At IGCSE, the question worth asking about a mock is how fast can I build it? At A-Level, that’s the wrong first question. The first question is does this paper actually look like the one my students will sit in May? — because at A-Level the gap between a rough approximation and a faithful one is the gap between a mock that predicts a grade and a mock that misleads everyone who reads it.

A-Level papers are harder to fake. They come in multiple components, each with its own character. They weight assessment objectives in fixed proportions. They expect synoptic thinking that pulls together content from across two years. And they hinge on extended, essay-length responses where a single question can be worth more marks than an entire IGCSE short-answer section. If your mock gets those things wrong, the score it produces is noise dressed up as a prediction.

This guide is about fidelity. It walks through how to create custom mock exams for A-Level that mirror the real paper faithfully — built from real Cambridge and Edexcel past-paper questions, with the structure, weighting and timing that make the result a genuine signal. It’s written for the individual teacher building their own mock, on their own, without waiting on a department or a committee.

Why fidelity matters more at A-Level than anywhere else

A weak IGCSE mock costs you a slightly inaccurate checkpoint. A weak A-Level mock can cost a student a university offer, because A-Level mock grades feed predicted grades, and predicted grades feed applications. The stakes pull the accuracy bar up.

Fidelity at A-Level means four things a generic quiz can’t give you:

• Multi-component structure. A real qualification isn’t one paper — it’s Paper 1, Paper 2 and often Paper 3, each testing different skills (recall and application, data handling, synthesis and extended argument). A mock that’s one undifferentiated question dump tests none of them properly.
• Assessment-objective weighting. Each component allocates marks across AOs — knowledge, application, analysis, evaluation — in fixed proportions. A paper that’s accidentally 80% recall tells you nothing about whether a student can evaluate.
• Synoptic content. A-Level deliberately tests connections across the whole course, not the last unit you taught. The hardest marks reward pulling threads together.
• Exam-accurate tariffs and timing. A 25-mark essay in 45 minutes is a completely different test from the same essay in 25 minutes. Get the tariff and the clock wrong and you’re measuring the wrong thing.

Miss any of these and you haven’t built a smaller version of the real exam — you’ve built a different exam that happens to share a subject.

Start from real past-paper questions, always

The principle that makes everything else possible: build from genuine Cambridge or Edexcel past-paper questions, not from scratch and not from worksheets. This matters even more at A-Level than at IGCSE, where I make the same case for IGCSE mocks.

Every real A-Level past-paper question already carries what you’d otherwise spend hours recreating:

• A full examiner mark scheme, including the levels-of-response descriptors that govern how extended answers are graded — the part that’s almost impossible to write convincingly yourself.
• Calibrated difficulty and exact command words — “discuss”, “evaluate”, “to what extent”, “assess” — at the standard your students will actually meet.
• A known mark tariff and AO breakdown, so you can assemble a paper whose weighting matches the spec instead of guessing.

Start from the past-paper bank and you’re assembling a faithful paper, not authoring an approximate one. That’s why a proper A-Level test builder from past papers beats any blank-page quiz maker: the questions arrive pre-loaded with the very things that make A-Level fidelity hard.

Step 1: Decide which component you’re mirroring

Don’t start pulling questions until you’ve decided what you’re building. A “mock” can mean very different papers:

• A full multi-component mock — Papers 1, 2 and 3 in sequence, mirroring the whole qualification. Use this for predicted grades and full exam-condition rehearsal.
• A single-component mock — Paper 3 only, for example, when you want to pressure-test essay technique specifically.
• A synoptic checkpoint — deliberately weighted toward cross-topic connections, to see whether students can integrate the course rather than recite units.

Each component has a distinct personality. Paper 1 might be application-heavy structured questions; Paper 2 might be source or data analysis; Paper 3 might be extended essays. Build them as the distinct things they are — don’t blur three components into one generic paper, because the real exam never does.

Step 2: Match the assessment-objective weighting

This is the step most home-built mocks skip, and it’s the one that separates a faithful A-Level paper from a flattering one. Pull up the spec’s AO grid for the component you’re mirroring and treat the percentages as a budget:

• If Paper 2 is, say, 30% AO1 (knowledge), 40% AO2 (application) and 30% AO3 (analysis/evaluation), your selection of questions should land near those proportions.
• Watch for the silent drift toward recall. Knowledge questions are easy to find and easy to mark, so unguarded papers quietly over-weight AO1 and under-test the higher objectives — exactly the skills that decide A* from B.
• Use the command words as your proxy for AO level when you’re pulling: “state” and “describe” sit low; “explain” and “apply” sit in the middle; “evaluate”, “discuss” and “to what extent” carry the top objectives.

Get the AO balance right and the mock stresses the same muscles the real paper does. Get it wrong and a student can score well on your mock and badly in May, or the reverse — either way you’ve lost the signal.

Step 3: Build in the synoptic and extended content

A real A-Level paper climbs. It opens with accessible marks and ends with the high-tariff, synoptic, extended-response questions that actually separate the grades. Your mock should have the same shape:

1. Anchor the structure first. Lay down the opening structured or short-answer questions, then the mid-tariff application questions, then the high-tariff extended responses. Build the skeleton, then fill it.
2. Pull genuinely synoptic questions for the back end. Use the question bank’s topic filters to find questions that span units rather than testing one in isolation. These are the marks that reward two years of teaching, and a mock without them lets weak synoptic skills hide.
3. Respect the essay tariff. A 20- or 25-mark extended answer isn’t four 5-mark questions stacked up — it’s one sustained argument graded on levels. Include the real high-tariff questions, with their real mark allocations, so students rehearse building an argument under pressure.
4. Confirm every question brings its mark scheme. Each real past-paper question should arrive with its examiner scheme attached — including the levels descriptors for the essays. This is what makes the paper markable later without you reconstructing the scheme from memory.

Step 4: Get the timing and tariff exam-accurate

Timing is where A-Level mocks quietly fail. Students don’t lose A-Level marks only because they don’t know the content — they lose them because they can’t produce a 25-mark argument in the minutes the real paper allows. So mirror the clock as carefully as the content:

• Match the total marks and the total time to the real component. If the real Paper 3 is 60 marks in 90 minutes, your mock should be too — not 75 marks in the same window because you got enthusiastic pulling essays.
• Watch the running total as you build. A builder that totals marks and estimated timing as you add questions saves you from the “this is 82 marks and three hours” surprise after you’ve already assembled it.
• Keep the marks-per-minute realistic. That ratio is what trains exam pace. If it’s off, you’re testing a different kind of stamina than the real paper demands.

Step 5: Set it up to mark — and decide what you’ll review by hand

The way you build the paper determines how much of it can mark itself. At A-Level the split is sharper than at IGCSE because the extended answers carry so much weight:

• Auto-mark the point-based bulk. Objective, short-answer and point-based structured questions can be marked instantly against the mark scheme.
• Flag the extended and essay responses for your review now. Levels-of-response marking needs your judgement against the descriptors, and you should decide before you see scripts which questions you’ll check by hand. The accuracy A-Level demands is exactly why extended answers get a human pass — I go deeper on that in AI marking for A-Level past-paper extended responses.
• Choose online or printed deliberately. Typed answers mark the cleanest; printed scripts are closer to the handwritten reality of the exam hall but add a capture step.

This is where building and marking connect — but the focus here is the build. A faithfully built paper, with schemes attached and review points decided, is a paper that’s ready to grade efficiently the moment students finish.

Step 6: Save it as a component template

The first faithful mock takes real care; the second shouldn’t. Once you’ve built a Paper 3 with the right AO weighting, synoptic balance, essay tariffs and timing, save that structure as a template. Next cycle you swap in fresh past-paper questions on the topics you want, and the hard-won fidelity carries over automatically. A good A-Level mock exam builder turns your first careful build into a reusable blueprint per component, so mock season stops being a from-scratch event every term.

What to watch out for

• Don’t let breadth collapse into recall. The easiest questions to find are the low-AO ones. If you don’t watch the weighting, your paper drifts toward knowledge and stops testing the skills that decide the top grades.
• Don’t shrink the essays to save marking. The extended responses are the point of an A-Level mock. Cutting them because they’re slower to grade defeats the exercise — flag them for review instead.
• Keep a real paper open as your reference. When in doubt about structure or tariff, lay your mock next to an actual recent component and check that the shape, the marks and the AO spread line up.

How this looks in practice

If you want to do all of this in one place, Tutopiya’s platform for teachers is built around exactly this build-faithfully-then-mark flow. The Test Builder assembles full and multi-component A-Level mocks from real Cambridge and Edexcel past-paper questions — filtered by topic, difficulty and command word — totals your marks and timing as you build, and brings each question’s mark scheme with it, including the levels descriptors for extended answers. The point-based questions mark instantly against the actual mark scheme; the essays get examiner-style feedback with a review-and-override step so you keep control of the high-stakes marks. It’s free to start, which is enough to build and run a full component mock end to end. If you also want the board-alignment fundamentals, see building exam-board-aligned tests for Cambridge and Edexcel, and for tool choices by subject, AI tools for A-Level teachers, subject by subject.

FAQ

How do I create custom mock exams for A-Level that genuinely mirror the real paper? Build from real Cambridge or Edexcel past-paper questions, then match four things the real qualification has: its multi-component structure (Paper 1/2/3), its assessment-objective weighting, its synoptic and extended-response content, and its exact marks and timing. Assemble from a past-paper bank so the questions arrive with mark schemes and known tariffs already attached, and check your draft against an actual recent paper.

Why does A-Level fidelity matter more than at IGCSE? A-Level mock grades feed predicted grades, which feed university applications. A paper that’s the wrong shape — too much recall, wrong AO balance, essays that are too short or mistimed — produces a score that looks like a prediction but isn’t. The higher stakes pull the accuracy bar up.

What’s the difference between an A-Level test builder from past papers and a normal quiz maker? A past-paper builder gives you questions that already carry examiner mark schemes, calibrated difficulty, exact command words, known tariffs and AO breakdowns — the things that make A-Level fidelity possible. A blank-page quiz maker gives you none of that, so you’d have to reconstruct the mark schemes and guess the weighting yourself.

How do I handle the extended, essay-style answers in a custom A-Level mock? Include the real high-tariff questions at their real mark allocations so students rehearse sustained argument under time pressure, and bring the levels-of-response mark scheme with each one. Mark the point-based bulk automatically, but flag the extended answers for your own review against the descriptors — more on that in AI marking for A-Level extended responses.

Can I reuse the structure for the next mock? Yes. Once you’ve built a component with the right AO weighting, synoptic balance, essay tariffs and timing, save it as a template and swap in fresh past-paper questions next cycle. The fidelity carries over, so the second mock takes a fraction of the time the first did.

The bottom line

At A-Level, speed is the second question; fidelity is the first. Build from real Cambridge and Edexcel past-paper questions, mirror the component structure, match the AO weighting, include the synoptic and extended content, and get the marks and timing exam-accurate. Do that and the mock predicts instead of misleads — and the score it produces is one you, your students and the people reading their predicted grades can actually trust.

Build a faithful A-Level mock free →

Ready to Excel in Your Studies?

Get personalised help from Tutopiya's expert tutors. Whether it's IGCSE, IB, A-Levels, or any other curriculum — we match you with the perfect tutor and your first session is free.

Book Your Free Trial