Start from the format, not the content
Before any schedule, get concrete about what the exam is. How long, how many questions, what proportion multiple choice against written, whether anything is permitted in the room, how it is weighted across topics, whether wrong answers cost anything. Instructors and awarding bodies publish this and it is routinely ignored by students who then prepare for a different exam than the one they sit.
This matters because retrieval practice is specific to what you practice retrieving. Preparing for an essay exam by doing multiple-choice questions trains recognition when you need production. Preparing for a closed-book calculation exam with the formula sheet open trains a skill you will not have. Match the practice to the format before you match it to the content.
Policies, weightings and permitted materials differ between institutions and change from year to year, so take them from the official source for your specific exam rather than from a friend's memory of last year.
Working backward
| When | What that period is for | What it is not for |
|---|---|---|
| 6+ weeks out | First coverage of everything, with self-testing built in from the start. Identify the topics you have never actually understood. | Past papers, which will only tell you what you have not covered yet |
| 4 weeks | One full timed practice paper to calibrate. Spaced review of covered topics. Build the error log. | Starting new topics from scratch, unless something was missed entirely |
| 2 weeks | Weighted practice: hours go to the topics with the worst first-attempt accuracy and the highest exam weight. | Comfortable review of things you already get right |
| 1 week | Full papers under timed conditions, at the time of day the exam happens. Fix what the papers expose. | Reading new chapters |
| 2 days | Light retrieval on the error log, one-page summaries, logistics | Anything that produces a new list of things you do not know |
| Night before | Short review, then stop. Sleep is the productive activity. | Cramming that costs you tomorrow's attention |
Doing this by hand across four subjects is where good intentions die. Study plan calculator divides a fixed amount of material across the days remaining, study hours target works backward from a total-hours goal to a daily number, and deadline backward schedule lays out milestones from the date rather than from today. Fill the resulting slots on a timetable and the plan stops being a feeling.
Budget honestly. The most common planning failure is assuming eight clear hours on a Saturday that has never once contained eight clear hours. Count the hours you have actually managed in recent weeks and plan for that number, because a schedule you abandon in week two is worse than a modest one you keep.
Practice papers are the highest-yield hours you have
A timed paper under real conditions does three things that reviewing cannot. It gives you retrieval practice in exactly the format that counts. It tells you about pacing, which is a large fraction of the marks lost in timed exams and is invisible in untimed practice. And it surfaces a specific list of your gaps rather than a general sense of unease.
Real conditions means real: the whole paper, in one sitting, timed, no notes, no phone, no pausing to look something up when you get stuck. The version where you do half of it, check the answers as you go, and take a break in the middle produces a flattering score and teaches you very little. When you hit a question you cannot do, leave a mark and move on, which is also what you will need to do on the day.
The marking pass matters more than the paper. For every item you got wrong, write down which of these it was: did not know the content, knew it and misread the question, knew it and ran out of time, knew it and made an arithmetic or careless error. Those four have four different fixes, and lumping them together as "got it wrong" is how people spend weeks reviewing content when their actual problem was pacing.
The error log
The single most useful document in a revision period is a running list of what you got wrong, why, and what the correct approach was. Not the questions themselves — the pattern. Three entries that all say "confused these two tests" is information; thirty photocopied questions is not.
| Column | What goes in it |
|---|---|
| Topic | Narrow enough to be actionable. Not "statistics" but "choosing between paired and unpaired comparisons". |
| Error type | Content gap, misread, timing, or careless. Keep these separate. |
| The correct approach in one line | Written by you, not copied from the solution. If you cannot write it, you have not fixed it. |
| Retest date | A few days out. Then again later if you get it right. See spaced repetition. |
| Result on retest | The only column that tells you whether the fix took |
Track first-attempt accuracy by topic rather than hours studied. Hours measure diligence and are easy to feel good about; first-attempt accuracy measures readiness and is not. Where the two disagree, the second one is right. If you are keeping running scores across practice sets, test average calculator and grade rank calculator do the arithmetic, and GPA calculator handles the term-level version.
Anxiety, and the part of it that is procedural
Exam anxiety is common and it is not uniformly bad — moderate arousal generally helps performance, and the problem is the level at which it starts consuming working memory that you needed for the questions. Some of it is genuinely reduced by preparation and by practice under conditions that resemble the real thing, which is a large part of why full timed papers are worth so much: the room feels less novel when you have simulated it six times.
Procedural things that help and cost nothing: know the logistics precisely — room, time, what to bring, how long the journey takes with something going wrong — so that none of it is a live question on the morning. Arrive with time to spare but do not stand in a group comparing revision, which reliably raises everyone's anxiety and teaches nobody anything. In the room, read the whole paper first and start with a question you can definitely do, because getting something on the page early settles the physiology.
Where anxiety is severe — panic symptoms, blanking on material you demonstrably know, avoidance of the exam itself — that is worth raising with a doctor or your institution's student services rather than treating with study advice. Institutions generally have processes for this, they differ, and the only way to find out what yours offers is to ask them directly.
The last 48 hours
There is a point where adding material stops paying and starts costing, because every new topic you open produces a fresh list of things you do not know, at exactly the moment you can least afford to feel unprepared. For most exams that point is roughly two days out.
What earns its place in the last two days: light retrieval over the error log, reading one-page summaries you made earlier, one timed section rather than a full paper, and getting the logistics settled. What does not: opening a chapter you skipped, comparing notes with a panicking classmate, or reorganizing your files, which is procrastination with a productive costume on.
Sleep across both nights, not just the last one, since a single good night does not undo a week of four-hour ones. The all-nighter deserves its own sentence because people keep doing it: you gain a few hours of low-quality study and pay with degraded attention, slower processing and worse recall on the morning that counts, and for most exams that trade loses. If your sleep has already been wrecked for a fortnight, sleep debt calculator will quantify it and sleep hygiene covers the repair, though the honest answer is that the repair takes longer than you have.
Afterward
Write down, within a day, what actually happened: which topics came up, where the time went, which of your four error types cost you marks. You will not remember this in three months and you will be preparing for another exam then. That page is the most accurate revision advice you will ever receive, because it was written about you.
And then leave it alone. Post-mortems conducted with friends in the corridor, reconstructing answers you can no longer change, are a way of extending the exam by several hours at no benefit to anyone. Write the page, close the notebook, go outside.
Questions people ask
How far in advance should I start revising?
Far enough that spacing is available, which for a substantial exam means at least four to six weeks rather than a week of intensity. The mechanism matters more than the number: material reviewed at widening intervals is still there on the day, and material covered once in a block decays fast. If you genuinely only have a week, the schedule compresses — cover by exam weight, self-test rather than review, and accept that you are optimizing for the exam rather than for keeping the material.
Are practice papers better than more revision?
For most people, considerably, and it is not close. A timed paper gives retrieval practice in the exact format that counts, exposes pacing problems that untimed work hides entirely, and produces a specific list of gaps instead of a vague sense of not being ready. The condition is that it has to be done properly: whole paper, timed, closed book, no pausing to check. A half-paper done with the notes open produces a flattering score and very little information.
What should I do the night before an exam?
Light retrieval on your error log and your one-page summaries, the logistics sorted so nothing is a live question in the morning, and then sleep. Two days out is roughly where opening new material stops paying, because each new topic produces a fresh list of things you do not know at the worst possible moment. Sleep is not the passive option here — consolidation happens during it, and sitting an exam on four hours costs you attention and recall on the morning it counts.
Does cramming work?
For a short-term bump, yes, and if the exam is tomorrow and you have nothing, hours in front of the material beat no hours. What it does not do is last: crammed content is substantially gone within a week or two, which makes it a poor plan for cumulative subjects or anything you will build on next term. The specific version that usually loses is the all-nighter, which buys a few hours of low-quality study in exchange for a measurable next-day deficit.
How do I handle exam anxiety?
Some of it is procedural and responds to preparation: full timed practice papers make the real thing less novel, knowing the logistics precisely removes a set of live worries, and starting the paper with a question you can definitely do settles things physically. Moderate nerves are normal and not harmful. Where it is severe — panicking, blanking on material you demonstrably know, avoiding the exam — that belongs with a doctor or your institution's student support rather than with study advice, and what support is available differs by institution, so ask yours directly.