Two weeks out, the arithmetic changes. Anything you begin now has to be learned and retained with no time for the spacing that makes retention work. Meanwhile everything you learned in the previous six months is decaying, and a reminder costs a fraction of what new material costs.
Stop doing
- Starting subjects you have never opened. The return is poor and the anxiety cost is high. Accept the gap.
- Chasing rare facts. At this range, commonly examined material you are shaky on is worth several times more per hour.
- Long single-subject blocks. They feel productive and leave the other nineteen subjects decaying while you work.
- Comparing with other candidates. It has no informational content at this point and reliably costs sleep.
Keep doing
- Your confident errors. Highest value per minute of anything available. These are marks you are currently guaranteed to lose.
- Timed sets of 36. Pacing is a skill and it decays too. Two or three a week keeps the rhythm.
- Scheduled review. Exactly what it was built for. Let the queue decide.
- Arbitrary material. Values, criteria, classifications. Rehearsal works on these and works quickly.
In the last fortnight, breadth beats depth. A shallow pass over everything beats a deep pass over one thing.
The plan for the last fortnight
Coverage over depth, retrieval over reading, and nothing new. A shallow pass across everything you have already met beats a deep pass over any one thing, because at this range the marginal mark comes from stopping decay rather than from adding understanding.
- Daily: the review queue, and a timed set of 36.
- Every few days: your confident-error list. It is short, specific, and worth more per minute than anything else available.
- Weekly: one full mock, until the final three days.
- Never: a subject you have not opened. Accept the gap and defend what you have.
Sleep is a revision technique
Consolidation happens during sleep, which makes the choice between an extra hour of questions and an extra hour of sleep less obvious than it feels at midnight. In the last week, when the marginal hour of study is worth very little and the marginal hour of sleep protects everything already there, the trade is clearly one way.
The day before
Little or nothing. Your own error notes if you want the reassurance of doing something. Check the logistics — centre, documents, timing — because a logistical failure costs more than any fact you could learn. Then stop; nothing learned on that day will be retrievable the next, and the attempt costs composure that will be.
In the final week you are protecting what you have, not adding to it.
The final few days
Reduce input. Sleep matters more than an extra hour of questions, and it is increasingly likely that an extra hour makes things worse rather than better. Go over your own error notes rather than any book — they are already filtered to the things you specifically get wrong, which is the highest-density revision material you will ever have.
Do not sit a full mock in the last two or three days. A bad score at that point cannot be acted on and can cost you the composure that is worth more than anything you would have learned from it.