Mock Paper Protocol

#WBS13 Mock Paper Protocol

Provenance — written 2026-09-28. Every specific claim below traces to research/veridian/WBS13-verified-facts.md (2,057 lines, built from primary Pearson mark schemes and examiner reports read directly via pdftotext -layout/-raw, either from qualifications.pearson.com or, for two series flagged below, a third-party mirror with its own standing provenance caveat), or to a direct check of this course's own existing WBS13 files: content/veridian/curriculum/wbs13/paper-anatomy.ts, content/veridian/technique-hub/wbs13/forge-exam-technique-guide.md, content/veridian/technique-hub/wbs13/signal-frequency-and-boundaries.md, and content/veridian/question-bank/wbs13.ts. Short quotations (under ~15 words) are attributed inline; longer mark-scheme or examiner-report content is paraphrased, never reproduced at length. No likelihood percentage or grade boundary appears anywhere below — this course's WBS13 sources contain neither (see §7).

#1. What this document is, and what it deliberately doesn't repeat

This is an operational protocol: how to choose a paper, sit it under real conditions, and mark it against this paper's real conventions afterward. It assumes the reader already has the tariff-by- tariff technique in forge-exam-technique-guide.md (command words, level ladders, a tariff-specific self-mark checklist in its §3, and an exam-day pacing list in its §5) and the topic-frequency table in signal-frequency-and-boundaries.md. Neither is repeated here. What follows is new: which paper to actually use for a mock given what this course has already taught from the available past-paper pool, how to log a sitting so the log itself is useful afterward, and a self-marking sequence built around holding yourself to the same holistic standard an examiner report describes, not a generous one.

#2. Choosing a paper: the blindness problem

Twelve WBS13 series have been located and read for this course's research (per signal-frequency-and-boundaries.md §1): Oct 2021, Jan 2022, June 2022, Oct 2022, Jan 2023, June 2023, Oct 2023, Jan 2024, Summer 2024, Oct 2024, Jan 2025, Oct 2025. That reference-work has a real side effect worth being honest about before picking a paper to mock: most of those same series have already been turned into named worked exemplars somewhere in this course's own curriculum, question bank, or technique-hub. Picking "a random real WBS13 past paper" is therefore not equivalent to sitting genuinely unseen content the way it would be for a paper this course had never touched — for several series, the specific company, figures, and even the mark scheme's own indicative-content wording are already quoted in a lesson you may have read.

Checked against this course's own citations (curriculum lessons, the technique-hub, and the question bank — not an exhaustive re-audit of every sub-part of every series, so absence here means "no worked exemplar found in this course's material," not a guarantee of true novelty):

Two practical conclusions follow, both grounded in the table rather than asserted:

1. Oct 2022 and Jan 2025 are the worst choices for a content-blind mock — every part of both has already been built into this course's teaching material. Use them instead as timed self-marking rehearsals: you already half-know the content, which isolates the self-marking process itself (§5–§6 below) from content recall, a genuinely different and still useful exercise — just not a cold sit, and it should not be logged or trusted as one.
2. Jan 2022 and Oct 2024 are the closest this course has to a genuinely unseen full paper, and Oct 2021's Section B/C (Q2 and Q3) are specifically unmined even though its Section A is well covered — worth isolating as a standalone 20-mark Evaluate blind attempt if a full sitting isn't available. Reserve these for an actual timed, blind attempt before working through the more heavily taught series.

Provenance caveat carried forward for anyone printing Oct 2025 specifically: this series was reached only via a third-party mirror, not directly from qualifications.pearson.com — see the research file's own standing note (letterhead, Publications Code convention, and level-descriptor wording independently checked against every confirmed-genuine document). Re-confirm against the official host before treating a mock built from it as beyond dispute if that matters for the sitting's purpose.

#3. Sitting conditions that actually match the real exam

Three real, spec-confirmed facts should shape how a mock is set up, not just how it's timed:

• Two hours, 80 marks, one sitting (spec 3.2, Assessment information). Don't split it across two sessions and call it one mock — the fatigue and pacing pressure across the full two hours is part of what a mock is meant to test.
• Calculators are permitted (spec Appendix 10) — sit with one, as in the real exam.
• The accounting ratios and formulas in Appendix 9 are explicitly "not supplied in the exam" (spec 3.1, verbatim). Sitting with a formula reference sheet open in front of you is a fair way to learn a formula for the first time, but it is not a faithful mock once the goal has shifted to rehearsing the real exam — decide consciously which mode this sitting is, rather than defaulting to "formulas open" out of habit.
• Section A is keyed to a separate Source Booklet of extracts in the real paper. If your copy of a past paper has extracts and questions merged into one continuous document, keep them physically separate anyway (print or fold) — flicking back to a separate booklet costs real seconds per question, the same seconds the timing budget in §4 assumes are being spent.

#4. What to time, and how to log it

A single stopwatch for the whole two hours tells you that you ran out of time; it doesn't tell you where. Write down the actual clock time at the start of each question, not just an internal sense of pacing — a two-second habit that turns a vague "Section A felt rushed" into a specific, fixable line item afterward.

Use the timing budget already established in paper-anatomy.ts and the FORGE guide's §1 rather than re-deriving it here: roughly 60 minutes for Section A's 40 marks, roughly 30 minutes each for Q2 and Q3, on the course's own 1.5-minutes-per-mark heuristic (120 minutes ÷ 80 marks — this course's own arithmetic, not a Pearson-published instruction). The one number worth restating because it's easy to under-budget from habit: Q1(d) and Q1(e) are each a confirmed 12-mark Assess on every series this course has read, not the lighter 10-mark version Assess carries on Units 1/2 — budget roughly 18 minutes each, not the shorter allowance the same command word gets elsewhere.

A minimal log, filled in live during the sitting:

#5. Self-marking: what to check, in order

Mark cold — at least an hour after finishing, not immediately, so the answer is read the way an examiner reads it rather than the way it felt to write it. Work through every answer in this order, not topic-by-topic:

1. Command-word and tariff audit, before reading content. For every question, write down which command word and tariff you planned for versus what the paper actually asked. A wrong tariff assumption (treating a 12-mark Assess as if it were an 8-mark Discuss, or vice versa) invalidates the depth-matching judgement in every later step, so catch it first.

2. Calculation audit, checked against the four confirmed traps, not just the final figure. For every Calculate answer: is a working shown (a correct method with visible working still earns most of the marks even when the final figure is wrong — confirmed repeatedly across examiner reports)? Does a percentage answer carry its % sign? Has ×100 actually been applied, not just implied? Is the answer given to the specified number of decimal places? Does a non-percentage numeric answer carry its own unit — confirmed on Tesla's labour productivity (Oct 2025 MS Q1a): "13.14 cars per employee" earns full marks with no working shown, but the bare "13.14" earns one mark less. Mark each trap pass/fail individually rather than scoring the calculation as a single right/wrong.

3. Level-descriptor self-placement — one cold read, matched to the whole answer, not a checklist tally. For every Discuss, Assess, or Evaluate answer, re-read it in one pass and decide which level band it most resembles as a whole, the same way an examiner report insists marking actually works: "The levels-based mark schemes are applied in a holistic way rather than looking for individual Assessment Objectives" (near-identical wording confirmed in five separate examiner reports — June 2022, Oct 2022, Jan 2023, June 2023, Jan 2025). Write down the one phrase from that level's descriptor your answer most clearly satisfies, and the one phrase from the level above it your answer does not yet reach. Then assign a mark honestly within the band — not automatically the top of it. An answer that squarely fits Level 3's "accurate knowledge... developed chains of reasoning" but stalls before "an awareness of competing arguments/factors leading to a supported judgement" (the Level 4 language) is a mid-to-high Level 3, not a low Level 4 — the gap between those two bands is a specific missing move, not a rounding error.

4. Discuss-specific honesty check. Confirm no conclusion was added where none was required or rewarded — Appendix 6 states this directly, and the Oct 2021 examiner report confirms candidates lost nothing by stopping at a balanced two-sided argument. If a conclusion is present, it isn't a bonus; Discuss caps at Level 3 regardless, so time spent writing one was better spent developing the second side of the argument instead.

5. Assess- and Evaluate-specific honesty check. Confirm the closing sentence is a genuinely conditional judgement — stating what would have to be true for the conclusion to hold — rather than a flat, unconditional preference asserted at the end. This is the specific, named difference between Level 3 and Level 4 on both tariffs (see the FORGE guide §2 for the full level-ladder quotes). A flat "Company X should therefore do Y" with no stated condition is a Level 3 answer even if everything above it reads as Level 4 — self-mark down one band rather than crediting the sentence for existing.

6. Trap-taxonomy cross-check. Read your own script back against the named traps already catalogued for this paper (FORGE guide §4, sections A–D — calculation traps, command-word and holistic-marking traps, topic-specific misconceptions, and the three newly-found traps: extracts paraphrased rather than used as evidence, reaching for a taught framework the question didn't ask for, and evaluating a technique generically rather than tied to the named company). Tag every match you find in your own answer by name — a vague "I think I lost marks somewhere" is not a fixable finding; "I paraphrased the extract in Q1(c) instead of using it as evidence" is.

7. On a paper already used as a worked exemplar in this course (§2's first conclusion), mark yourself against the mark scheme itself, not against the version of the answer you already half- remember from the lesson. Recognising content you were taught is not the same skill as constructing it cold, and self-marking generously because a sentence "sounds right" defeats the purpose of the exercise on a paper chosen specifically for self-marking-process rehearsal rather than content blindness.

#6. The self-audit log

Fill one row per question, immediately after marking it in §5, while the reasoning is still fresh:

The last two columns are the point of the log: a trap named once is a coincidence; the same named trap recurring across two or more separate mocks (logged, not remembered) is the actual priority for the next revision session — a specific, checkable finding rather than a general feeling of weakness on a topic.

#7. Aggregating results honestly

Total the self-assigned marks out of 80. Do not convert that total into a percentage-likelihood of a grade, or compare it to a grade boundary. This course's own frequency reference states the position plainly: "No grade-boundary data exists anywhere in this course's WBS13 sources" (signal-frequency-and-boundaries.md §3) — no raw-mark-to-grade conversion has ever been part of the material mined for this paper, and a level descriptor's mark range is not a grade boundary. If a grade boundary is genuinely needed, it has to be fetched fresh from Pearson's own published tables and cited with its own dated provenance note — nothing in this course's existing research supports stating one here, so this protocol doesn't manufacture one.

What the raw total and the log in §6 legitimately support tracking across repeated mocks instead:

• The raw mark trend itself (out of 80, series to series) — useful as a relative measure against your own past sittings, not as an absolute grade signal.
• Time-budget adherence — how many of the seven timed rows in §4 finished at or under budget, tracked sitting to sitting.
• Trap recurrence — which named traps from §5's step 6 reappear across two or more logged mocks. This is the most actionable of the three: a recurring trap identifies a specific habit to fix, where the raw mark alone only tells you that something needs fixing.

#8. Repeat-mock cadence, grounded in a finite real paper pool

Twelve located series is a finite resource, and §2 already spends two of the closest-to-blind ones (Jan 2022, Oct 2024) plus Oct 2021's Q2/Q3 early. Once those are used, any further full-paper mock necessarily draws on a series already partially built into this course's own teaching material — that's a legitimate self-marking-process rehearsal (§5's step 7 exists for exactly this), but it is a different exercise from a cold content test, and the log in §6 should record which mode a given sitting actually was, not treat every mock as equivalent. Before assuming the pool is exhausted, check qualifications.pearson.com directly for any series released after Oct 2025 — the newest series in this course's current sources, and one only reached via a third-party mirror at that (§2) — since a genuinely new series is, by construction, the best available material for a real blind mock.

#9. Genuine gaps in this protocol

Flagged rather than smoothed over, per this course's standing rule against manufacturing certainty past what the sources support:

• No grade-boundary conversion is possible — covered in full in §7; this is a limitation of the source material, not of this protocol specifically.
• Analyse (6 marks) and Construct (4 marks) have no confirmed real WBS13 exam question in any series this course has read (FORGE guide §2/§6), so this protocol has nothing paper-specific to check a self-marked Analyse or Construct answer against beyond the Appendix 6 definition itself.
• The "already taught" table in §2 reflects what this course's own material happens to cite, not an exhaustive re-audit of every sub-part of every series — signal-frequency-and-boundaries.md §1 itself flags that reading depth varies series to series (Jan 2022 and June 2022 in particular were mined less completely than most). Treat a "no confirmed worked exemplar" cell as "not found in this course's material," not as a verified guarantee of novelty.
• The Oct 2025 mirror-provenance caveat (§2) applies to every row above that cites that series, not only to the ones repeated there.