You now have the tools, and you know how to tell whether they're working. The campaign is the question the tools can't answer on their own: given more to learn than there are hours, and a date you can't move, what do you spend tonight on? Most study books either skip this or bury it under elaborate planner systems. We'll do the opposite — name the few honest decisions that decide most of the outcome, and then send you straight back to the tools.
Hold onto one thing as you read this part: a plan is not a technique. It's the scaffolding that makes sure the techniques actually happen, on the right things, in time. It earns no marks by itself. So spend just enough on it to be realistic, then stop planning and go retrieve. Five decisions: who gets your hours (triage), in what order (passes), what the room itself demands (the day), what to protect around the work (conditions), and when to walk away from a topic (stopping).
Your hours are the scarce thing — not your willingness, not your tools, your hours — so the first campaign decision is also the most consequential: decide depth by the marks, not by your interest. Some topics are worth a lot and asked often; some are worth little and asked rarely; some are fascinating and barely tested. The hardest honesty in this whole book is admitting that the topic you enjoy and the topic that pays are often not the same one.
A simple triage does the work. Sort every topic by two questions — how much is it worth? and how shaky am I on it? — and four quadrants fall out, only one of which deserves the bulk of your effort.
The mistake almost everyone makes is pouring hours into high-value-but-solid — it feels wonderful, you already know it, every rep confirms how much you know — and into low-value-but-interesting, because it's enjoyable. Meanwhile high-value-and-shaky, the only quadrant that actually moves your score, stays scary and quietly avoided. That avoidance is not laziness; it's the familiarity trap wearing a productive face. Re-studying what you own feels like progress because it's fluent. The quadrant that moves your score is the one that feels worst to open. Name the four quadrants honestly and the plan more or less writes itself — your hours go where you're weak and it counts.
Put your whole syllabus on one page — every topic, nothing left off. Beside each, write two letters: value (H or L — how much is it worth, how often asked) and state (S for solid, shaky otherwise). Now find your high-value, shaky topics. That short list — not the whole page, that short list — is where this week goes. It will be the list you least want to look at. That's how you know it's the right one.
"But what if they ask the topic I deprioritised?"
They might — and a light pass to not-zero on the low-value topics is exactly your insurance against that, which is why triage keeps them rather than dropping them. What triage refuses is letting a rarely-asked topic eat the hours that a high-value-shaky one needed. You are not gambling the syllabus; you are refusing to over-insure the cheap risks at the cost of the expensive ones. The candidate who covers everything to the same shallow depth, terrified of every gap equally, is the one who walks in with no topic strong enough to carry them.
With the hours aimed, the next question is order — and the instinct here is almost always wrong. The instinct is to start at topic one, study it to perfection, and only then move to topic two. It feels thorough. It is the worst order there is, because a blank topic scores zero no matter how brilliant your best topic is. You almost always score more by getting every topic to a passing, connected grasp first, and only then deepening where the marks say it's worth it.
So go wide before you go deep, and plan in passes, not in one perfect sweep. You will not learn the syllabus once and be done; you'll go over it several times, each pass deeper and more retrieval-heavy than the last. Pass one builds a connected map of the whole ground (encoding — Part 2's building tools). Pass two and beyond drill it closed-book and space the returns (the engine — retrieval and spacing). This is not extra work bolted on; it is exactly the shape spacing wants anyway. A topic you touched in pass one and return to in pass three has been spaced for free. Planning in passes is the one move that lets you go wide and still get deep.
Then comes the part that wrecks more plans than laziness ever did: the calendar. Estimate from what things actually took, not from what you hope. When you lay the passes against real dates you will badly underestimate how long each takes — everyone does, every time; it is one of the most reliable findings there is. The fix is the outside view: don't ask "how long should this revision take?"; ask "how long did a revision like this actually take me last time?" Plan from your own history, then build in slack on top — because you'll still be too optimistic even after reading this warning.
Sketch your next three passes over the whole syllabus. Against each pass, write the time it would really take — not from hope, but from memory of the last time you did something like it. No honest memory to draw on yet? Then your first job this week is to time one real revision block, so your next estimate has a fact under it instead of a wish. Now multiply your total by 1.3 and see if the calendar still closes. If it doesn't, that's triage telling you what to cut — better to learn it now than in the final fortnight.
"Going wide and shallow feels like I'm not really learning anything."
It feels that way because a shallow first pass is shallow by design — that's the point, not a flaw. You are not trying to learn it deeply on pass one; you are laying down a connected map so that the deep passes have somewhere to attach, and so spacing has the whole syllabus to work on instead of one over-polished corner. The deep, it-finally-sticks feeling you're missing arrives on passes two and three, on top of the map. Skip the wide pass and the deep work has nothing to hang on — you learn each topic in isolation and forget it before you reach the next.
Everything so far has been about getting knowledge in and making it stick. The exam is a different, narrower skill: getting it out, under time, onto the page, in the form the marker rewards. Part 4 already did the upstream half — the room is a skill, and you train it by rehearsing the real conditions until it's boring. This section is the downstream half: assuming you've rehearsed, here is the minute-by-minute playbook for the room itself.
Empty your head onto the page first. In the opening minute of a paper — or of a section — write down what you're most afraid you'll forget: the framework, the formula, the dates, the structure you planned. Get it out of your anxious working memory and onto the margin where it can't escape, before the questions crowd it out. It's a retrieval you've practised a hundred times; do it once more, on demand, the instant you're allowed to write. The minute it costs is repaid many times over by the panic it prevents.
Answer the question the paper actually asked, in the shape it rewards. This is where the whole task-matching lesson lands. A recognition paper wants the fast, correct pick — don't argue an essay in your head over a multiple-choice question. A construction paper wants a built, structured answer — don't dump a pile of facts where a judgement was asked for. Read what the question is demanding ("explain", "evaluate", "to what extent"), give it exactly that, and stop. An answer that's brilliant about something slightly different earns the marks for the something-different: none.
Spend time in proportion to the marks. The commonest unforced error in any exam is over-investing in a question you like and starving three you don't. A question worth twice the marks deserves roughly twice the time — and no question deserves the time that pushes you past its share into the next one's territory. Decide your per-question budget from the mark scheme before you start, glance at the clock at planned checkpoints, and move on when the budget's spent, even mid-sentence. A half-answered high-value question beats a perfect low-value one sitting next to a blank.
Before your next study block, do one timed, closed-book past section instead. Full conditions, clock running, no peeking. Open with a one-minute brain-dump of what you fear you'll forget. Hold yourself to a per-question time budget set from the marks, and move on when each runs out. Afterwards, mark it ruthlessly and feed every miss back into spaced retrieval. That single session rehearses the room, trains the timing, and refreshes the content — three things at once.
"Brain-dumping and budgeting time just burns minutes I could be answering in."
Both pay for themselves several times over, which is exactly why the calmest high-scorers do them. The brain-dump costs sixty seconds and saves you the five minutes you'd otherwise lose to the mid-paper blank when a fact you "definitely knew" vanishes under pressure. The time budget costs a glance at the clock and saves you the disaster that loses real exams — the beautiful answer to question two written in the time that questions five and six needed. The minute spent planning the room is the cheapest insurance on the paper.
This section is short on purpose. Most study books are mostly this — sleep, focus, planners, motivation, stationery — blown up to fill a book. The honest evidence here is thin and the field is full of attractive myths, so the most useful thing this section can do is name the two or three conditions actually worth protecting, then name the filler so you can stop spending hours on it. The subtraction is the service.
Nothing here is a technique. The tools were what to do with your study time; this is what to protect around it. Get the tools right and the conditions wrong and you lose some of the gain; "optimise" the conditions and skip the tools and you have a beautifully scheduled waste of time. The tools come first. These two guard them.
Guard your attention while you encode. Of everything in this section, this has the firmest evidence: divided attention wrecks the moment of learning. A notification, a half-watched video, a conversation in the room while you take something in — the encoding comes out measurably worse. And it lands with a revealing asymmetry: divided attention damages putting new material in far more than pulling already-learned material out.
That asymmetry is the whole rule, and it's liberating rather than puritanical. It does not say "never have your phone near you." It says: single-task the hard new learning — defend that one window fiercely, phone in another room — and let the drilling of half-known material survive a noisier setting. This is also the answer for a reader with no quiet room: you don't need silence all day, only one protectable window for the hard part. A defended twenty minutes beats a distracted two hours. The bus, the queue, the noisy kitchen — fine for retrieval of what you half-know; useless for taking in something new and hard.
Protect sleep around hard learning — the honest version. You know sleep matters; the well-supported finding is narrow and worth stating exactly: sleep loss hurts the brain's ability to take in new information. So don't try to learn hard new material on a badly slept brain. The grander, popular claim — that a night's sleep does deep magic to reorganise what you studied that day — is weaker and far more contested than it's sold, and we won't lean on it. Keep the safe version: don't study hard new material exhausted; a decent night before a heavy learning day is an investment in that day.
Then cut the myths — this is hours handed back to you. Each of these is either disproven or far weaker than its popularity suggests, and cutting them is reclaimed time, not deprivation. Learning styles (visual / auditory / kinaesthetic) is the most popular idea in study advice and one of the most thoroughly disproven — study in the form the material demands, not a label you pinned on yourself, and drop it entirely. Motivation and confidence as a study method — believe-in-yourself spirals and motivation videos, done in isolation — don't raise learning; confidence is an output of competence built by the tools, not an input you install separately. The grand sleep-consolidation story — keep "don't learn exhausted," drop the elaborate REM-sorts-your-memories version. Elaborate planner systems — calendars are admin, not learning; plan only enough to protect the passes, then stop. And re-reading and highlighting — named one last time so the cut is complete: the defaults that feel like studying and build the least. Replace them with retrieval.
Name the single daily window you'll defend for hard new learning — when, where, and where the phone goes during it. Then write down three things from the myth-cut list above that you currently do, and stop them this week. Put the reclaimed hours straight into spaced retrieval. You've just upgraded your conditions without buying, scheduling, or believing anything.
"I study fine with music / the TV on / my phone beside me."
For re-drilling things you half-know, you might — that's the half this rule deliberately leaves alone. But the feeling of studying fine with distraction is itself unreliable, because the cost lands at encoding, the one moment you can't feel going wrong: the new material simply goes in thinner, and you discover the thinness later, when you can't retrieve it and blame the topic instead of the noisy room it was learned in. The test isn't whether it feels fine while you do it. It's whether the hard new thing is still there, cold, a week later. Defend the encoding window and find out.
One last campaign skill, and it's the one almost nobody practises: knowing when to stop working on a topic. Effort on any single topic shows sharply diminishing returns — the climb from blank to a solid, connected grasp is steep and cheap, and the climb from solid to flawless is shallow and ruinously expensive. Past a certain point you are buying marks that mostly aren't on the paper, with hours that a blank topic two columns over would have repaid many times over.
So the discipline is to stop a topic at "solid and connected" and move on, deliberately, before it's perfect — and to treat the urge to keep polishing a topic you already half-own as a signal that you're back in the comfortable corner of the triage grid. With finite, often broken time, what you choose not to study is as much a decision as what you study. A clean, honest "good enough — next" on a topic you've brought to passing strength is not a failure of thoroughness. It is the campaign working: thoroughness is a property of the whole syllabus at passing strength, not of one beloved topic at perfection while three others sit blank.
Look at what you studied this past week. Find the one topic you spent the most time on, and ask honestly: was it already at "solid and connected" before half those hours went in? If it was, you found your over-polish — name it. Then find the topic you most avoided (your triage will have flagged it), and move the over-polish hours there next week. You don't need more hours. You need them in the right column.
"Leaving a topic at 'good enough' feels like exactly the gap they'll punish."
The opposite is true, and the maths is on your side: marks are roughly proportional to coverage at passing strength, so a syllabus full of solid-and-connected topics scores far higher than a few perfect ones beside several blanks. The polish you're tempted to add is on a topic already scoring most of its available marks; the topic you're tempted to leave blank scores zero. Stopping early on the strong one to rescue the blank one is the move that protects you, not the one that exposes you. The real punishment comes from over-insuring what's already safe.
A campaign is not a longer to-do list; it's a few honest decisions made once and held. Aim your hours at what's weak and counts. Go wide before deep, in passes, on an honest calendar. Treat the room as its own skill. Protect attention and sleep; cut the rest. And stop polishing what's already strong. Decide these five well and the daily studying — which is where the real work lives — finally has somewhere worth going.