number invented here.
• content/veridian/question-bank/wma14.ts already carries two fully marked-up real questions (both October 2024) for self-marking drill practice at line-by-line resolution. This document tells you when those two questions are usable and when they're already spent (§3 below).
• forge-exam-day-protocol-extended.md already covers reading-time use, question-order strategy, and a derived timing-checkpoint table for the live exam sitting itself, plus a "running behind" playbook. That document is about exam day; this one is about the mock as a rehearsal-and-diagnostic tool beforehand — §5 below asks you to build your own per-question timing log from a real sitting, which is different from, and a genuine input into, that document's checkpoint table.
• forge-forty-eight-hour-checklist.md already tells you when, in a revision countdown, to run or avoid a full past paper. It assumes the how-to described here.

#1. What counts as a real WMA14 mock

A WMA14 mock is a genuine past series' question paper, sat under real conditions, marked against that same series' genuine mark scheme — a real publication code on the cover (WMA14_01_...), not a retyped or "reconstructed" practice paper. A generic "Pure Maths 4 practice test" from a third-party bank doesn't carry Pearson's own M/A/B/dM1/cso/isw conventions, and self-marking against anything else teaches you to grade yourself by a system you won't be marked against on the real day.

#2. The real series-availability picture

WMA14-verified-facts.md §6 states this plainly: 11 series with at least one genuine primary document reviewed — January 2021, June 2021 (provisional), June 2022, October 2022, January 2023, June 2023, October 2023, January 2024, June 2024, January 2026, plus October 2021 (examiner's report only, no mark scheme obtained). Only 4 of those have a full Examiner's Report: January 2021, October 2021, October 2022, January 2024. Only 3 have a full question paper verified in that file: June 2022, October 2022, January 2026.

"Not confirmed" means this course's own facts bank does not state a full question paper was obtained for that series — it does not mean Pearson never published one. If you can source the real question paper yourself (Pearson's own past-papers portal, or a checked mirror such as PMT) for January 2023, June 2023, October 2023 or June 2024, you can still self-mark it against the mark-scheme wording this course's research already verified.

#A genuinely new series found this session: January 2025

No document in this course's WMA14 research — not the facts bank, not FORGE, not SIGNAL, not the battle card — names January 2025 (WMA14/01, Publications Code WMA14_01_2501_MS, Question Paper P76196A). It exists as a real Pearson PDF pair (question paper and mark scheme; no examiner's report found alongside it, matching the same publication-lag pattern already documented for January 2026) in this machine's personal exam-archive at the path given in the provenance note above — not inside this repository. That is itself worth flagging plainly: this is a real, usable mock paper this course's own verified-facts document doesn't yet know about.

Its full structure, read directly off the real paper and cross-checked against the real mark scheme's own per-part totals (they sum to 75, matching the cover page):

Two things worth reading directly from this table, both new evidence rather than restated evidence:

• The calculator-technology-banned rubric appears on 4 of this paper's 9 questions (Q1, Q4, Q7, Q9) — not the single occurrence the facts bank's existing anchor (January 2026 Q7) suggested. This is a second, independent real series confirming the same pattern battle-card danger #5 already names, and it meaningfully strengthens that pattern's evidence base rather than merely repeating it: on this paper, read the rubric line for every question before assuming your calculator's numeric answer is enough.
• The tariff spread inside one real paper is 3× — Q6 (4 marks) next to Q8 and Q9 (12 marks each) — concretely demonstrating, with a fresh series, the same point forge-exam-technique-guide.md §1 already makes in the abstract about why a flat per-mark average misleads on this specific paper.

This finding is not reflected in WMA14-verified-facts.md, which should be updated in a future research pass to add this series properly (with its own topic-by-topic citations) rather than left standing only in this protocol document — flagged here, not silently fixed, matching this course's own stated practice for the Oct 2024/§6 inconsistency above.

#3. Choosing which series to sit, and in what order

A mock only measures live performance if you sit it cold. This pass ran the same reproducible check as the WMA13 protocol: grep -irE for each series' month/year (abbreviated and full) across the 12 files in content/veridian/curriculum/wma14/*.ts and separately across the 33 files in content/veridian/technique-hub/wma14/*.md that predate this one (counted at time of writing — see the provenance note above on this being a shared, growing folder).

Read this as a proxy, not a precise measure — a raw substring count picks up a passing citation the same way it picks up a full worked exemplar, and the technique-hub side of this count will keep climbing as more of this batch's sibling documents land in the same shared folder. It is, however, real and reproducible at any given moment: re-run the same grep over the same files and you get the same numbers for that moment.

January 2025 has zero exposure anywhere in this course's own WMA14 material — sit that one first, while it is provably unspoiled, using the topic breakdown in §2 above only to check you've covered every spec point, never to preview the actual questions. After that, June 2021, January 2023, June 2023 and October 2023 are the next-least-exposed — but per the table in §2, this course's own archive holds only mark schemes for the January/June/October 2023 series and only a provisional mark scheme for June 2021 (itself flagged as a disrupted-assessment series), so sitting any of those as a timed mock means sourcing the real question paper yourself first.

Save January 2021, October 2021, October 2022, January 2024 and January 2026 for open-book review or re-marking practice rather than a blind timed sitting — every one of them now exceeds 220 combined mentions across this course's own lessons and technique-hub documents, and January 2021 and October 2021 specifically exceed 295 each.

One specific exception: October 2024's Q1 and Q2 are reproduced as complete, fully marked-up real questions — question text, mark scheme, and examiner-report commentary — in content/veridian/question-bank/wma14.ts. If you've drilled that file, those two questions specifically are spent for cold-sitting purposes regardless of the series' own exposure score; the rest of that paper (marks 3–9, if fully sourced) is unaffected.

#4. Before you start: replicate the real conditions

Grounded directly in WMA14-verified-facts.md §2–§3, independently confirmed on the real January 2025 cover page read this session:

• 90 minutes, 75 marks, every question compulsory. There is no section to skip or reorder around.
• The question count is not fixed — confirmed at 9 (June 2022, January 2026, and now January 2025) and 11 (October 2022). Don't assume a specific count going in; read the cover page's own stated count and total before planning.
• Calculator: an ordinary scientific calculator only. Appendix 6 (spec pp. 85–86) bans a calculator with "the facility for symbolic algebra manipulation, differentiation and integration" — no CAS. This spec has no separate calculator/non-calculator section anywhere — confirmed two independent ways in the facts bank (§2) — so a mock never needs a non-calculator mode; it needs the question-level rubric check below instead.
• Formula sheet: only the real Mathematical Formulae and Statistical Tables booklet (Issue 2, January 2021). It is cumulative across P1–P4 — the general binomial expansion for rational $n$, the integration-by-parts formula, and the $\sec x$/$\csc x$ integrals are all given, not memorised. The spec's P4.2.3 list of what must be memorised (and will not appear in the booklet) contains exactly one item for this paper: the scalar-product-in-components formula. The volume-of-revolution formula ($\pi\int y^2\,dx$) sits in a genuinely ambiguous middle ground — not on the must-memorise list, not in the booklet's P4 section either, only descriptive spec guidance (facts bank §8) — so write it down from memory for a mock regardless of its unresolved formal status; no examiner report reviewed has ever flagged a candidate for not knowing the formula itself, only for misapplying it once written.
• Watch for the per-question calculator rubric before you start each question, not only when it's visually obvious. On the real January 2025 paper, four of nine questions (§2 above) carry their own printed instruction overriding the paper's general calculator-allowed policy. Read every question stem before answering it, exactly as in the real exam — don't assume a rubric only shows up on the paper's biggest question, which was true of the one existing anchor (January 2026 Q7) but is not the general rule.

#5. Timing the sitting

forge-exam-technique-guide.md §1 already gives the real planning average — 1.2 minutes per mark — and the caution that it's the wrong tool alone, since question count and per-question tariff both vary. The January 2025 paper is fresh, real evidence for exactly that caution: a 4-mark proof question (Q6) sits on the same script as two 12-mark questions (Q8, Q9) — a 3× spread the flat average hides completely. Don't re-derive the tariff table here; use it, and read the sat paper's own printed marks before committing to a plan.

What's specific to a mock rather than exam day is generating your own timing data to audit afterward:

1. Before starting, write your own per-question time budget in the margin, using each question's own printed tariff × 1.2 minutes, as a running cumulative clock time ("Q1 by 0:06, Q2 by 0:16…") rather than a countdown — a glance at a wall clock then tells you instantly whether you're ahead or behind.
2. Check the clock only at question boundaries. A mid-question glance measures anxiety; a between-question glance produces the number §7 below actually needs.
3. No per-question alarm. The real exam has no bell between questions; an alarm trains a reflex the exam doesn't reward.
4. If a question is running well past its budget, note where you are and decide deliberately — at the next boundary, not mid-question — whether to push on or move on.

#6. The moment time is called

• Stop writing exactly at 90 minutes. Finishing "just this line" corrupts the timing data the mock exists to produce, and that option isn't available on the real day either.
• Mark, in the margin, exactly where you were before doing anything else — the raw material for telling a timing failure apart from a technique failure once you mark.
• Set the script aside before opening the mark scheme. Marking from memory of what you meant to write marks a different candidate than the one who actually sat the paper.

#7. Self-marking honestly

forge-exam-technique-guide.md §2 already defines M/A/B, dM1, cso, isw, ft and "quote the formula first" in full, with verbatim sourcing — this section is the order of operations for applying those definitions to your own script without quietly being generous to yourself.

1. Mark every M line first, straight through the whole script, before any A mark. A marks cannot stand without their M mark — marking A first on a line that "looks about right" risks crediting an answer you recognise rather than checking the method mark it actually depends on.
2. For a dM1 line, count how many independent method marks genuinely precede it in your own script — don't assume the depth from a different question. A dependent mark is a distinct, separately-checkable stage, not folded into the mark before it.
3. Check the whole line for a cso mark, not just the final number. The convention is explicit: no errors anywhere in that part of the question — a correct final value reached via a script with an earlier slip doesn't earn a cso mark just because the number is right.
4. Know the difference between a genuine misread and an incomplete method — they're marked differently, and self-marking that conflates them is either too harsh or too generous. The general misread rule (facts bank §4): for a misread that doesn't change the character of the question, deduct two from the A/B marks gained in the affected part, not the whole part — forge-exam-technique-guide.md §2 already cites the real anchor for this (October 2024 Q1, a misread binomial-index sign, capped at B1M1A0A0 rather than 0/4). That is a different failure mode from an incomplete method: October 2024 Q1(b)'s own commonWrongPath in question-bank/wma14.ts shows a candidate who correctly substitutes into their expansion but never takes the required reciprocal — that scores M0 because the method itself was never finished, not because anything was misread, and it costs only the two marks in that part; the independent B1/M1/A1/A1 already earned in part (a) stand untouched either way.
5. Two tiers of "what to do with crossed-out work" both apply, and they're complementary, not contradictory. The maths-specific rule (facts bank §4): if all but one attempt is crossed out, mark the one that isn't; if all or none are crossed out, mark every attempt and keep the best. Independently, the January 2025 mark scheme's own general marking guidance states the same underlying default from the other direction — crossed-out work still gets marked unless you've actually replaced it with something else. Read together: never assume crossing something out erases it from consideration unless a real replacement stands in its place.
6. Re-check every rubric-flagged question against what your script actually shows, not what you know is true. If a question carried the calculator-technology-banned rubric (§4 above), any mark chained to a step your script shows only as a bald final value is wrong regardless of whether the value itself is correct — this is exactly what the real October 2024 Q2 mark scheme does: a calculator-only jump to the roots scores M0, and the mark scheme states plainly that "if the roots are just stated then M0A0 follows." This is the easiest place to over-credit yourself, because you already know your number is right.
7. Keep exact values exact. Facts bank §4, verbatim: marks are "normally lost" where the candidate "resorts to using rounded decimals" and an exact or surd answer was required — independent of whether the method itself was right.
8. Don't penalise yourself for skipping working that could genuinely be done in your head — but treat this as a narrow, named exception (facts bank §4), not a general licence. Most real scripts show working; an "occasional awkward case" is exactly that, occasional.

For topic-specific traps on the actual spec areas your sat paper tested, use forge-exam-technique-guide.md §3's checklist boxes and codex-formula-and-method-bank.md Part 2's method skeletons — don't re-diagnose from scratch what's already named there.

#8. What to check first, before you trust a total

Mark the whole script (§7), then audit in this order:

1. Write your score question-by-question, not just a final total. A single number hides whether marks were lost in one place or spread across the paper.
2. Check forge-battle-card.md §4's ranked top-5 danger list, in its existing rank order, for every question you dropped marks on. That ranking is built from how many independent topics/series each pattern is confirmed across, so it is already the cheapest, highest-yield check to run first — don't build a new ranking here.
3. Only then, check forge-exam-technique-guide.md §3's topic-specific checklist for the actual spec areas your sat paper tested.
4. Calibrate the resulting total against real grade boundaries only if you find a primary Pearson boundary document yourself. signal-frequency-and-boundaries.md §2 confirms none exists anywhere in this course's own WMA14 research — do not estimate one, and do not treat a single mock's raw score as more precise than that gap allows.
5. Cross-check your §5 timing log against where marks were actually lost. A question you ran well over budget on and also lost marks on points at a pacing problem; a question finished on time that still lost marks points at a technique gap §8.2–8.3 should already have caught.

#9. Logging the result

Keep this minimal and consistent across mocks. Per mock, record:

• Date sat; series used, and whether it was genuinely cold (§3) or already exposed.
• Raw score, question-by-question, out of the paper's real total.
• Which of forge-battle-card.md §4's ranked dangers recurred, and on which question(s).
• Which named topic-specific trap from forge-exam-technique-guide.md §3 recurred, and on which question(s).
• Per-question time budget vs. actual, from the margin notes made during the sitting (§5).

A single occurrence of a pattern is not yet evidence of anything; the same pattern recurring across two or more separate mocks is real, specific evidence about where revision should go next.

#Flagged gaps

• January 2025's discovery is not yet reflected in WMA14-verified-facts.md. This protocol confirms its structure and mark tally directly, but a proper topic-by-topic citation pass (matching the depth §5 of that file gives every other series) has not been done and should be, in a future research pass — flagged here rather than silently patched into a file this document didn't set out to rewrite.
• No examiner's report was found for January 2025, matching the same publication-lag pattern already documented for January 2026 — self-marking that series draws on mark-scheme wording only, never report commentary.
• June 2021 is a disrupted-assessment series per the facts bank's own characterisation (§6) — treat a mock sat on it with caution; it may not reflect standard full-paper pacing or structure.
• For January 2023, June 2023, October 2023 and June 2024, this course's own archive verified a mark scheme but not a full question paper. A mock sitter needs to source the real question paper independently (Pearson's own past-papers portal, or a checked mirror) before this course's mark-scheme-derived facts can be used to self-mark it.
• The exposure table in §3 is a proxy built from this course's own materials, not a claim about what any individual student has personally read — a student who has skipped the lessons and technique-hub entirely would find no series contaminated by this course's own writing, regardless of the counts above.
• No grade-boundary primary source exists anywhere in this course's WMA14 research. Restated from signal-frequency-and-boundaries.md §2 because it matters enough to repeat: never estimate one.

#Sources

• research/veridian/WMA14-verified-facts.md — §2 (assessment structure, calculator policy, AO mark ranges), §3 (formula-booklet contents, the P4.2.3 memorise list, the volume-of-revolution ambiguity), §4 (mark-scheme conventions, verbatim — misread rule, crossed-out-attempts rule, cso/isw, exact-answer rule, "in your head" exception), §6 (series/document-count breakdown), §8 (flagged issues — narrower evidence base, no January 2026 report, October 2020's dead mark-scheme link).
• content/veridian/technique-hub/wma14/forge-exam-technique-guide.md — §1 (tariff table and timing caution), §2 (mark-code definitions), §3 (topic-specific checklists), §4 (fudged-reverse-fit cross-topic pattern).
• content/veridian/technique-hub/wma14/forge-battle-card.md — §4 (ranked top-5 danger list, reused directly rather than re-ranked).
• content/veridian/technique-hub/wma14/signal-frequency-and-boundaries.md — §2 (confirmed absence of any real grade-boundary source), §3 (the Oct 2024/§6 inconsistency this document repeats rather than resolves).
• content/veridian/technique-hub/wma14/codex-formula-and-method-bank.md — Part 2 (method skeletons, referenced for topic-specific self-marking, not restated).
• content/veridian/question-bank/wma14.ts — read in full (237 lines, 2 real October 2024 questions); source for the October 2024 Q1(b) misread example and the Q2 calculator-only-jump example in §7, and for the "already spent" caveat in §3.
• content/veridian/curriculum/wma14/*.ts (12 files) and content/veridian/technique-hub/wma14/*.md (33 pre-existing files, counted at time of writing in this shared, actively-written folder) — direct grep -irE series-mention audit for §3's exposure table (methodology stated in full there, reproducible).
• content/veridian/technique-hub/wma14/forge-exam-day-protocol-extended.md and forge-forty-eight-hour-checklist.md — read in full for scope-boundary purposes (exam-day pacing and revision-countdown timing, respectively; this document is about the mock itself, not either of those).
• January 2025 (WMA14/01), verified directly this session: question paper and mark scheme PDFs at /Users/neosteinhoff/Documents/SOL/Papers/IAL_Paper4_Complete/IAL_Materials/Mathematics/January_2025/ (cover page opened and visually checked: genuine Pearson Edexcel branding, "Tuesday 21 January 2025", publication code P76196A, PMT/Physics & Maths Tutor watermark, real barcode) and their text conversions at /Users/neosteinhoff/Documents/SOL/Papers/IAL_Paper4_Complete/Master_Consolidated_Docs_U4/Mathematics_Unit4_Master_{QP,MS}.md, both read in full — source for §2's topic/mark table, the four-rubric-questions finding, and the crossed-out-work general-guidance wording in §7. This location is outside the current repository.

Not affiliated with or endorsed by Pearson Edexcel. Every quoted fragment above is under fifteen words with its own attribution; all else is paraphrase or this course's own synthesis of material verified against a named primary or previously-verified source.