Most students who miss an A* in A Level Mathematics do not miss it because a question defeated them. They miss it by four or five marks scattered across four papers — a rounding convention here, an omitted constant of integration there, two minutes lost to a question they should have abandoned.
That is the useful thing to understand about the A*. It is rarely a wall. It is a leak.
What follows is not “work hard and do past papers.” It is the specific list of where those marks go, and what to do about each one.
First: know where the marks actually live
For Cambridge A Level Mathematics (9709), the four papers are not equally weighted. The Pure papers carry 75 marks each; the applied papers carry 50 each. So of the 250 marks in the full A Level:
- Pure Mathematics 1 and Pure Mathematics 3 together are 150 marks — 60% of your grade.
- Your two applied papers (Mechanics 1, Statistics 1, Statistics 2 — whichever combination you sit) share the remaining 40%.
Students routinely spend disproportionate time on the applied paper they find least comfortable. That is understandable and often wrong. A student who is shaky in Pure and excellent in Statistics has arranged their weakness in the most expensive place available.
If your revision time is limited, weight it 60/40 toward Pure. That is not a preference; it is arithmetic.
One caveat worth stating plainly: Cambridge revises syllabuses and grade thresholds periodically, and how the A* is calculated depends on your route through the qualification. Confirm the current structure and the A* rule for your session with your school’s exams officer — do not plan around a forum post.
The rounding rule that quietly costs A-grade students their A*
This is the single most common avoidable loss I see, and it has nothing to do with mathematical ability.
Cambridge expects final answers to three significant figures unless the question says otherwise, and angles in degrees to one decimal place. Two things follow, and the second is the one students get wrong:
- Give the final answer to 3 s.f. — not 2, not 5, not an unrounded decimal string.
- Do not round intermediate values. Carry the full accuracy in your calculator and round only at the end.
Round a value to 3 s.f. in the middle of a multi-step question and your final answer can drift outside the accepted range. The method was perfect. The mark is gone. In a long question this can cost two or three marks, and it will happen on every paper if the habit isn’t fixed.
Fix it mechanically: use your calculator’s Ans key or store values in memory rather than retyping a rounded number. If you find yourself writing down 4.67 and then typing 4.67 back in, you have just introduced the error.
Method marks are the cheapest marks in the paper
Read a mark scheme properly — not the answers, the M and A annotations. You will find that a six-mark question typically awards four or five marks for method and one for the final value.
This has a direct consequence: a wrong answer with clear working is worth far more than a right answer with none. Students who work in their heads to save time are optimising for the wrong thing.
Three habits that convert directly into marks:
- Write the rule before you use it. State the quotient rule, or the identity you are applying, then substitute. That line is often a mark on its own.
- Never erase. Cross out with a single line. Examiners cannot award marks for work that isn’t there, and crossed-out work is still marked if nothing replaces it.
- If you are stuck, write down what you know. Setting up the correct integral and failing to evaluate it still earns method marks. A blank space earns nothing.
The recurring Pure 3 losses
Certain topics account for a disproportionate share of dropped A* marks. These are worth targeted drilling rather than general revision.
Choosing the wrong integration technique. Most P3 integration difficulty is not execution — it is recognition. Before integrating anything, ask in order: is it a standard form? Does substitution simplify it? Is the numerator the derivative of the denominator (giving a logarithm)? Is it a product needing parts? Is it a rational function needing partial fractions? Practise classifying twenty integrals without solving them. It is a faster drill than solving five, and it fixes the actual weakness.
The constant of integration. In differential equations, + c is not decoration. It must be there, and it must be evaluated using the given condition — and, critically, applied before you rearrange into the final form, not after. This is a mark that A-grade students lose repeatedly.
Complex number conventions. The argument is conventionally given in the range −π < θ ≤ π. Sketch the Argand diagram before computing the argument; the calculator will not tell you which quadrant you are in, and it will hand you a plausible wrong sign without complaint.
Partial fractions with a repeated factor. A repeated linear factor needs both A/(x−a) and B/(x−a)², and a quadratic factor needs a linear numerator. Getting the form wrong wastes the whole question.
Vectors — reading the question. Most lost vector marks come from answering a slightly different question than the one asked: the distance from a point to a line rather than to a plane, the acute angle rather than the obtuse one. Underline what is actually being asked before you start.
Timing: about a minute and a half per mark
The papers are built to a consistent rhythm. Pure 1 and Pure 3 give you 110 minutes for 75 marks; the applied papers give 75 minutes for 50. Both work out at roughly 1.5 minutes per mark.
That gives you a usable rule in the exam hall: a four-mark question deserves about six minutes. If you are at ten minutes with no route through, you are no longer solving it — you are donating marks elsewhere in the paper.
Leave it, mark it clearly, move on, and come back. Students lose more A*s to one stubborn question than to any topic they never learned. Returning with fresh eyes after ten minutes also works far more often than grinding does.
How to practise so it actually transfers
Past papers are necessary but most students use them badly. Three corrections:
Start earlier than feels reasonable. Begin timed papers around three months out, not three weeks. Early papers will be uncomfortable. That is the point — you want to discover your failure patterns while there is still time to repair them.
Mark your own work against the mark scheme, harshly. Award yourself only what the scheme actually awards. Generosity here is self-deception, and it is the reason students are surprised on results day.
Keep an error log. One page, three columns: the question, what went wrong, and the category (careless / method not known / misread / timing). After five papers the pattern will be obvious, and it is almost never “I need to revise everything.” It is usually two topics and one habit. Fix those and the grade moves.
Redo every question you got wrong a week later, from blank paper. Reading your corrected solution creates a comfortable feeling of understanding that does not survive an exam hall.
The final six weeks
Stop learning new content and start consolidating. Full papers under strict timed conditions, one or two a week, marked properly. Rework your error log. Drill the two topics it identifies.
In the last week, do no new papers. Reread your error log, revise the standard results you are expected to know from memory, and sleep. A tired student makes exactly the careless errors this entire article is about eliminating.
The short version
An A* is not a different kind of mathematics. It is the same mathematics with the leaks sealed: full accuracy carried through, working written down, technique chosen deliberately, time budgeted per mark, and errors logged rather than forgotten.
Every item on that list is a habit, and habits are trainable. That is genuinely good news — it means the A* is available to considerably more students than currently get it.
If you are aiming for an A* and want an honest assessment of where your marks are actually going, you are welcome to book a free class. Bring a marked past paper — that single document usually reveals more in twenty minutes than a term of general revision.