Night Before Guide

#WST01 — Night Before Guide

#Pearson Edexcel IAL Mathematics, Unit S1: Statistics 1

Provenance: compiled 2026-09-29. No new primary-source work was done for this document — it is a same-evening distillation of material already verified in research/veridian/WST01-verified-facts.md and already organised in this folder's forge-battle-card.md (top-5 danger warnings, last-minute checklist), signal-frequency-and-boundaries.md (topic-frequency table), and forge-forty-eight-hour-checklist.md (§4, "what not to start learning fresh"). Every citation below traces to one of those four files; none is re-verified here. Not affiliated with or endorsed by Pearson Edexcel.

#What this is, and isn't

The 48-hour checklist already schedules the final day in blocks, down to a T-1-to-T-0 "morning of" pass. This document doesn't repeat that schedule. It's the version for one specific moment — the evening before, when there's time for a last look and nothing more — compressed to the handful of things worth that look, and the handful worth deliberately leaving alone.

#Worth a last look tonight

• Var(aX+b): square the constant. The same squaring slip recurs across Jan 2021 Q4 and twice within the Jun 2024 paper (Q2(c), Q3(d)) — confirmed in two series, three years apart.
• Conditional probability combined with the Normal distribution. Flagged as the most discriminating step in all five fully-read examiner reports that reach one (Jan 2021 Q3(c); Oct 2021 Q6(c); Jun 2022 Q6(c); Jan 2023 Q5(e)). Draw the curve, shade the region, then decide what to divide by.
• Subtracting from 1 — check both directions. Confirmed wrong both ways across series (under: Jan 2021 Q3(a), Jan 2023 Q5(a); over: Jun 2024 Q5(a)). Pearson's own fix, stated twice: sketch the curve before touching the table.
• Percentage Points table to 4 d.p., never a rounded z. 1.6449, not 1.64; 1.0364, not 1.03 — the same substitution costs a mark across three separate series (Jan 2021 Q3(b); Oct 2021 Q6(b); Jun 2024 Q5(b)).
• "Show that" needs the intermediate line written down. Pearson's single most repeated meta-complaint, independent of topic, named in every general-comment section read.
• Section 2 (data representation) is the one safe assumption — confirmed tested in all six fully-read series. Section 6 (Normal distribution) carries the heaviest single-question tariff seen — 17 of 75 marks where a full breakdown exists (Jan 2023 Q5).

#What not to start learning fresh tonight

• Hypothesis testing. Confirmed absent from this paper's spec guidance, stated explicitly twice. If a different unit's habit pulls you toward $H_0$/$H_1$ tonight, it doesn't belong here.
• Deriving anything — the Normal p.d.f. and its mean/variance/CDF, or the regression coefficients $a$, $b$, $r$. The spec states plainly that none of these derivations are required; know how to use the formulae, not prove them, and don't spend tonight trying to.
• A new calculator, or new settings. If you haven't practised with it on a past paper, tonight isn't the time to switch.
• A grade-boundary number. None exists anywhere in this course's verified record for this paper — there is nothing real to calibrate a target against.

#One honesty note

Nothing above is new evidence. It's a same-evening compression of the battle card's top-5, the frequency table's two clearest findings, and the checklist's clearest "don't" items — chosen because they're the ones a tired read the night before is most likely to actually retain. Fuller citations and the complete record sit in the four files named at the top.