How to review an Exam P practice exam

Keep the evidence before opening the key

Save the answer you chose and the calculation that led to it. Mark whether you could explain the model, narrowed the choices, or guessed. Record the time if you tracked it during the attempt; leave it blank if you did not. Reconstructing a fluent solution after reading the key can hide the exact step you needed to learn.

After checking the key, flag three kinds of answer: wrong, correct with uncertain reasoning, and correct but slow. For each flagged question, compare your work with the solution from the top. Stop at the first line where your reasoning diverges. A later arithmetic error may be a consequence of an earlier incorrect model.

A reusable review log

Copy these columns into a notebook or spreadsheet. The entries are hypothetical diagnoses of the free counting sample, not measured student results. Replace them with your own work.

Worked diagnosis: why “halve it again” fails

Attempt the free question 1 and inspect its full solution before continuing; this section reveals the answer. Its two restrictions concern precedence between mortality and morbidity reviews and nonadjacency of lapse and expense reviews.

There are five distinct reviews. Swapping the mortality and morbidity labels pairs every unrestricted schedule with one having the opposite precedence, so exactly half of the 120 arrangements have mortality first. That gives 60. The adjacency restriction needs its own count: there is no matching symmetry that says it eliminates exactly half of those 60.

To count what must be removed, combine lapse and expense into one block. The block and the other three reviews form four objects. The block can have either internal order. Swapping mortality and morbidity still divides these arrangements equally by precedence.

The incorrect choices identify different decisions to inspect. Use this table to locate a possible error, then check whether that error actually appears in your work. A selected distractor alone does not prove how you reasoned.

Check the model with a different count

The six nonadjacent unordered slot pairs are (1,3), (1,4), (1,5), (2,4), (2,5) and (3,5). Each pair allows two orders of lapse and expense. The remaining three reviews have 3!/2 = 3 orders with mortality before morbidity. Thus 6 × 2 × 3 = 36. This check starts from allowed slot pairs rather than forbidden blocks.

A useful correction in the log is: “Count adjacency with a two-order block, subtract from the precedence-valid set, and check using nonadjacent slot pairs.” Writing “revise permutations” gives you much less to act on.

Choose practice that addresses the first error

• Model or condition error: restate the requested event and its support before calculating. For a conditional probability, name the conditioning event and its denominator. For independent variables, identify where the independence assumption enters.
• Algebra or arithmetic error: keep the valid setup and redo the faulty line. Preserve an exact fraction until the last step, and check whether the result lies in the quantity’s possible range.
• Correct guess: close the solution and derive the answer. If you cannot explain why the other choices fail, keep the question flagged.
• Slow correct answer: identify the step consuming time. Practise that step on a small set before spending another three hours on a full paper.

Retest the repaired skill on unfamiliar wording or parameters. Record the new setup and result. Returning to the same question can check recall of a method, but it cannot show by itself that you can choose that method on a new problem.

For a specific topic gap, start with the free Exam P topic library. Use the free formula sheet when retrieving a formula is the obstacle, or the study guide when several prerequisite topics need work. The practice-exam and workbook comparison separates mixed timed papers from focused topic practice.

Use six minutes as an average, then inspect the pattern

The November 2026 SOA syllabus specifies 30 questions in three hours. Dividing 180 by 30 gives six minutes per question on average. That includes time to revisit marked items; it is not a requirement to spend six minutes on each problem. A short counting question can leave time for a longer calculation.

Look across your log. Several slow questions in one topic suggest a method you need to practise. Delays scattered across otherwise sound solutions may call for work on calculation fluency or deciding when to move on. In your next practice sitting, mark a problem when progress stops and return after attempting the others. Review whether that decision left enough time for the final questions.

For the actual exam, the syllabus says unanswered questions are scored incorrect, so enter an answer for every question. It also describes unscored pilot questions and statistical scaling across forms. A fixed raw score on an independent practice PDF is therefore not a direct conversion to an SOA exam result. Track errors, confidence and timing across unfamiliar sets rather than treating a single practice percentage as a pass guarantee.

Use a free sample before choosing another full paper

Try the complete question 1 sample with its worked solution and wrong-choice analysis. SOA also provides a free online sample exam through its official study page.

If you want an additional printable paper, Probability Practice Examination I contains 30 original questions, a 180-minute format and full solutions in a 77-page English PDF. Version 1.0.1 costs $39 USD plus applicable tax as a one-time purchase. Read its three actual page previews and delivery terms before deciding whether it fits your practice plan.

Use the next sitting to test the changes written in your log. A helpful review leaves you with a concrete action and a way to see whether it worked.