A mock is a diagnostic, not a score
The instinct is to treat a mock as a rehearsal — sit it when you're ready, do well, feel reassured. That gets the value backwards. The information is in the failures, so a mock sat when you're ready is a mock that tells you the least.
Sit it earlier than feels reasonable. The discomfort of doing badly on a paper three weeks out is the price of finding out what to do with those three weeks, and it is cheap at the price.
Timing
| When | What it's good for | What it can't do |
|---|---|---|
| 4–6 weeks out | Discovering which topics you can't do at all | Nothing about timing or stamina — you'll be too slow to finish, which is fine. |
| 2–3 weeks out | The main event. Content gaps plus technique, with time to fix both | This is the one that matters most. Sit it properly. |
| 1 week out | Timing, stamina, question selection | Too late for content gaps. Anything it finds now is triage, not repair. |
| 2 days out | Confidence, if it goes well | Nothing actionable. If it goes badly it costs you more in morale than it gives. |
Conditions that make it a real test
Every comfort you allow yourself moves the result away from the thing you're trying to measure. The point of the conditions isn't discipline for its own sake — it's that a mock sat with notes open measures a different skill than the exam does, so its result can't be trusted.
- Full duration, in one sitting, including the reading time. Stamina is a real variable and paper three of a day is where marks vanish.
- No notes, no phone, no looking anything up. The moment you look one thing up, the score becomes meaningless as a diagnostic.
- The same time of day as the real exam where possible. A 9am paper practised at 10pm trains a state you'll never be in.
- Written by hand if the exam is handwritten. Typing speed and handwriting stamina are not interchangeable.
- Somewhere that isn't your desk. Context-dependent recall is a modest effect, but an unfamiliar room is a free approximation of the hall.
Marking is where the value is
- 1
Mark it the next day, not immediately
Immediately after, you remember what you meant, and you'll award marks for intentions that aren't on the page. A day's gap makes you read what you actually wrote — which is what the examiner will do.
- 2
Use the official mark scheme, harshly
If a mark requires a term you didn't write, you don't get it. Generous self-marking is the single commonest way a mock becomes worthless, because it produces a reassuring score and no diagnosis.
- 3
Categorise every lost mark
Four buckets: didn't know it, knew it but wrote it badly, ran out of time, misread the question. This categorisation is the actual product of the mock — the score is almost irrelevant next to the distribution.
- 4
Count the marks per bucket and act on the biggest
If half your losses are timing, more content revision is the wrong response and would have been your instinct. If half are misreading, the fix is a two-minute reading protocol, not another week of study.
- 5
List the three topics that cost the most marks
Not the topics you found hardest — the ones with the largest mark loss. Those are different, and the mark loss is the one that matters.
- 6
Rewrite one answer properly, to the mark scheme
One. Take your weakest answer and write the version that would have scored full marks. Doing this once teaches you more about what the scheme rewards than reading ten model answers.
What the categories tell you
| Dominant loss | What it means | The next two weeks |
|---|---|---|
| Didn't know it | Genuine content gaps | Targeted revision on the three highest-cost topics. Test, don't re-read. |
| Knew it, wrote it badly | Technique, not knowledge | Practise answer structure against mark schemes. Plans, not full essays. |
| Ran out of time | Pacing and question selection | Timed single questions with a stopwatch. Write finish times in the margin. |
| Misread the question | Reading protocol | Three-pass question reading, underline the instruction verb, every time. |
The reason this matters: all four feel like "I need to revise more" from the inside, and only one of them is. Students who lose marks to timing and respond with more content revision go into the real exam with the same problem and more facts.
The mock that goes badly
A bad mock three weeks out is good news delivered unpleasantly. It has found, at the only point where finding it is useful, exactly what would have happened on the day. The failure mode isn't the bad score, it's the response — abandoning mocks because they're demoralising, which restores the comfortable ignorance the mock interrupted.
If it's genuinely destabilising, separate the marking from the sitting by a day, and mark it as though it were someone else's script. And note the score somewhere you can compare it to the next one — the trend across two mocks is far more informative, and usually far more encouraging, than either number alone. Exam anxiety covers the case where the reaction is disproportionate to the result.