IFoA CB2 study plan

Map the CB2 syllabus into a one-page topic checklist covering data, modelling, and actuarial applications, and rank each topic by perceived difficulty

Set up a working R environment with the packages you will need for data manipulation, visualisation, and basic modelling, and confirm each loads without error

Work through core R data structures (vectors, factors, data frames, lists) and write at least 15 short snippets demonstrating indexing, subsetting, and coercion

Practise importing and tidying a messy CSV: handle missing values, fix variable types, and reshape one wide table to long format

Implement grouped summaries and aggregations on a realistic dataset using base R and a tidyverse approach, and compare the outputs

Build at least four static plots (histogram, boxplot, scatter, line) and refine axis labels, legends, and titles to a presentable standard

Write custom functions with sensible argument defaults and apply them across a list of data frames

Debug three deliberately broken scripts by reading error messages and tracing the fault to its source line

Derive the likelihood for a simple frequency model and code a numerical maximum likelihood estimation routine from scratch

Fit and interpret a generalised linear model for claim counts, checking the dispersion assumption and explaining coefficient meaning in actuarial terms

Simulate from a fitted model and compare simulated summary statistics against the observed data

Explain the distinction between model fitting and model validation, and list the diagnostics you would run for each

Complete two full past-paper-style questions under timed conditions using only the permitted reference materials

Mark your own answers against the model solutions and classify every lost mark as knowledge gap, method error, or presentation error

Rewrite the weakest answer so it would earn full marks, then re-attempt the same question cold after 48 hours

Build a personal error log with one line per mistake and the corrective action

Summarise the assumptions and limitations of the main survival and lifetime distribution models used in the syllabus, in your own words

Fit parametric survival models to a dataset and compare them using information criteria and residual checks

Code a Kaplan-Meier estimator and a Cox proportional hazards fit, then interpret the hazard ratios in plain language

Explain how censoring changes the likelihood and demonstrate the effect on a small worked example

Implement a simple claims reserving calculation (for example chain ladder) from first principles and reconcile it against a package output

Explain the assumptions behind the reserving method and identify which diagnostics would reveal a breach

Work through a credibility or experience-rating calculation by hand and then verify it in code

Describe the regulatory and professional context in which reserving results are reported, without overstating what you know

Sit a full timed mock paper under exam conditions, including the data analysis component, and record time spent per question

Identify the two topics where your accuracy is lowest and produce condensed revision sheets for each

Rehearse explaining your code and modelling choices aloud in under three minutes per question

Review the error log from week 4 and confirm each listed mistake is now resolved

Re-attempt every question missed across the previous mocks and score above your earlier mark on each

Practise the exam-day workflow: reading the data analysis brief, planning the script, and leaving time to write up results

Condense the whole syllabus into a final two-page formula and method sheet you can recall from memory

Confirm your exam logistics, permitted materials, and software conventions so nothing is a surprise on the day