On this page
Review a CFA® mock by diagnosing each lost mark before doing another full paper. Separate a missing concept from a misread vignette, a calculation slip, a timing failure and an answer that did not follow the required format. Then assign a specific repair and test it on fresh practice. Repeating the same mock immediately can raise familiarity without improving the skill that the next exam will demand.
Preserve the conditions that make the score informative
CFA Institute says its mock exams aim to reflect the timing, topic weights, difficulty and question format of the registered exam level. That makes a mock useful as a rehearsal, but only if the candidate records how it was taken. Before opening explanations, note the date, whether the attempt was timed, which breaks were taken, any interruptions, and whether reference material was used. Record unfinished items separately from wrong answers. A score from an untimed, open-book attempt answers a different question from a timed simulation; both can be useful, but they should not be plotted as if they were identical measurements.
Do not reduce the review to one percentage. The same total can hide a concentrated curriculum weakness or a broad pace problem. A candidate who knows the concepts but leaves a cluster of questions unanswered needs a different repair from one who finishes comfortably but repeatedly chooses the wrong model. CFA Institute describes practice questions as a way to assess topic-level strengths and weaknesses, with explanations attached, on its mock and practice page. Use that feedback to diagnose the cause rather than to copy the correct option into a notebook.
Create a review record with five fields for each item worth revisiting: topic or learning outcome, what you chose or wrote, the reason you chose it, the correct principle, and the next test. The reason matters most. “I knew it after seeing the answer” is not a diagnosis. Write the first incorrect assumption or missing step. If you cannot explain why the chosen answer looked plausible, you may still be vulnerable to the same distractor in a fresh form. If you got the item right by guessing, include it; a correct guess is weak evidence of mastery.
Use a small set of error causes
Mark a knowledge gap when you could not retrieve the relevant rule, relationship or definition. Repair it by returning to the current curriculum, writing a short explanation in your own words, and solving a new item without notes. Mark an application error when you remembered the rule but selected the wrong inputs or applied it to the wrong setting. Repair that by identifying the condition that decides which method is appropriate. A candidate who can recite a formula yet misses the trigger for using it needs examples that contrast neighbouring cases, not another reading of the formula alone.
Mark a reading error when the necessary information was in the stem or vignette but you overlooked a qualifier, reversed a comparison or answered a different question. Repair it by practising a deliberate two-pass read: first identify the command and decision, then mark the facts needed to make that decision. This is particularly relevant when a long vignette supplies many facts that are true but not needed for the item. Mark an execution error when the model and inputs were sound but arithmetic, units, calculator keystrokes or transcription failed. Repair it with a repeatable calculation layout and a final reasonableness check, then test it under time pressure.
Mark a pacing error when you knew how to solve an item but spent so long on it that later work was rushed or blank. Repair it by setting a decision point at which you move on and return if time permits. A slow answer may reflect a hidden knowledge gap, so examine the work before calling it “just time”. Mark a format error when the response did not supply what the prompt requested. The error may be especially costly in a constructed-response setting: knowing the concept and producing an answer are different tasks. One item can have a primary and secondary cause, but choose the cause that changes the next practice session.
Work through an illustrative review
Suppose an invented candidate completes a 100-item practice set under self-imposed time conditions. The candidate records 68 correct answers, 20 wrong answers and 12 unanswered items. This is a made-up teaching example, not a CFA Institute mock structure or a prediction of an exam result. The raw 68% tells us less than the record of why 32 items were lost. After review, suppose 10 wrong answers show missing concepts, six show misapplied conditions, four show arithmetic slips, and the 12 blanks mostly follow two long detours. Those counts sum to the 20 wrong answers, while the blanks remain a separate pace signal.
The candidate’s next session should not consist of memorising 32 answer keys. The 10 knowledge items point to a small group of curriculum passages and fresh topic questions. The six application errors suggest side-by-side cases that differ in one decisive condition. The four arithmetic slips call for a compact calculation routine. The two detours need timed practice in moving on when a route is unclear. If the candidate merely redoes the same 100 items, the apparent score may rise because the answers are familiar. A stronger check uses unseen items and compares the rate of each error type under similar conditions.
This example also shows why a full mock is a scarce diagnostic resource. A review can be longer than the attempt because it must connect wrong answers to repairable causes. CFA Institute’s study guidance after a deferral recommends reflecting on how practice problems vary the application of a concept, paying attention to weak areas, and using mocks for readiness and pacing. The practical inference is to alternate full simulations with narrower practice that targets the problems uncovered. There is no magic number of mocks that substitutes for correcting a repeated misunderstanding.
Change the review for the exam level
At Level I, review whether the error came from a free-standing concept, a calculation, or a misread choice. The short item format makes careless selection tempting; record why the rejected options were wrong rather than simply memorising the selected letter. At Level II, ask whether a vignette fact was missed, taken from the wrong part of the case, or carried across items where it did not apply. The repair may be a better way of locating relevant facts quickly, alongside subject knowledge. In both levels, keep the timing record separate from conceptual accuracy so you know which intervention to test.
At Level III, constructed responses require another layer. CFA Institute states on its Level III exam guide that candidates must attend to bold command words and the number of answers requested; additional answers beyond the requested number are not evaluated. In review, mark whether you answered the actual command, used the relevant vignette facts, gave the required number of distinct points, and spent time proportionately. A response that explains a neighbouring idea at length may demonstrate study effort while leaving the requested point unstated. Rewrite a weak response in concise form, then test the same command with a fresh scenario.
Do not treat a constructed-response marking guide as a script to memorise. The aim is to recognise the kind of evidence the command requires. “Identify” needs a selection; “justify” needs a reason tied to the case; a calculation needs a defensible numerical result. If a mock’s marking scheme allows partial credit, inspect where the missing link occurred. If it does not show item-level marks, avoid inventing them. The candidate can still classify the response as complete, incomplete, off-target or blank, and practise making the necessary reasoning explicit.
Decide what to redo and what to drop
Redo a missed concept after a delay, using a fresh question or a newly constructed variation. Redo a calculation with changed inputs and a check of units. Redo a reading error under time pressure, aiming to locate the decisive qualifier. Redo a weak constructed response by answering the command in fewer, more relevant lines. If a topic keeps failing after one repair, move from question drilling back to the underlying curriculum and rebuild the explanation. The CFA programme page gives the wider course context; the methodology page describes how study notes connect to official learning outcomes.
Drop tasks that create little new evidence: copying an entire model answer, redoing a memorised mock for a flattering percentage, or rewriting a chapter that you can already apply on unfamiliar items. Drop a low-priority tangent when the main error came from a specific condition elsewhere, but keep a note if it recurs. Dropping a task is an editorial decision about study time, not a licence to ignore an examined area. When the next fresh set is complete, compare it with the earlier attempt only if timing, support and format were comparable. Look for fewer errors of the targeted type, fewer blanks and clearer answers. If the total score does not rise, the review still tells you whether the intended repair worked and what to change next.
A mock is therefore best treated as an experiment: record the conditions, diagnose the failure mode, make one targeted change, and check that change on unseen work. The useful output is a better next decision, not the comfort of recognising questions you have already seen.
Sources
- CFA Program Mock Exam and Practice Questions · cfainstitute.org · Retrieved
- Studying for the CFA Exam After a Deferral · cfainstitute.org · Retrieved
- CFA Program Level III Exam · cfainstitute.org · Retrieved



