The hardest GCSE Maths questions, ranked

Most GCSE Maths papers have a handful of questions that can decide the difference between a grade 7 and a grade 9. They tend to be clustered in the final third of the paper, they are often worth four to six marks each, and they tend to be the ones examiner reports describe as having a low pickup rate.

These questions are not necessarily testing harder content. Many of them sit on the Higher Tier specification you have already covered. What makes them difficult is the combination of unfamiliar wording, multi-step reasoning, and the assumption that you will spot a connection between two or three topics on your own.

This guide ranks eight of the harder question types on the AQA and Edexcel Higher Tier papers, explains what makes each one tricky, and tells you what the mark scheme rewards. Treat it as a hit list for the final weeks of revision.

The ranking is based on three signals. First, examiner reports from AQA and Edexcel for the last three exam series, which call out specific questions where the cohort underperformed. Second, the topics that consistently appear in the final third of past papers, where the hardest marks live. Third, the topics that students aiming for grade 9 mention most often as confidence gaps when they sit mock papers.

We have stayed away from inventing precise pass rates because examiner reports rarely publish them in clean percentages. Where we have used numbers, they come directly from the official mark schemes and grade boundary documents. The rest is qualitative and based on consistent patterns across exam series.

Vector questions are a staple of the final pages of Higher Tier papers. The harder ones ask you to prove that two vectors are parallel, that three points are collinear, or that a particular ratio holds within a geometric figure.

What makes them tough is that the answer is not a number. It is a written argument. You have to express vectors using the notation the question sets up, manipulate them algebraically, and finish with a clear conclusion in words. Students who can find the right vector but cannot communicate the proof lose method marks they should have banked.

The AQA mark scheme typically awards marks for the correct vector expressions, the correct manipulation, and the explicit conclusion (something like "AB is parallel to CD because both are multiples of the same vector"). Skipping the conclusion is one of the more common slips.

Circle theorems on their own are not the problem. Students learn the seven theorems, practise spotting them in single-step questions, and feel confident. The hard version chains two or three theorems together in a single problem, and asks you to find an angle that depends on each step holding.

What catches students out is the order. Knowing the theorems is not enough: You also need to know which one to apply first. If you start in the wrong place you cannot reach the answer. Examiner reports flag this regularly. Students often write down the right theorems but cannot link them into a sequence.

The fix is to practise problems where the diagram looks busy and you have to identify which theorem unlocks the first angle. Once you have that one, the rest tend to follow.

Algebraic proof questions ask you to show that an expression is always even, always a multiple of some number, or always satisfies a particular identity. They typically appear once per paper at Higher Tier and they often appear in examiner reports as low-scoring.

The difficulty is structural. You need to represent the general case correctly (using 2n for any even number, 2n+1 for any odd number, three consecutive integers as n, n+1 and n+2 and so on), then manipulate the algebra cleanly, then explain why the final form proves the claim. Many students can do one or two of those steps. Doing all three under exam pressure is the bar.

Mark schemes reward the correct setup, the correct algebraic manipulation, and the explicit conclusion. Writing "therefore it is always a multiple of 3 because 3(n+1) is divisible by 3" is the marks-bearing line that students often forget.

Iteration is a topic that rarely lives in textbook exercises but regularly appears in exams. You are given a recurrence relation, an initial value, and asked to find the next few terms, or to show that the iteration converges to a root of a given equation.

What makes iteration hard is unfamiliarity. The notation looks intimidating (the subscript n and n plus one notation throws students), and the calculator work has to be done carefully without rounding intermediate values. Round too early and your final answer drifts away from the expected mark scheme tolerance.

A solid preparation route is to drill 10 to 15 iteration questions in a row so the notation becomes automatic. Examiner reports often flag students for stopping after two iterations when the question asked for the value correct to a certain number of decimal places, which usually requires more steps.

Histograms are widely flagged in examiner reports as a low-scoring topic. The trap is that students confuse frequency with frequency density, or draw the bars at the wrong height, or read frequency densities off a finished histogram without multiplying by the class width.

The core formula (frequency density equals frequency divided by class width) is not difficult on its own. What catches students out is reading the question correctly. "Estimate the number of values between 12 and 18" requires you to multiply the frequency density by the class width for that section, then add the totals together. Skipping the multiplication, or guessing the class width from the bar position rather than the axis, loses marks fast.

Drill at least five histogram questions before the exam and check every one against the mark scheme to make sure you are using frequency density correctly.

Functions are a regular Higher Tier topic, but the harder questions test composite and inverse functions together. You might be asked to find fg(x) where f and g are given, then solve fg(x) equals some value, then find the inverse of the result.

Multi-step manipulation is where errors cascade. A small slip on the first substitution propagates through everything that follows, so students who lose one mark on setup often lose the whole question. The notation also catches students out, because fg(x) means apply g first, then f, which is the opposite of the visual order most students assume.

The fix is mechanical: Always write out fg(x) as f(g(x)) before substituting anything, and always check the final answer by plugging a value back in. The mark scheme rewards method marks if your setup is correct, so write every step.

Surds appear on the non-calculator paper, which is where they bite. You have to manipulate irrational numbers using a small set of rules (the product rule, the quotient rule, the conjugate trick for rationalising denominators), and you have to do it without a calculator to check yourself.

The harder questions combine surds with algebra: Expanding brackets containing surds, simplifying surds in fractions, rationalising denominators that contain a sum of a rational and an irrational number. These are 3 to 4 mark questions where every line of working has to be exact.

Students often lose marks for not simplifying their final answer. The mark scheme typically expects the fully simplified form, so always check whether the surd can be reduced further (for example, root 12 should generally be rewritten as 2 root 3).

Quadratic graph questions test whether you can complete the square, find the turning point coordinates, and sketch the resulting parabola with intercepts labelled. The harder versions ask you to do all three in one extended question.

Completing the square is often the bottleneck. Students who try to substitute values to find the turning point often miss it entirely. The mark scheme rewards the correct completed-square form, then the correct turning point coordinates derived from that form, then a sketch that matches.

Link completing the square with solving quadratics by the formula and with the discriminant. Examiners often combine these in a single multi-part question, and the parts tend to depend on each other.

There is no shortcut, but the route is well-trodden. Past papers under timed conditions, marked honestly against the official mark scheme, tend to be one of the highest-leverage activities in the last 12 weeks before the exam.

Work through every paper your board has released for the current specification. For each one, score yourself against the boundaries, then categorise every dropped mark into three buckets: Topic gaps, careless errors, or timing issues. Each bucket gets a different fix. Topic gaps need targeted re-revision and 10 to 15 fresh practice questions on the topic. Careless errors need slower, more deliberate working with every step shown. Timing issues need speed practice on the easier opening questions so the hard ones at the end get the time they deserve.

Examiner reports are underused. They tell you, in the examiners' own words, where the cohort dropped marks last year. Read the reports for your board across the last three years and you will often spot the same mistakes appearing again and again.