Deep Dive: Sequences and Arithmetic Series

Provenance (2026-09-29). Read in full for this pass: research/veridian/WMA12-verified-facts.md (§2's verbatim transcription of spec items 4.1–4.3; §3's formula-booklet section; §4's mark-scheme conventions; §6.4 in full; §7's "topics with enough verified material" entry for this topic; §9's sequences-and-arithmetic-series.ts mark-scheme-bullet-coverage audit dated 2026-09-14 (PR #441, including its explicitly flagged, not-yet-actioned Q8(ii) lead); and §9's later geometric-series.ts audit entry, specifically its bullet-by-bullet comparison of the real arithmetic-proof mark scheme against the real geometric-proof mark scheme); the complete lesson file content/veridian/curriculum/wma12/sequences-and-arithmetic-series.ts (all 730 lines, read to confirm exactly what it already teaches — its recurrence-vs-formula teach block, its increasing/decreasing/periodic teach block, both mechanism derivations, one worked chain, one chain-drill, one method-comparison, a 6-mark marked-solution with wrong-path, and seven trap-taxonomy entries — so nothing below restates them); geometric-series.ts (grepped for "overlap," "sigma," and "mixed," then read in the relevant sections, to confirm the mixed arithmetic-then-geometric worked exemplar and its two named traps already live there in full); content/veridian/technique-hub/wma12/codex-formula-and-method-bank.md (§4, its sequences/series formula table and its explicit note that spec 4.3 has no primary-source-verified real error pattern); content/veridian/technique-hub/wma12/forge-exam-technique-guide.md (grepped for "arithmetic" and "sequence"; its trap-taxonomy locator line for this lesson); content/veridian/technique-hub/wma12/signal-frequency-and-boundaries.md (this topic's frequency row: seen in 4 of 20 series checked); and content/veridian/question-bank/wma12.ts (grepped for "arithmetic" and "sequence" — one incidental, unrelated match, confirming no worked past-paper item exists there for this topic to avoid duplicating). Not affiliated with or endorsed by Pearson Edexcel. Every quotation below is short (under ~15 words) and attributed to a named document; no full question, mark-scheme extract, or examiner-report passage is reproduced at length. All new arithmetic in this document (the Part 3 decomposition sum, and the two closed-form cross-checks in Part 2) was independently verified by direct computation before being written in.

#CODEX Deep-Dive — Sequences and Arithmetic Series (WMA12, spec 4.1–4.3)

#What this document is, and isn't

The lesson file already teaches this topic in real depth: the recurrence-vs-formula distinction (spec 4.1) with its own "(n−1)-not-n" counting warning; a full teach block on increasing, decreasing and periodic sequences (spec 4.3); two independent mechanism derivations — the nth-term formula falling out of repeated substitution into the recurrence, and the reverse-and-add pairing proof of the sum formula; a worked chain and chain-drill that each turn "two given terms" into a simultaneous-equations problem; a method-comparison between factorising and the quadratic formula for a sum-equals-a-target-value question; a 6-mark marked-solution reproducing the real January 2020 Q8(i) bookwork proof verbatim, with a wrong-path branch; and seven trap-taxonomy entries, six of them tied to named, quoted examiner-report findings on that same real question. codex-formula-and-method-bank.md already holds the compact formula statement and the honest note that spec 4.3 has no primary-source-verified real error pattern of its own. forge-exam-technique-guide.md and signal-frequency-and-boundaries.md already carry this topic's exam-craft framing and its real, floor-level frequency count.

None of those four documents does what this one does: derive, once and in general, the closed-form $a$/$d$ pair that the lesson's worked chain and chain-drill each re-derive from scratch by elimination; work through the one real, examiner-report-named WMA12 question this course's own research has explicitly flagged and left unactioned — a non-arithmetic, formula-defined sequence requiring spec 4.1's general machinery, not spec 4.2's arithmetic one — closing a gap the facts bank itself calls "a real candidate to close... in a future pass"; and give spec 4.3's increasing/ decreasing/periodic claims an actual proof, honestly labelled as reasoned from the definitions rather than sourced to any examiner report, since none exists for this sub-item.

The honest evidentiary picture, stated up front. signal-frequency-and-boundaries.md records arithmetic sequences and series as confirmed present in 4 of the 20 series checked for a question paper and mark scheme (June 2019, general; January 2020 Q8(i); October 2020, general; January 2022, mixed context), carrying ~3 marks where the standard proof is isolated. That is a real floor, not a rarity signal — spec 4 is compulsory every series — but it means the exam-angle ledger at the end of this document has a small, fixed number of real anchors, and does not stretch that into more than it is. One of those four series, January 2020, is now shown below to carry two distinct real angles on this topic, not one, once its second sub-question is read.

---

#Part 1 — The theory model, compactly, plus one structural comparison the lesson doesn't make

An arithmetic sequence is the one member of spec 4.1's general family $u_{n+1}=f(u_n)$ where $f$ is addition by a fixed constant: $u_{n+1}=u_n+d$, $u_1=a$. Repeated substitution gives the closed form $u_n=a+(n-1)d$; writing the sum forwards and backwards and adding pairs gives $S_n=\tfrac12n[2a+(n-1)d]$, with $a=1,d=1$ as the named special case $1+2+\cdots+n=\tfrac12n(n+1)$. Both formulas are printed in the P2 formula booklet (WMA12-verified-facts.md §3) and need not be memorised; the proof of the sum formula is the one part spec 4.2 explicitly requires to be known — "The proof of the sum formula should be known" — and is not reproduced again here; see the lesson's own mechanism block and its six-entry trap-taxonomy for the full, real-anchored treatment.

A structural fact the lesson states in passing but the facts bank makes explicit only in its geometric-series audit entry, worth surfacing here because it belongs to this topic too. The real Pearson mark scheme for the arithmetic proof (January 2020 Q8(i)) gives an independent B1 for writing $S_n$ correctly before any method is applied; the real mark scheme for the parallel geometric proof (October 2020 Q8(a)) does not — it bundles writing $S_n$ and multiplying by $r$ into a single M1, with no standalone mark for the first line alone. The two proofs sit one paragraph apart in the same spec item and look structurally identical at a glance, but Pearson's own marking treats them asymmetrically: the arithmetic proof rewards a correct opening line on its own merits; the geometric proof only rewards that same opening line once it has been carried on into the next step. A student who writes a correct $S_n$ for the geometric proof and then stalls has, on the real scheme, nothing banked — the identical action on the arithmetic proof already has one mark secured.

---

#Part 2 — Sub-case: the closed form behind "two given terms," stated once instead of re-solved each time

The lesson's worked chain ($u_5=33,\,u_{12}=68$) and chain-drill ($u_3=20,\,u_8=50$) each turn two given terms into a pair of simultaneous equations and eliminate $a$ by subtraction. That elimination always follows the identical pattern, because $u_n=a+(n-1)d$ is linear in $n$: for any two terms $u_p$ and $u_q$ ($p\neq q$),

$$u_q-u_p=(q-p)d \quad\Longrightarrow\quad d=\frac{u_q-u_p}{q-p}, \qquad a=u_p-(p-1)d.$$

Cross-checked against both of the lesson's own instances, not asserted in isolation: for $u_5=33,u_{12}=68$: $d=\frac{68-33}{12-5}=\frac{35}{7}=5$, $a=33-4(5)=13$ — identical to the worked chain's own stage-3 values. For $u_3=20,u_8=50$: $d=\frac{50-20}{8-3}=\frac{30}{5}=6$, $a=20-2(6)=8$ — identical to the chain-drill's own model completion. The formula is not a different method from the one the lesson teaches; it is that same elimination, carried out symbolically once so it never has to be re-derived term-by-term under exam pressure. It is worth knowing as a check, not as a replacement for showing the elimination working a real mark scheme's M1 line asks for — quoting a memorised closed form with no visible method line would not, on its own, satisfy a mark scheme that credits "a correct method for eliminating one unknown from the pair of equations" (the lesson's own stage-2 wording, matching this course's general M1 conventions).

---

#Part 3 — Sub-case: a formula-defined sequence that is not arithmetic — closing a named research gap

The real evidence, and why it matters. WMA12-verified-facts.md flags, honestly and explicitly, that January 2020 Q8(ii) — the second half of the same question whose Q8(i) anchors the lesson's whole proof treatment — was never checked in the depth needed to teach it, naming it as "a real candidate to close that gap" — but leaving it for a future pass rather than acting on it. The question itself, as transcribed in the facts bank, is real and specific: given $u_n=5n+3(-1)^n$, find $u_5$ and $\sum_{n=1}^{59}u_n$. This is genuinely different mathematics from everything else in this lesson — it is spec 4.1 territory (a sequence given by a direct formula for the $n$th term, not the arithmetic recurrence of spec 4.2) crossed with spec 4.3 (the $(-1)^n$ factor is periodic with period 2), and no lesson file in this course currently teaches how to sum a sequence of this shape. What follows is this document's own worked treatment of the mechanism such a question requires — not a claim to have the real mark scheme, which the facts bank confirms was not extracted for this exact question.

The mechanism: split the formula into a piece the arithmetic-series formula already handles, and a piece that is finite by periodicity, then sum each separately. $u_n=5n+3(-1)^n$ is the sum of an arithmetic term, $5n$ (itself $u_n=a+(n-1)d$ with $a=5,d=5$, or just "5 times $n$"), and a periodic term, $3(-1)^n$, which takes only the two values $-3$ (odd $n$) and $+3$ (even $n$), forever. A direct sum of the whole sequence splits cleanly by linearity of summation:

$$\sum_{n=1}^{59}u_n=\sum_{n=1}^{59}5n+\sum_{n=1}^{59}3(-1)^n.$$

The arithmetic piece is the sum of the first 59 natural numbers, scaled: $\sum_{n=1}^{59}5n=5\times \tfrac12(59)(60)=5\times1770=8850$ — the a=1,d=1 special case from Part 1, scaled by 5, not a new formula.

The periodic piece uses periodicity directly rather than any series formula: 59 terms of a period-2 pattern split into 29 complete pairs plus one leftover term, because 59 is odd. Each complete pair is one odd-indexed term and one even-indexed term, summing to $-3+3=0$ regardless of where the pair falls in the sequence — so all 29 pairs contribute exactly $0$. The one leftover term is the 59th, and 59 is odd, so it contributes $3(-1)^{59}=-3$. The periodic piece therefore sums to $-3$, not because of any arithmetic-series or geometric-series machinery, but purely because a period-2 sequence with an odd number of terms always has one unmatched term of one parity.

Combining: $\sum_{n=1}^{59}u_n=8850+(-3)=8847$. And separately, $u_5=5(5)+3(-1)^5=25-3=22$. (Both values independently verified by direct term-by-term computation for this document, not merely by the shortcut algebra above.)

Why this genuinely extends the lesson's own coverage, not merely restates it. The lesson's own header note, quoted in the facts bank, states plainly that "no real examiner-verified error pattern was found for 4.1/4.3" when it was written — a real, named content gap in a compulsory spec sub-item. This sub-case does not manufacture an error pattern that isn't there (none is claimed here, and none should be invented for it), but it does supply the one thing that was missing and is real: a worked mechanism for the exact shape of question spec 4.1/4.3 is actually examined with, anchored to a real, named WMA12 question rather than a generic textbook example. The one honest limit on this sub-case: this document does not have access to the real mark scheme for Q8(ii), so no mark allocation, method-mark wording, or examiner-confirmed error is claimed for it anywhere above — only the mechanism and the verified arithmetic.

---

#Part 4 — Sub-case: increasing, decreasing, periodic — proved, not just asserted

codex-formula-and-method-bank.md names spec 4.3 as a genuine, checkable gap: the only unit-content row whose "guidance column for 4.3 is empty." The lesson states the arithmetic-specific consequences correctly but informally ("an arithmetic sequence with $d>0$ is always increasing," "a periodic sequence... is almost never arithmetic"). Both claims have short, clean proofs worth giving in full, precisely because no primary source does the proving for this sub-item — flagged here on that honest basis, not sourced to any examiner report.

Monotonicity is read directly off the sign of $d$. $u_{n+1}-u_n=(a+nd)-(a+(n-1)d)=d$ for every $n$, by direct subtraction of the closed form — so the defining inequality for "increasing" ($u_{n+1}>u_n$ for all $n$) holds for every $n$ simultaneously exactly when $d>0$, and for "decreasing" exactly when $d<0$. There is no arithmetic sequence that is increasing for some values of $n$ and decreasing for others: the difference between consecutive terms is the same constant $d$ everywhere, so its sign cannot change partway through. This is a genuine structural feature of arithmetic sequences specifically, not of sequences in general — a formula-defined sequence like Part 3's $u_n=5n+3(-1)^n$ has no fixed sign of $u_{n+1}-u_n$ at all (it alternates), which is exactly why "increasing/decreasing" is not even a meaningful label to attach to it.

An arithmetic sequence is periodic if, and only if, $d=0$. Suppose an arithmetic sequence is periodic with period $k\geq1$: $u_{n+k}=u_n$ for every $n$. Substituting the closed form, $a+(n+k-1)d=a+(n-1)d$, which rearranges to $kd=0$. Since $k\geq1$, this forces $d=0$ — the sequence is constant. Conversely, a constant sequence ($d=0$) trivially satisfies $u_{n+k}=u_n$ for every $k$, so it is periodic (with every positive integer as a valid period, period 1 being the smallest). This proves the lesson's "almost never" hedge is, in fact, an exact "if and only if": the only arithmetic sequences that are periodic are the constant ones, with no partial or approximate case in between.

---

#Part 5 — Sub-case: the mixed arithmetic-then-geometric context (pointer only)

Already fully covered, with a complete original worked exemplar and two named real traps, in geometric-series.ts — not reproduced here. The real anchor is January 2022 Q(c)/(d) (WMA12-verified-facts.md §6.4): an arithmetic series with common difference $-2$ transitions into a geometric series starting from the arithmetic series' own final term, where candidates confirmedly "double counted" the shared term. The arithmetic-side contribution to this trap is identifying the shared term's exact value using $u_n=a+(n-1)d$ before doing anything else with it — the same discipline Part 2's closed form above formalises — and then treating that one value as belonging to exactly one of the two sums, never both and never neither. geometric-series.ts's own overlap-term-double-counted and geometric-series-restarted-from-the-original-first-term trap-taxonomy entries carry the full, real-quoted detail; this document adds nothing beyond cross-referencing them so the exam-angle ledger below is complete.

---

#Every real exam angle this topic has actually been tested from — consolidated ledger

Read every row as "confirmed seen in the series named," never as a probability. No likelihood percentage and no grade boundary specific to this topic is stated anywhere in this document, per this course's standing discipline; signal-frequency-and-boundaries.md carries the paper-wide grade boundaries.

Cross-cutting observations, each traceable to a specific row above, none invented:

• One series, two real angles. January 2020 is counted once in signal-frequency-and-boundaries.md's "4 of 20" floor, but this document shows it actually carries two structurally distinct real questions on this topic — a bookwork proof (Q8(i), fully taught) and a formula-defined-sequence summation (Q8(ii), taught here for the first time). The floor count is a count of series, not of distinct real angles, and should be read that way.
• This topic's best-evidenced angle (Q8(i)) is bookwork, not novel problem-solving — the real failure mode, confirmed six ways, is executing a known proof under exam conditions, not misunderstanding what is being asked. This is a different shape of difficulty from, say, the binomial expansion's dominant real error (index/coefficient mismatch) or the mixed-series overlap trap (context misreading) — worth naming because the intuitive assumption that arithmetic series are trivially easy is not what the real evidence shows costs marks.
• Spec 4.1 and 4.3 remain the thinnest-evidenced parts of this topic. Q8(ii) closes part of the gap the facts bank named, but only the mechanism; no primary-source error pattern for either sub-item exists anywhere in this course's research as of this document, and none is claimed here. Part 4's proofs are sound mathematics, honestly labelled as reasoned from the definitions, not as examiner-confirmed findings.
• No grade-boundary or likelihood claim is made anywhere in this document. The topic's own confirmed floor is "4 of 20 series checked" (signal-frequency-and-boundaries.md) — stated as a floor on what has been individually cited, never converted into a percentage or a prediction.