Exam Strategy: Time, Question Selection, Guessing
Two students who know the same amount get different marks
The exam doesn't measure what you know. It measures what you can produce, in order, under time pressure, on that specific paper — and those are separate skills from knowing the material. Every year students who understood the content walk out having lost marks they didn't need to lose: ran out of time, spent twenty minutes on a four-mark question, left a section blank, answered the question they wished had been asked instead of the one that was.
This lesson is the tactical layer on top of actually knowing the content (that's LN-01/LN-02). It won't save you if you didn't learn the material — nothing will — but if you did learn it, these tactics are free marks that most people leave on the table. It splits into three: before (know the paper), during (time and triage), and the guessing maths.
What you'll have at the end
- A time budget you compute before the exam, not mid-panic
- A triage order that banks easy marks first
- A guessing rule that matches the paper's mark scheme
Before: know the paper before you sit it
Half of exam strategy is done in advance, by removing surprises:
- Get the format cold. How many sections, how many questions, which are compulsory, how long, total marks. Walking in already knowing the shape means zero time lost orienting.
- Study the mark scheme. This is the highest-leverage and most-skipped move. Mark schemes tell you exactly how marks are awarded — a "6-mark" question wanting six distinct points, a "compare" question needing both sides, the specific command words (describe vs explain vs evaluate mean different things and score differently). Practising against the mark scheme is practising the actual target.
- Do past papers under time. The single best exam-technique practice there is: it rehearses the format, the timing and the question style simultaneously, and turns the real thing into a repeat.
During: the marks-per-minute rule
Time is the currency, and marks are what you're buying with it. So the rule is simple: spend time in proportion to marks.
- Compute your budget up front. Total minutes ÷ total marks = minutes per mark. On a 120-minute, 100-mark paper that's ~1.1 min/mark, so a 20-mark question gets ~22 minutes and a 2-mark question gets ~2. Write the per-question time on the paper as you start.
- The opening-minutes tax. Spend the first two minutes reading the whole paper and noting time targets. It feels like lost time; it prevents the disaster of discovering a whole section with ten minutes left.
- Hard cap every question. When a question's time is up, move on even if it's unfinished. This is the discipline that saves the most marks: the last 20% of a hard question costs the same time as the first 60% of two easy ones you haven't reached. Leave space and come back if time allows.
The classic failure is the perfectionist one — grinding a hard question to completion while easy marks in the next section go unclaimed. A finished exam beats a perfect first half.
During: triage — bank the easy marks first
You don't have to answer in order. Do a fast triage pass:
- First pass — the certainties. Go through and answer everything you know cold, fast. This banks guaranteed marks early (protecting you if you run out of time) and, as a bonus, calms the nerves and often jogs loose the harder answers.
- Second pass — the workable ones. Return to questions that need thought but you can do.
- Last — the hard/uncertain ones, with whatever time and the guessing rule below.
On "choose N from M" questions: decide fast and commit. Pick the ones where you can score the most, not the ones that look most interesting, and don't start one then switch three paragraphs in — that's double the time for one answer.
The guessing maths
Whether to guess depends entirely on the mark scheme, and it's a straight expected-value calculation (the general tool is DT-10):
- No penalty for wrong answers (most exams). Never leave a multiple-choice or short answer blank. A blank scores zero for certain; a guess has a positive expected value, however small. Even a pure random guess on a 4-option question earns you 25% of those marks over the paper. Always put something.
- Eliminate first. Crossing off two obviously-wrong options turns a 1-in-4 guess into a 1-in-2 — the same guess, double the odds. Do this before guessing.
- Negative marking (penalty for wrong). Now guessing has a cost, and you weigh it: if wrong answers lose ¼ mark and there are 4 options, a random guess breaks even (so guess only if you can eliminate at least one option, tipping expected value positive). If you can eliminate two, guess. Blind guessing under heavy negative marking can genuinely cost you — read the rule and do the small sum.
- On written questions, "blank" is always wrong. Even a partial, structured attempt picks up method or knowledge marks. Write the relevant formula, define the terms, show the setup — mark schemes reward the pieces.
And your first instinct: despite the folklore about second-guessing, changing an answer is roughly as likely to fix a wrong one as to break a right one — so change an answer only when you have a concrete reason (misread the question, spotted an error), not on vague doubt.
After you finish early — don't leave
If you have time at the end, you are not done — you're in the checking phase, which is worth real marks:
- Re-read each question stem and confirm you answered that question (not the one you assumed).
- Check the mechanical stuff: units, signs, decimal points, "all that apply", word limits.
- Fill any blanks with your best guess (per the rule above).
- Only leave when there's genuinely nothing left to gain.
What this means for you
- The exam measures production under time pressure, a separate skill from knowing the content.
- Read the mark scheme — it's the actual target, and almost nobody studies it.
- Budget time by marks (total min ÷ total marks) and hard-cap every question. A finished paper beats a perfect first half.
- Triage: bank the certainties first, then the workable, then the hard.
- No penalty → never blank; eliminate then guess. Negative marking → do the small expected-value sum (DT-10).
- Change an answer only with a concrete reason, not vague doubt.
- Finishing early means checking, not leaving.
Exercise (40 min, verifiable output)
- Pick one real upcoming exam. Find its format and mark scheme.
- Compute the time budget: total minutes ÷ total marks, and the target minutes for each section/question type. Write it as a one-line plan.
- Write your triage order and your guessing rule for this paper (check whether it has negative marking; if so, do the break-even sum).
- Do one past paper (or section) under strict time, running the protocol: two-minute read, budget, triage, hard caps, guessing rule, check.
- Note where you lost time or marks to technique rather than knowledge — that's the fixable part.
✅ Finish check: a one-page exam-day protocol for a real exam (time budget + triage order + guessing rule), tested once on a timed past paper with the technique-losses noted.
Summary card
- The exam scores production under time, not knowledge alone. Both are trainable.
- Before: know the format and study the mark scheme; do timed past papers.
- Time by marks: minutes ÷ marks; hard-cap every question; a finished paper beats a perfect half.
- Triage: certainties first (banks marks, calms nerves), then workable, then hard.
- Guessing: no penalty → never blank, eliminate then guess. Negative marking → do the EV sum (DT-10).
- Change answers only for a concrete reason. Finishing early = checking, not leaving.
Sources
- Dunlosky, J. et al. — Improving Students' Learning With Effective Techniques, 2013
- Kahneman, D. — Thinking, Fast and Slow, 2011
- Benjamin, L. et al. — Staying With Initial Answers on Objective Tests, 1984
Next lesson: LN-04 — Deep Work and Focus: Training Attention as a Skill (L2) Related: LN-08 Test Anxiety and Thinking Under Pressure · LN-01 Active Recall · DT-10 Decision Frameworks (Expected Value) · LN-02 Building Your Own Study System Path: related — exam-season core