Exam Day Protocol Extended

Provenance: written 2026-09-29, against research/veridian/WST01-verified-facts.md (read in full — the primary, provenance-checked source for every spec/front-cover/mark-scheme/examiner-report fact below), content/veridian/technique-hub/wst01/forge-exam-technique-guide.md (read in full — this document builds on and does not repeat its §1 timing budget or its §6 short exam-day protocol), codex-formula-and-method-bank.md and signal-frequency-and-boundaries.md in the same folder (read in full to confirm neither already covers this ground), the eight signal-topic-brief-*.md files in this same folder (read in full — each already establishes a real, question-numbered position finding for its own topic; this document's genuinely new contribution is combining those into a single whole-paper position picture, not re-deriving them), the 13 audited lesson files in content/veridian/curriculum/wst01/, and content/veridian/question-bank/wst01.ts (checked — no exam-day-protocol material, no duplication risk). No new PDFs were fetched for this pass. Every claim below traces to a citation already verified in one of those sources, cited by series and question number where the source gives one. Any exam-board wording quoted directly is kept under fifteen words per quote, attributed; everything longer is paraphrased.

#WST01 (Statistics 1) — Exam-Day Protocol (Extended)

#What this document is, and how it differs from what already exists

forge-exam-technique-guide.md §6 already gives a nine-point exam-day protocol. This document does not repeat those nine points. It goes deeper on four things that guide leaves thin: how much of the 90 minutes is genuinely reading time, what a question-order strategy grounded in this paper's real mechanics looks like, a worked elapsed-clock checkpoint table rather than a per-question minute budget, and a concrete sequence of moves for the specific situation of finding yourself behind schedule. Where a claim below is a verified, paper-specific fact, it is cited by series and, where the source gives one, by question. Where advice is ordinary test-taking practice with no WST01-specific source behind it, it is labelled (general practice). Two sections end early rather than invent specifics to fill them out — §2 (reading time) and §5 (running behind) — because the real, paper-specific evidence for both runs out sooner than a document with this title might be expected to fill.

---

#1. What's fixed, and what the paper's real question-order picture actually is

Fixed: 90 minutes; 75 marks; every question compulsory — confirmed on the spec's own Assessment Information and independently on every front page reviewed, already established in forge-exam-technique-guide.md §1 and not re-derived here.

Not fixed: the number of questions (6, 7 or 8 depending on series — same source), and, this document's own added layer, which topic sits where. The eight per-topic signal-topic-brief-*.md documents in this folder each already establish a real, question-numbered position finding for their own topic. No existing document combines them into one picture of the whole paper. Pulling every confirmed question number together, for the five series with examiner-report-level detail (all six questions each):

(Sources: WST01-verified-facts.md §5 throughout; signal-topic-brief-reading-and-comparing- distributions.md §3; signal-topic-brief-elementary-probability-and-conditional-notation.md §3; signal-topic-brief-conditional-probability-independence-vs-mutually-exclusive.md §3; signal-topic-brief-normal-distribution-standardisation-and-conditional-probability.md §3; signal-topic-brief-discrete-random-variables-probability-function.md §3; forge-exam-technique-guide.md §1's Jan 2023 breakdown. Cells marked "not cited" are genuinely unconfirmed in the material this course has read — not assumed blank, not filled with a guess.)

Three real, checkable patterns fall out of that table, and one non-pattern worth stating just as plainly:

1. Q1 is reliably Section 2 (data representation/comparison) or Section 3 (probability) — never Normal distribution, regression, or discrete random variables, in any of the five series with a confirmed Q1 topic. signal-topic-brief-reading-and-comparing-distributions.md §3 independently confirms the data-representation version specifically opens the paper in 3 of the 4 series it checked (Jan 2023, Jun 2022, Jun 2024) — 5 of 5 once Jan 2021 and Oct 2021's probability openers are added. There is no confirmed case of Q1 being anything else.
2. The Normal-distribution question never opens the paper, and is confirmed at Q5 or Q6 of 6 in four of the five series with a confirmed position — the sole exception is Jan 2021, where it sits at Q3 of 6, exactly the paper's midpoint. This exact finding already exists, independently derived, in signal-topic-brief-normal-distribution-standardisation-and-conditional-probability.md §3 — cited here, not re-derived.
3. The Normal-distribution question is not always the last question — sometimes regression is. Jan 2023 places Normal distribution at Q5 (17 of 75 marks, the paper's largest single question, forge-exam-technique-guide.md §1) and regression at Q6. Do not assume the paper's final question is a wind-down; check what it actually is once you have it in front of you.
4. Regression's own position is genuinely inconsistent — Q2, Q2, Q6, Q4 across the four series that test it (signal-frequency-and-boundaries.md row 4.1–4.3) — and signal-topic-brief- correlation-coefficient-and-regression-calculation.md §5 independently notes that within a regression question, the PMCC calculation specifically tends to sit as a middle part rather than opening or closing it. No safe "regression is always question N" rule exists; this document will not invent one.

The practical rule this table supports directly: don't walk in expecting a fixed running order. What you can rely on is that the first question draws from data representation or probability, and that the Normal-distribution question — reliably one of the paper's largest — sits somewhere in the back half, not the front. Treat everything else as unpredictable series to series.

---

#2. Reading time: none confirmed for this paper

Checked directly, and the answer is a genuine absence, not an unchecked gap. A dedicated, timed reading-only phase — a period before the 90 minutes starts, during which you may read but not write — is described nowhere in the spec's Assessment Information, nowhere on any front-cover instruction the facts bank transcribes, and in none of the six examiner reports read for this paper. Sibling Pearson IAL Maths units in this same course (WMA11 through WMA14) show the identical pattern in their own facts banks — worth noting as context, not as evidence for WST01 specifically, since each was checked independently against its own material. What follows is about the first few minutes of the 90, not a separate allowance on top of it.

Verified, paper-specific, and worth doing in those first minutes:

1. Confirm the booklet. The formula and statistical-tables booklet is named verbatim on every front page reviewed as "Mathematical Formulae and Statistical Tables (Yellow)" — check it is on the desk before committing time to the paper itself.
2. Check the real question count before planning anything. It has been 6, 7 or 8 across the 14 series this course's research reviewed — the front page states the real number directly. forge-exam-technique-guide.md §6 already makes this point as its first line; it is restated here only because it is the precondition for §4's checkpoint method below, which needs the real count to build from.
3. Note every question's own printed mark tariff as you skim, not as you reach it — feeding directly into the running checkpoint built in §4.

One genuine, flagged absence: the sibling paper WMA13 carries a real, recurring front-cover rubric naming specific questions where calculator-only working is not accepted. No equivalent question-specific rubric line is confirmed anywhere in the WST01 facts bank. This is stated as an absence in the material checked, not as proof no such line has ever appeared on any WST01 paper.

(General practice, not sourced to this paper specifically): starting with whichever question you personally feel most confident about, rather than strictly in printed order, is ordinary exam habit and is not contradicted by anything above — compulsory answering fixes what you attempt, not the order. No examiner report or mark scheme for this paper recommends a particular attempt order.

---

#3. Question-order strategy, given what §1 actually shows

Because every question is compulsory, "order strategy" here is less about which question to attempt first and more about not losing marks to an assumption §1 shows is unreliable. Three real, paper-specific consequences follow directly:

1. Do not treat the opening question as a warm-up you can rush. It is reliably data representation or elementary/conditional probability (§1) — exactly the material a rushed first few minutes is most likely to fumble on a "read the diagram," "simplify the fraction," or "state the independence condition" step. A strong, automatic foundation in those two areas pays off in the first ten minutes of every series checked, not just occasionally.
2. Do not assume the Normal-distribution question is behind you early, and do not assume whichever question is printed last is the easy one to finish on. §1 shows it lands at Q5 or Q6 of 6 in four of five series, and that the paper's actual final question is sometimes regression instead (Jan 2023). Check the printed topic and mark tariff of what's still ahead once you're a few questions in, rather than assuming from position alone.
3. Where a question's own mark scheme separates a standalone judgement from a chained calculation — documented directly for conditional probability vs. independence (signal-topic-brief-conditional-probability-independence-vs-mutually-exclusive.md §3: a standalone independence judgement scores as an independent B mark, not a chained M/A pair) — that standalone mark does not require the harder calculation around it to be finished. Reading a whole question before starting it, so a mark of this kind is not skipped because it looked buried inside a harder part, is a real, low-cost habit this specific mark-scheme mechanic supports.

(General practice, not sourced to this paper specifically): always finishing one question fully before reading the next, or always doing short questions before long ones, are reasonable exam habits in their own right, but no WST01 mark scheme or examiner report recommends either, and this document will not dress either up as paper-specific.

---

#4. Timing checkpoints, built from a real mark breakdown

The method matters more than the numbers, because question count and per-question tariffs both vary (§1): the only checkpoint tool that survives an unfamiliar paper is a running total you build live — cumulative marks completed vs. cumulative time elapsed — checked against the fixed 72-seconds-per-mark average (90 min ÷ 75 marks, already established in forge-exam-technique-guide.md §1). What follows is a worked illustration of that method using the one series in the facts bank with a complete, independently confirmed, question-by-question mark total: January 2023 (10, 10, 11, 13, 17, 14 = 75 — forge-exam-technique-guide.md §1). This table extends that same breakdown into a running-clock tool it was not originally built to be:

How to use this on a different paper, because it will very likely have a different shape. The table above is one real series, not a template — a 7- or 8-question paper produces different numbers at every row. The transferable habit is the method itself: after finishing each question, add its marks to a running total, and check the clock against (running total × 72 seconds). Running ahead of that figure means earned slack; running behind means §5 below is for you.

One further, real signal worth watching for, not just clock time. On this one fully-documented series, the single largest question on the paper (Q5, Normal distribution, 17 marks — over a fifth of the entire exam in one block) does not land until roughly 70 minutes in, with only a 14-mark question left after it. If you reach that point significantly behind the ~73-minute checkpoint, you are entering the paper's heaviest question already short on time, not its lightest — the opposite of what "nearly done" might otherwise suggest. §1 already shows this isn't a one-series accident: the Normal-distribution question sits in the back half on four of five series checked.

---

#5. What to do if you are running behind

No examiner report read for this paper frames a candidate's incomplete or rushed answers in terms of time pressure specifically — the facts bank's error findings (a missed value in a probability domain, a forgotten square in Var(aX+b), a subtract-from-1 slip, an unshown "show that" step) are about method and completeness, not a stated cause of running short on time. This document will not invent a "candidates who ran out of time" quote that does not exist in the material reviewed. What it can give, honestly, is what this paper's own marking rules make the mathematically correct response to being short on time — real, structural facts, not general encouragement:

1. A correctly-started method still earns its M mark independent of whether the final accuracy mark chained to it is ever reached — the M/A/B mechanism already established in forge-exam-technique-guide.md §2.1. Consequence: writing the method line — the standardisation set-up, the conditional-probability ratio, the equation for probabilities summing to one — before time runs out on that question earns something real. Leaving it blank earns nothing.
2. A standalone (B) judgement does not depend on a harder calculation elsewhere in the same question being finished (§3, item 3). Consequence: triaging a multi-part question under real time pressure, take the independent judgement or "state" part first if you can see it, rather than spending remaining minutes on the hardest part and leaving a free standalone mark unclaimed.
3. The misread penalty is a fixed deduction, not a disqualification. Already established in forge-exam-technique-guide.md §2.5: misreading a given figure costs a fixed two-mark penalty on the affected marks, confirmed in active use (Oct 2021, where an alternative mark scheme was applied after exactly this kind of misread). Consequence for a rushed reading of a question under pressure: a misread number does not zero the part outright if the method from that point on is otherwise sound — it is a bounded, known cost, not a reason to abandon the part.
4. The multiple-attempts rule protects a rushed, unfinished switch of method. Already established in forge-exam-technique-guide.md §2.3: if every attempt is left uncrossed-out, or all are crossed out, every attempt is marked and the best one counts. Consequence: if you start one approach, realise it isn't working, and switch with limited time left, there is no version of "tried two things and ran out of time to tidy up" that scores worse than leaving the part blank.
5. One real, paper-specific caution, framed carefully: Jun 2024's examiner report notes, of the hardest confirmed Normal-distribution sub-skill (a multi-step inverse-Normal problem), that it was "not unusual for students to give up at this point" (Jun 2024 Q5(c) — signal-topic-brief-normal-distribution-standardisation-and-conditional-probability.md §2). This is evidence about candidates abandoning a difficult step, not explicitly about running out of time — the distinction matters, and this document does not blur it. But the practical consequence is the same either way: per item 1 above, a partial standardisation line written before giving up still banks real marks that a blank space does not.

(General practice, not sourced to this paper specifically): the broader advice to move on and come back later if stuck is standard exam technique, not contradicted by anything here, but not something any WST01 mark scheme or examiner report states as guidance either — labelled general rather than dressed up as paper-specific.

---

#6. Genuine gaps and limits of this document

• No separate reading-time allowance is confirmed for this paper (§2) — checked directly against the spec, every front-cover instruction transcribed in the facts bank, and all six examiner reports read. If a future series' front cover states otherwise, this section needs revisiting.
• §1's whole-paper position table covers only the five series with question-numbered examiner-report detail (Jan 2021, Oct 2021, Jun 2022, Jan 2023, Jun 2024) out of 14 series with a verified mark scheme — a lower bound, per the same honest denominator signal-frequency-and-boundaries.md already states for this paper. Q5 in Jan 2021 and Oct 2021, and Q6 in Jun 2024, are marked "not cited" rather than guessed.
• §4's checkpoint table is a worked example from one series (January 2023) — the only series in the facts bank with a complete, independently confirmed, question-by-question mark total. This document declines to guess a second full breakdown from partial data.
• §5 has no paper-specific examiner-report evidence about time pressure itself, only the structural marking rules that follow from being short on time, plus one comment about abandoning a difficult step that is not the same claim. If a future examiner report for this paper directly discusses time management, this section should be extended with that citation rather than left as inference alone.
• This document does not repeat forge-exam-technique-guide.md's existing structure section, mark- coding section, or nine-point exam-day protocol, nor any of the eight signal-topic-brief-*.md files' own topic-level content — read those alongside this one; this document covers only reading-time use, question-order strategy, timing checkpoints, and running-behind contingency, synthesised across documents that had not previously been combined this way.

---

#Sources

• research/veridian/WST01-verified-facts.md — primary source for the assessment structure (§2), question-count table (§2), and every per-series/per-question topic and error citation compiled into §1's table (§5 throughout).
• content/veridian/technique-hub/wst01/forge-exam-technique-guide.md — read in full; its §1 (timing budget, Jan 2023 mark breakdown), §2.1/§2.3/§2.5 (M/A/B mechanics, crossed-out-work rule, misread penalty) and §6 (existing short protocol) are cited above rather than re-derived, and are not repeated.
• The eight signal-topic-brief-*.md files in this same folder — each read in full for its own question-numbered position finding, combined here for the first time into §1's whole-paper table: -reading-and-comparing-distributions.md, -outliers-boxplots-comparing-distributions.md, -elementary-probability-and-conditional-notation.md, -conditional-probability-independence-vs-mutually-exclusive.md, -normal-distribution-standardisation-and-conditional-probability.md, -discrete-random-variables-probability-function.md, -correlation-coefficient-and-regression-calculation.md.
• content/veridian/technique-hub/wst01/signal-frequency-and-boundaries.md — read in full to confirm the honest "N of 6 examiner-report series" denominator used throughout this document, and to confirm no grade-boundary data exists to add to this protocol.
• content/veridian/technique-hub/wst01/codex-formula-and-method-bank.md — read to confirm it contains no reading-time, question-order, or running-checkpoint material before writing this document.
• content/veridian/curriculum/wst01/paper-anatomy.ts — read to confirm its own scope (structural facts and M/A/B definitions) does not overlap with this document's exam-day-strategy focus.
• content/veridian/question-bank/wst01.ts — checked; contains no exam-day-protocol material.