Start with the exam you are actually taking

SOA actuarial exams are typically computer-based, proctored assessments with a large number of time-pressured objective questions, often multiple choice. Some exams also include written-answer or constructed-response components. The balance of topics, question types and requirements differs by exam. Before you build a plan, check the Society of Actuaries’ official website for the current syllabus, exam format, permitted tools and candidate instructions for your specific exam. Do not rely on a study schedule or calculator rule written for a different exam.

Across the exams, candidates need more than formula recall. Questions test core theory and accurate, efficient calculations under time pressure. A reliable approach is to identify what is given and what is required, select a suitable method, set up the calculation clearly, and check whether the result makes sense. Misreading a condition or making a small algebra or arithmetic error can undo otherwise sound understanding.

Make your study material usable

Start from the syllabus and your own course notes, textbook annotations, slides and worked examples. Bring them together by topic rather than keeping separate piles of material. For each topic, identify:

  • the definitions and assumptions that make a result valid;
  • the formulas you need, including what each symbol means;
  • the standard calculation patterns and when to use them;
  • common traps, such as confusing a probability with a rate or using the wrong time period;
  • at least one example you can redo without looking at the solution.

Build a revision board from those materials. Keep the board concise enough to review, but detailed enough to reconstruct a method. A formula without its assumptions or variable definitions is often not enough to answer a question correctly. MySummaries can organise material into a board by topic; a board on discrete random variables might look like this:

SOA Exams Discrete random variables: expectation and varianceStudy
Probability modelsDiscrete random variables: expectation and variance5 sections · 3 columns
Probability mass function
  • Support: list every possible value of X.
  • Total probability: sum p(x) over the support = 1.
  • Use P(X=x), not a density, for a discrete variable.
Worked distribution
x0100200
P(X=x)0.50.30.2
xP(X=x)03040
x²P(X=x)03,0008,000
Expected value

For a discrete X, E[X] = Σ x p(x). It is a probability-weighted average; it need not be one of the possible values.

Checks
  • Probabilities are non-negative and sum to 1.
  • E[X] = 70 lies between the smallest and largest possible values.
  • Variance cannot be negative; here it is 6,100 dollars squared.
Second moment and variance4 due
  • E[X²] = Σ x²p(x).
  • Var(X) = E[X²] − (E[X])².
  • Variance is in squared units; standard deviation is in the units of X.

Learn the minimum set of facts and methods

For each board, make a short “must not miss” checklist. This is not a replacement for learning the topic: it is a test of whether you can recall the conditions, setup and checks that prevent avoidable errors. Keep it specific. “Know variance” is too vague; “use E[X²] − (E[X])² and square the values before weighting” is actionable. MySummaries can pull this core checklist from a board; the key points for this example are:

Key pointsDiscrete random variables: expectation and variance
For a discrete X with pmf p(x), E[X] = Σ x p(x); first check Σp(x)=1.
Compute E[X²] as Σ x²p(x), squaring each support value before multiplying by its probability.
Use Var(X) = E[X²] − (E[X])²; variance has squared units and must be non-negative.
For P(X=0,100,200)=(0.5,0.3,0.2), E[X]=70, E[X²]=11,000 and Var(X)=6,100.

Practise a repeatable question method

Use a fixed sequence on practice problems, even when the question looks familiar:

  1. Read the task. Identify exactly what is requested: a probability, expectation, variance, present value, parameter, or explanation.
  2. List givens and constraints. Record units, time periods, independence assumptions and any boundary conditions. Do not infer a condition that the stem does not provide.
  3. Choose the method. Ask what principle fits the structure. Avoid starting a long calculation before deciding what quantity the question needs.
  4. Set up before computing. Define variables and write the formula with the given values in place. This makes a wrong setup easier to catch.
  5. Calculate and validate. Check signs, magnitude, units, bounds and whether the answer fits the situation. For a probability, for example, it must lie between 0 and 1.

For objective questions, also read every option carefully. Options can reflect a correct method with one changed sign, a missing square, or a mistaken interpretation. If your result is not listed, check the stem and setup before rounding or choosing the nearest option. For constructed-response work, show enough of the setup and assumptions for another person to follow your reasoning; consult the current official exam guidance for the requirements of your exam.

Review errors as data

Do not label every lost mark “careless.” Record what actually failed: reading, concept, method selection, notation, calculation or checking. Then write a brief correction that changes what you will do next time. For example: “I calculated E[X] correctly but subtracted 5,000 instead of 4,900. Recompute (E[X])² separately before subtracting.” Revisit that correction in a later session and solve a fresh problem using the same principle.

A marked written exercise should help distinguish a sound setup from an incomplete calculation. MySummaries can mark a practice response against your notes; this example shows a candidate who has the main method but slips on the final arithmetic:

Paper — Discrete random variables: expectation and variance06:35
78%Discrete random variables: expectation and variance — marked3/4 marks · 06:35 taken

A claim amount X has P(X=0)=0.5, P(X=100)=0.3 and P(X=200)=0.2. Calculate E[X] and Var(X).

3/4

E[X] = 0(0.5)+100(0.3)+200(0.2)=70. E[X²] = 100²(0.3)+200²(0.2)=11,000. Var(X)=11,000−70²=6,000.

The probability-weighted expectation and second moment are set up correctly. The final subtraction is not: 70² is 4,900, not 5,000.

Missed

Arithmetic accuracy in the final variance calculation; report the squared units as well.

Model answerGiven: P(X=0,100,200)=(0.5,0.3,0.2). E[X]=Σxp(x)=0+30+40=70. E[X²]=Σx²p(x)=0+3,000+8,000=11,000. Therefore Var(X)=E[X²]−(E[X])²=11,000−4,900=6,100. The variance is 6,100 dollars squared; the standard deviation, if required, is √6,100 dollars.

That is a MySummaries paper, filled with SOA Exams material. Yours is written from your own notes. Start free

Train for time pressure without practising mistakes

Timed practice is useful only when you can still learn from it. First practise a new method untimed until you can set it up correctly. Then use short timed sets and review every item, including correct guesses. As accuracy becomes steadier, practise longer mixed sets under conditions that match the current official guidance for your exam.

During a timed set, keep moving when a problem is consuming too much time. Mark it for return if the platform or practice system allows, and use the time remaining to secure questions you can solve reliably. Avoid spending several minutes refining an answer before you have checked that you are answering the question asked. Afterward, classify the error and redo the problem without looking at the solution. Timing targets should come from your exam’s current format and your own practice data, not a guessed universal pace.

Use a short audio review to consolidate a method

Audio can be useful when you are walking through a familiar topic, but it should not replace calculation practice. Listen for the assumptions and sequence, pause to predict the next step, then solve a related problem on paper or in your practice system. A good review explains why a formula applies and how to check the result, not just what to memorise. MySummaries can generate an examiner-voiced lecture from your board; a short review of the same distribution could sound like this:

Lecture — Discrete random variables: expectation and variance8 min
From a probability table to varianceWorks through a three-point claim distribution, from checking the pmf to validating the variance.
03:1808:02
Speed1×1.25×1.5×2×

Transcript · tap any word to jump there

Start with the support: zero, one hundred and two hundred. The probabilities are one half, three tenths and one fifth; they sum to one, so the probability mass function is valid. Now calculate the expectation by multiplying each value by its probability. That gives zero, thirty and forty, for an expected claim amount of seventy dollars.

For the variance, do not square the expectation too early and do not forget to square each value in the second moment. The terms are zero, three thousand and eight thousand, so E of X squared is eleven thousand dollars squared. Subtract seventy squared, which is four thousand nine hundred. The variance is therefore six thousand one hundred dollars squared.

Finish with a check. The mean lies within the support, and the variance is positive. Keep the units straight: the variance is in squared dollars, while the standard deviation would be in dollars. If the question asks only for the variance, do not substitute the standard deviation for it.

Put the work into a weekly cycle

A practical study week alternates learning, retrieval and timed application. Set the total hours according to your available time and the official exam date; do not use a generic number as a guarantee of readiness. For each topic, repeat this cycle:

  • Learn: read your source material and build or update a concise board.
  • Retrieve: explain definitions and assumptions from memory; answer due flashcards.
  • Apply: solve fresh problems, first untimed and then under time limits.
  • Review: analyse errors, correct the board and schedule another attempt.

Keep a simple record of topic, question type, time spent, result and error category. Use it to choose the next session. If accuracy is poor because the concept is unclear, return to the notes and rebuild the method. If the setup is sound but arithmetic fails, practise calculation and independent checks. If the main issue is misreading, slow down at the stem and underline the requested quantity and constraints in your practice work.

In the final stretch, prioritise weak syllabus areas and mixed practice while preserving time to sleep and review. Do not spend the last sessions collecting new materials instead of using them. Confirm logistics and permitted tools through the official SOA instructions for your specific exam. On practice questions, keep using the same sequence: frame the problem, select the method, set it up, calculate, and check.

How MySummaries helps

MySummaries turns your own actuarial notes into topic boards, flashcards for spaced review, written practice marked against those notes, and examiner-voiced audio lectures. Use the board to organise assumptions and methods, then use missed marks to decide what to practise next.