Candidates assess themselves constantly and almost always with the wrong instrument. "I have done Civil Law twice" is not a measurement. Neither is "I feel behind" — although, to be fair, that one is at least honest.
Readiness is measurable, in the same way that any performance is: run the task, score the output, compare over time.
The four checks
Each takes an afternoon at most, and together they fit into one weekend a month.
1. The closed-book rebuild
Take one subject. Blank paper, no notes, ninety minutes. Rebuild the subject's outline from memory — headings, rules, elements — then mark your own rebuild against your outline and count what is missing.
What it measures: structure and recall, at the level the examination assumes. What a good result looks like: the shape is right and the misses are specific topics, not whole sections. What a bad result tells you: whole missing sections mean the subject was read, not learned, and the repair is retrieval rather than more reading.
Score it as a simple fraction: topics present / topics in the syllabus. Track the number monthly.
2. The timed, marked answer
Take five questions from one subject — mix the shapes: a distinction, a requisites question, a multi-part hypothetical, a status question, a remedy question. Twelve minutes each, no notes. Then mark them yourself against the rubric's three questions:
- Did I state the conclusion first?
- Did I state the rule with its elements, precisely?
- Did I apply it to these specific facts, and conclude?
What it measures: the conversion of knowledge into a scorable answer — the skill described in how a bar answer is scored. What a good result looks like: five answers where you can tick all three boxes, with the rule correct. What a bad result tells you: you know more than you can write. That is a fixable, and completely different, problem from not knowing enough.
Track two numbers: answers with all three elements out of five, and rules stated correctly out of five.
3. Archetype coverage
List the question shapes you have practised for a subject (the past-question method builds this list) and mark which ones you have answered under time. A subject where you have written ten distinction answers and no multi-part hypothetical is a subject with a blind spot the paper will find.
What it measures: whether your practice matches the paper's actual variety.
4. Delayed retention
This is the check candidates never run, and it is the one that predicts September. Pick ten doctrines from a subject you finished three weeks ago. Closed book. State the rule and one application for each.
What it measures: whether your schedule is spaced, not merely sequential — the failure described in interleaving versus blocking. What a good result looks like: seven of ten, with the misses being detail rather than whole rules. What a bad result tells you: the retrieval cadence is too thin, and the fix is twenty minutes a day, not more reading.
A scorecard you can keep on one page
| Month | Rebuild (topics present / total) | Timed answers (3/3 elements) | Archetypes practised | Delayed retention (out of 10) |
|---|---|---|---|---|
| Nov | ||||
| Dec | ||||
| Jan |
Three rules for using it:
- Compare months, not candidates. Someone else's numbers are not your baseline.
- Record the misses, not only the score. The list of what failed is the work plan for the next month.
- Do not convert the score into a prediction. Nobody can tell you from a mock whether you will pass. You are tracking the direction of a curve, not foretelling a result.
The proxies that feel like measurement but are not
- Hours. The most popular and least informative. An hour of retrieval and an hour of highlighting are both an hour.
- Pages. See above, with more paper.
- "Covered" a subject. Coverage is a claim about input. The examination tests output.
- Mock scores, taken once. A single score is noise: the questions were a sample, the marking was you, and the day varied. The third data point is where a trend begins.
- Confidence at the end of a session. Fluency during study is the most reliable sign that you have not yet been tested; the effortful session is the one that taught you something.
What to do with a bad month
The temptation is to add hours. Usually the repair is different, not larger:
| Symptom | Likely cause | Repair |
|---|---|---|
| Rebuild misses whole sections | Subject read but never retrieved | Rebuild weekly, not monthly; stop adding new material until the map exists |
| Rebuild fine, answers poor | No practice at writing | Two timed answers a week; use the twelve-minute shape |
| Answers fine, delayed retention poor | Schedule is sequential, not spaced | Daily retrieval block; interleave subjects |
| Everything slips at once | Overloaded plan or burnout | Cut coverage of the lightest subjects, protect retrieval and sleep — see handling review burnout |
The one number that matters most
If you track nothing else, track the timed, marked answer count for the year. Twenty written-and-marked answers before September puts you in a different category from a candidate who wrote none, regardless of how much either of you read. Writing is the only activity that rehearses the actual task, and the count is the only honest measure of whether you rehearsed it.
Everything else in the review year is preparation for the moment when a question you did not choose appears on a screen and you have twelve minutes. Measure that, and the rest follows.