How to Predict Exam Questions From Past Papers

5 July 2026  ·  7 min read

Every student who's ever flicked through five years of past papers has had the same thought: this topic keeps coming up. That instinct is correct, and it's the basis of a legitimate revision strategy — using historical exam patterns to predict which topics are most likely to appear again, and prioritising your revision accordingly. Predicting exam questions doesn't mean guessing exact wording; it means identifying which parts of the syllabus examiners keep returning to, and making sure you know those cold.

Why Past Papers Aren't Random

Exam papers are written to a syllabus, and syllabuses don't change every year. Module leaders often reuse question structures, rotate through a fixed set of core topics, and are constrained by what's actually been taught that year. This means past papers contain a genuine signal about what's likely to reappear — not because questions are recycled verbatim, but because the underlying topics, question formats, and areas of emphasis tend to repeat.

This is different from cramming a "question bank" and hoping for a repeat. It's closer to reading the syllabus the way an examiner does: which topics are foundational enough that they get tested almost every year, and which ones only show up occasionally.

Step 1 — Collect as Many Past Papers as You Can

Start by gathering five to ten years of past papers for your module, if they're available. University repositories, module pages on Canvas or Moodle, and course reps are the usual sources. If your course has changed significantly in that time, weight the more recent papers more heavily — older papers may reflect a syllabus that no longer exists.

Step 2 — Map Every Question to a Topic

Go through each paper and label every question with the topic or theme it covers, using the same categories as your course syllabus or lecture list. For example, in a contract law module, questions might map to categories like "offer and acceptance," "consideration," "misrepresentation," and "remedies for breach." Build a simple table: rows are topics, columns are years, and you tick which topics appeared in which year.

This step is the most tedious part of manual topic prediction, especially across a decade of papers, but it's what turns a vague feeling ("this comes up a lot") into an actual frequency count you can act on.

Step 3 — Rank Topics by Frequency and Recency

Once you have your table, you'll usually see a clear pattern: a handful of topics appear almost every year, a larger group appears roughly half the time, and some appear rarely or only once. Rank topics using two factors:

• Frequency — how many of the last several years included this topic
• Recency — whether it appeared in the last one or two papers, which can indicate it's still active in the current syllabus

Topics that are both frequent and recent deserve the most revision depth. Topics that appeared once, five years ago, and never since are lower priority — not zero priority, since a genuinely comprehensive answer still requires baseline knowledge of the whole syllabus, but lower relative to your time budget.

Step 4 — Cross-Reference With Lecture Emphasis

Past papers tell you what's been asked historically; your current lectures tell you what your specific lecturer is emphasising this year. If a topic got an unusually long lecture, a dedicated seminar, or repeated mentions across several sessions, that's often a signal it matters to whoever is setting or moderating the exam — even if it hasn't appeared often historically. Combining historical frequency with current-year emphasis gives a much stronger prediction than either signal alone.

Step 5 — Revise Ranked Topics With Depth, Not Just Breadth

Once you have a ranked list, allocate revision time proportionally. High-ranked topics deserve full practice — writing out full answers, working through multiple past questions on that topic, and being able to discuss it from several angles. Lower-ranked topics still need baseline coverage (you can't skip parts of the syllabus entirely), but with less depth. For the actual revision technique once you've picked your priority topics, our guide on revising with past papers topic by topic covers how to structure that practice.

This kind of prioritised revision pairs well with spaced repetition — reviewing your highest-priority topics on a schedule that keeps them fresh right up to the exam, rather than cramming them once and hoping they stick.

Where This Method Struggles (and How to Fix It)

Manual topic mapping across a decade of papers, done properly, can take several hours — time many students don't have during a busy term. It's also easy to make mistakes: miscategorising a question, missing a paper, or letting recency bias skew your rankings because the most recent exam is freshest in your memory.

This is exactly the kind of pattern-recognition task that's well suited to automation. Corvo analyses the past papers and lecture material you upload — whether for a law module tracking recurring case-based questions, or any other subject — and ranks topics by how often and how recently they've appeared, building you a prioritised revision list without hours of manual cross-referencing.

FAQs

Can you actually predict exam questions?

You can't predict exact wording, but you can predict which topics are statistically more likely to appear, based on past-paper frequency, syllabus centrality, and recent lecture emphasis. This is topic prediction, not question prediction, and it's a widely used, legitimate revision strategy.

How many years of past papers should I analyse?

Five to ten years is a good range for most modules, provided the syllabus hasn't changed significantly. If your course has been restructured, focus on papers from after the change.

Is it risky to focus revision on predicted topics only?

Yes, if done in isolation. Use topic prediction to prioritise depth and practice time, not to replace broad syllabus coverage entirely. Cover everything at a baseline level, then go deeper on the most likely topics.

How does Corvo rank topic likelihood?

Corvo analyses the past papers and lecture material you upload, identifies recurring topics and question types, and ranks them by frequency and recency — giving you a prioritised revision list without manual cross-referencing.