How to get an A* in A-Level Maths

The short answer is a number. An A* in Edexcel A-Level Maths this summer took 254 marks out of 300 — a shade under 85% — across three two-hour papers of 100 marks each. Read it the other way round and it becomes useful: you can drop about 46 marks, roughly 15 a paper, and still walk away with an A*. An A took 210. So the whole distance between an A and an A* was 44 raw marks, and in over a decade of teaching this subject I have almost never seen a student lose those 44 marks to content they couldn't understand. They go in ones and twos, on questions the student could do.

Forget 90%. The boundary moves.

Under the old modular A-Level, the A* had a fixed definition: 80% overall plus 90% across the two core A2 modules. That rule is where the "you need 90%" folklore comes from, and it has been dead for years. The current linear A-Level has no UMS marks at all. Exam boards set one subject-level A* boundary after every script in the country has been marked, using statistical evidence about how that cohort performed, which means the boundary lands somewhere different every summer.

Edexcel's A* boundary was 251 out of 300 in 2024, 258 in 2025 and 254 in 2026. So "about 85%" is a fair working assumption, and the other boards set their own numbers, but you cannot revise towards a target percentage — if a paper turns out to be brutal the boundary drops and everyone's mark drops with it. The only thing you control is raw marks. That turns the goal from an abstract percentage into a question you can answer after every past paper: where did the marks go, and were they gettable?

An A* in maths is not exotic

Maths was the most-entered A-Level in the country this summer: 115,380 entries across the UK, up 3.9% on last year, and around one in eight of all A-Level entries. Just over 42% of those entries came out at A or A*.

Two things follow, and they pull in opposite directions. The first is encouraging: an A* here is not a prize handed to a handful of prodigies. A large, ordinary group of students collects one every August, and most of them are not remarkable — they are organised. The second is sobering: nobody drifts into A-Level Maths by accident, so the people you're ranked against are already self-selected and already working. Further Maths is starker still, with close to 58% of its 21,032 entries at A or A*. The marginal mark in this subject is genuinely hard to win, and it is won by not leaking the ones you've already earned.

Where the 44 marks actually go

Content you can't do costs you far less than you'd guess, because it's visible. You know the gap is there, your teacher knows, and it gets fixed. What separates 210 from 254 is a much duller list: a sign error midway through a long expansion, an answer not given to the accuracy the question specified, a mechanics problem attempted without drawing the forces, a missing constant of integration, a "hence" quietly ignored. I've written up the ten that do the most damage — they recur from C-grade students right up to A*-chasers, and every one of them has a specific, trainable fix.

The A* student is not the one who makes no mistakes. They're the one who keeps a short written list of their own recurring failures and has trained each one out on purpose. When I sat these papers myself I came out with 599 out of 600, and it wasn't because I knew more maths than the person next to me. It was because by the spring I knew exactly which four things I did wrong under pressure.

Fluency, not knowledge

Everyone sits the same papers and the content isn't a secret, so what separates the grades is speed and range. A C says the standard methods mostly work. An A says they work quickly and connect across topics. An A* says they work on a question deliberately dressed up to look like nothing in the textbook — and that capacity is built almost entirely through volume of varied practice. There is no shortcut into it, including being clever.

Which is why the foundation matters more than ambitious students want to hear. If GCSE algebra still costs you conscious thought — rearranging, surds, indices, quadratics — then every A-Level question is being solved on a fraction of your available attention, and the disguised ones will beat you. Making that automatic is the least glamorous and highest-return work in the whole course, whether you're doing it over the summer before Year 12 or quietly repairing it in Year 13. If you want to see what fluent working looks like at pace, the worked examples on my videos page are there for that.

Paper 3 is where most A* attempts quietly die

The A-Level is three papers of 100 marks each: two pure, one statistics and mechanics. The applied paper is therefore a full third of your grade, and it is routinely revised like a tenth of it. Students who love pure maths treat statistics as arithmetic and mechanics as somebody else's physics lesson, do a fraction of the practice, and hand back twenty marks on a paper worth a hundred.

Twenty marks is half the distance between an A and an A*. If you're aiming at the top grade and you've been avoiding the applied content since Year 12, that is not a detail to tidy up in May — it's probably the single biggest block of recoverable marks you own, and it deserves more of your revision time than pure does, not less.

What the final months look like

Beyond lessons, a student genuinely targeting an A* wants something like six to eight hours a week of their own work, front-loaded into fluency early in the course rather than saved up for a heroic spring. From January of Year 13 the shape changes: roughly sixteen teaching weeks remain, and how you spend the first fortnight of them sets the ceiling on everything after.

The endgame is past papers, and the way you use them decides whether they're worth anything. Ten to twelve papers marked harshly, with every lost mark categorised and every wrong question redone cold two days later, beats thirty skimmed and scored — the full protocol is here, and the marking half is where nearly all the value sits. Every board's papers are on the past papers hub, free, whether you ever speak to me or not.

When an A* is the wrong target

I'd rather say this plainly than sell you a plan. If you're sitting at a C in January of Year 13, "A*" isn't a target, it's a mood — and organising a year around it tends to produce scattered revision that delivers a B instead of the solid A a narrower plan would have got you. Going from a C to an A is a bigger real achievement and a much better use of sixteen weeks than falling short of a grade you were never within reach of.

It's also worth checking whether you need one. Before you build a year around an A*, look up the actual entry requirement for every course on your list, because offers vary far more than students assume and each course page states its own. And if you're carrying Further Maths as well and one of the two is visibly suffering, that's a real decision with a real deadline rather than a failure of character — I've written about how to make it honestly. An A* in Maths and a dropped Further Maths is a better outcome than two damaged grades.