Edexcel A-Level Maths past papers: the full archive, used properly
Pearson publishes every Edexcel A-Level Maths (9MA0) paper, mark scheme and examiner report free on its website, from the 2018 specimen set through the most recent June series. Work through 2018–2023 as timed practice with an error log, treat the examiner reports as required reading, and save the two newest series as final full mocks. Paper 3 needs separate attention: its statistics section assumes familiarity with Edexcel's large data set.
Edexcel has been running this specification since 2018, which means there is now a deep archive of real papers — more than most students will ever finish. The archive is a resource, but only if you sequence it deliberately instead of doing papers at random.
The archive, at a glance
| Series | Papers available | Best use |
|---|---|---|
| Specimen (SAMs) | Papers 1, 2, 3 | Early practice — written before the real exam, slightly off-style |
| 2018–2021 | Full sets each year | Topic drilling and untimed-to-timed transition |
| 2022–2023 | Full sets | Timed sections — the core of your practice |
| 2024–2025 | Full sets | Save as final mocks under strict conditions |
Everything lives on Pearson's qualifications site under A-Level Mathematics (9MA0): question papers, mark schemes and the examiner reports for each series. The reports are the underrated half of the archive — they list, paper by paper, the exact steps candidates refused to write down.
Paper 3 needs its own plan
Paper 3 is the only paper Edexcel students routinely under-practise, because it is 'only' applied maths. It is also the paper with the most repeatable structure: Section A statistics leans on the large data set, and Section B mechanics runs through the same skeletons — resolving, F = ma, moments, projectiles. Do Paper 3 as two separate one-hour halves before attempting it whole.
- Statistics section: practise every hypothesis-test question from 2019 onwards — they are near-identical in skeleton, and the conclusion-in-context mark is dropped every year.
- Large data set questions: they ask things like 'explain why this variable could not be modelled as normal' — the answer lives in knowing the data is categorical or has missing values.
- Mechanics section: the 'state an assumption' or 'give a limitation of the model' line appears on nearly every paper. Have the phrases ready.
What the mark scheme is really telling you
Edexcel mark schemes are built from M marks (method), A marks (accuracy) and B marks (independent results). The useful detail: an M mark can be earned with a wrong final answer, and an A mark usually requires the preceding M. When you self-mark, give yourself nothing for method you only thought — the examiner sees paper, not intentions.
A meteorologist claims more than 30% of days at a station are 'rainy'. In a sample of 60 days, 25 were rainy. Test at the 2.5% level.
- H₀: p = 0.30, H₁: p > 0.30. Under H₀, X ~ B(60, 0.30).
- P(X ≥ 25) = 1 − P(X ≤ 24). Using the cumulative binomial, P(X ≤ 24) ≈ 0.9404, so P(X ≥ 25) ≈ 0.0596.
- 0.0596 > 0.025 — not significant at the 2.5% level.
- Conclusion must be in context: 'insufficient evidence that more than 30% of days are rainy' — bare 'accept H₀' loses the mark.
- The large data set context is the point: this question type appears with weather variables on nearly every Paper 3.
Do not reject H₀ — 0.0596 > 0.025, no significant evidence for the claim at 2.5%.
Mistakes specific to Edexcel papers
- Doing Papers 1 and 2 endlessly while Paper 3 stays at two attempts — a third of the grade, chronically under-trained.
- Skipping the 'give a limitation' or 'state an assumption' line — worth a mark on almost every mechanics question.
- Self-marking with a generosity the real examiner won't share. If the method isn't written, it didn't happen.
- Treating the large data set as trivia — the statistics questions are built on it.
- Ignoring the October resit papers — they are real, recent, and a good source of extra volume once June papers run out.
The checklist
- I've downloaded the 9MA0 papers, mark schemes and examiner reports from Pearson's site.
- I've done at least four full Paper 3s, and the statistics section with the large data set in front of me.
- My error log separates 'didn't know' from 'rushed' from 'misread' — they need different fixes.
- I've read the two most recent examiner reports for each paper.
- The latest series is saved as an untouched final mock.
- I've practised writing hypothesis-test conclusions in context, every single time.
When the mark scheme compresses three lines into one and you cannot see how it got there, that is the moment to stop guessing. Upload the question to Lumi, point at the line, and it will draw the missing steps and then hand you a similar question to prove the fix held.
What to take from this
- All official papers are free on Pearson's site — you never need to pay for Edexcel past papers.
- There are now eight-plus real series for each of Papers 1, 2 and 3, plus specimen papers.
- Edexcel mark schemes are method-heavy — self-mark strictly, awarding nothing for unwritten working.
- Paper 3's statistics questions assume the large data set; practise them with it open.
- Examiner reports repeat the same complaints yearly — read the last two before each paper.
Questions people also ask
From the 2018 first series onwards there are eight-plus June series for each paper, plus specimen papers and October resit papers in some years — well over 50 full papers across Papers 1, 2 and 3. More than enough; sequencing matters more than volume.
Yes, but mid-revision, not last. SAMs were written before any real exam sat, so their style and difficulty are slightly off the real papers. Use them for extra volume once you have done the earliest real series.
Very — they are real Edexcel papers to the same specification, just sat by resit candidates. They are a good source of fresh unseen questions once you have exhausted the June archive.
Boundaries move each series, but as a working target an A has recently sat around 65–75% aggregate and A* around 75–85%. Your paper scores are an indication, not a guarantee — the real boundary is set after marking.