How to use past papers for GCSE (and when to start)

When to start: not week one, not the last week

A past paper is a test of whether revision has worked, so it needs something to test. Start when you can answer most of a topic's questions from memory, which for most GCSE students is two to three months before the exam.

Starting too early gives you a paper full of questions on content you have not covered, a score that tells you nothing except that you have not covered it, and a dented confidence you did not need. Starting too late — the classic first full paper the weekend before the exam — leaves no time to do anything about what you find. The whole value of a paper is in the follow-up, and follow-up needs days, not hours.

Before then, papers still have a use, just not as papers. Pull single questions on a topic you have just revised, or a section of a paper, and treat them as practice questions rather than as a mock. That is fine from the start of revision. What you are delaying is the full, front-to-back paper.

Get papers, mark schemes and examiner reports from your exam board's own website — AQA, Edexcel, OCR, WJEC, Eduqas and CCEA all publish past series free, usually a year or so after the exam sat. Check the exact specification code with your teacher first, because a paper from the old course will have a different structure. Your exam dates are on the JCQ key dates and timetables page, and the calendar you build backwards from should start there.

The first paper: untimed, and that is fine

Your first full paper in a subject should be untimed, with the notes shut. You are finding out what you know, not yet how fast you know it.

Timing is a separate skill from knowing the content, and trying to learn both in the same sitting means you learn neither well. An untimed first paper shows you which topics you can answer, which you half-know and which are blank — information a timed paper hides, because a rushed answer looks the same as an unknown one on the mark scheme. The notes stay shut for the same reason: an open-book paper tells you where the answers are in your folder, which is not the question.

Timed papers come once you are scoring reasonably on untimed ones — typically the second or third paper in a subject. From then on, every paper is timed, because the exam is, and the last two or three before the real thing should be done in a single sitting at the full time with the phone in another room.

Marking with the mark scheme, honestly

Marking is the half of the exercise that produces the learning, and it is the half most people rush. Give it as long as the paper took.

Use the board's mark scheme, not your sense of whether an answer was "basically right". Mark schemes are specific about which words earn marks and which do not, and the gap between what you meant and what you wrote is where GCSE marks disappear. The mark scheme is the only calibration you have.

• Give yourself nothing for a near-miss. If the scheme wants "increases the rate of reaction" and you wrote "makes it faster", ask whether an examiner would accept it. If you are not sure, it is a zero and an entry in the log.
• Mark the method, not just the answer. In maths and science, a right answer with no working and a wrong answer with sound working are both worth looking at, and the scheme tells you how each is credited.
• Read the whole scheme for long answers. Extended-response questions in English, history, geography and the sciences are marked against level descriptors. Find the level your answer sits in and read the one above it — that is the instruction for next time.
• Write the lost marks down before you move on. Not "revise topic 4" — the specific thing that lost the mark, in your own words.

The temptation to be generous is strong, because a higher score feels better. But the score is not the product; the list of lost marks is. A harshly marked paper with a full error log is worth more than a kindly marked one that tells you everything is fine.

The error log

Keep one table per subject, across every paper you do. It is the most useful document you will own by May, because it shows you the same mistakes coming back.

The "type" column is the one that matters. After two or three papers it tells you whether your problem is content you do not know, technique you have not learned, questions you are not reading, or care you are not taking under time pressure. Each of those has a different fix, and "revise more" is only the fix for the first one.

Before every new paper, read the log. The entries from the last one are the marks you are most likely to win back, and reading them first turns the new paper into a check on whether the fix worked.

Examiner reports: the cheapest marks available

Every exam series, the chief examiner writes a report on how students did, question by question. Almost nobody reads it, and it is the closest thing to being told what loses marks by the person who decides.

Examiner reports sit next to the mark schemes on the board's website. For each question they describe what strong answers did, what common wrong answers looked like, which words students used that did not earn marks, and where whole cohorts misread the question. When you have marked a paper and lost marks on a question, read the report's paragraph on that question before you write the log entry — very often your mistake is in there, along with the exact phrasing the examiner wanted.

A single evening reading the last two or three reports for a subject, with a highlighter, will change how you answer several question types.

Space the papers out

One paper, properly marked and followed up, per subject per week or fortnight beats three in a weekend. Spacing is not a scheduling convenience; it is where the learning happens.

The evidence on distributed practice is consistent: the same amount of practice spread across days produces better retention than the same amount massed together, and the gap should grow as the exam gets further away. For papers that means a rhythm — paper, mark, log, fix the gaps over several short sessions, then the next paper — rather than a papers weekend. The fix sessions between papers are the point; the papers are how you find out what to fix.

A workable schedule for one subject with two exam papers, starting ten weeks out: an untimed paper in week one, follow-up in weeks two and three, a timed paper in week four, follow-up, a timed paper in week seven, and the final timed paper under full conditions in week nine, leaving the last week for the error log and nothing new. Work back from the date on the JCQ timetable, not forwards from today.

Turn lost marks into flashcards and questions

A lost mark is a question you could not answer. Turn each one into a question you will be asked again, so the paper keeps working after you have finished it.

Content errors become flashcards: the definition, the equation, the four causes, the labelled diagram. Technique errors become a rule on a card you read before the next paper: "explain means say why", "two marks means two points". Reading errors become a ritual: underline the command word and the number before writing anything. Careless errors become a check you do at the end of each question.

Then actually test yourself on the cards, closed, a few days later and again a week after that. That is retrieval practice, and it is what turns a marked paper into memory rather than a score you once got. If you use a flashcard app, this is where it pays for itself: FacilStudy will take a photo of a page of corrected answers or an error log and draft flashcards and a quiz from it, then bring them back on a spaced schedule alongside the rest of your subjects. It is pre-launch, so there is a waiting list rather than something to download, and index cards and a wrong pile do the same job by hand.

How many papers is enough — and what they cannot do

Three to five full papers per exam paper, each properly marked, logged and followed up, is enough for most students. More than that with no follow-up is worse than fewer with it.

Some subjects only have a handful of papers on the current specification, and that is fine: use the specimen and sample assessment materials the board publishes, and go back to the older specification for individual questions on topics that have not changed. When you run out, redoing a paper you did two months ago is a legitimate test, because you will not remember the answers, only the topics.

The reverse is also true. A student who does ten papers without ever fixing the content gaps is practising being wrong faster. Papers find the gaps; the sessions in between close them; and the daily fifteen minutes on your weakest topics is what keeps the closed gaps closed. For how that daily session fits around papers in the final weeks, see the 8-week GCSE revision timetable.