CAS PCPA Guide

PCPA is the CAS predictive-analytics requirement for ACAS. It uses a two-part structure: a computer-scored exam first, then a project window where you apply predictive analytics to a business problem and submit a technical response.

The current CAS PCPA content outline describes a two-hour computer-based exam with 40 multiple-choice questions. The official PCPA page and FAQ then add the practical structure around it: once you pass the exam, you can register for the project component, which runs in four administration windows each year.

The CAS PCPA page also notes that exam registration is continual through Pearson VUE, that there is a mandatory two-week waiting period between exam attempts, and that candidates should pass the exam at least four weeks before a project window to allow time for official results and project registration.

The content outline frames PCPA as a property-casualty predictive analytics requirement, not just a statistics quiz. The exam checks foundational predictive-modeling knowledge, while the project asks you to work from a business problem, data, scope parameters, and submission guidelines.

The project format details emphasize exploratory data analysis, model construction and refinement, interpretation of results, and a technical summary in response to a business question. It assumes prior knowledge from MAS-I, MAS-II, Exam 5, and the practical coding skills needed to execute modeling work.

PCPA is not the beginning of the CAS statistics sequence. It sits after the candidate has already built the MAS-I and MAS-II base and moved through Exam 5. That is why the content outline explicitly treats predictive-analytics foundations as assumed rather than newly introduced.

In other words, MAS-I and MAS-II build much of the statistical language. PCPA then asks whether you can actually use that language in a P&C predictive-analytics setting with project discipline and communication.

PA and PCPA both care about data exploration, model choice, interpretation, and communication. But they test those ideas differently. PA is an SOA exam built around a business-facing analytical writeup. PCPA is a CAS requirement with both a separate exam and a hands-on project component tied to property-casualty predictive analytics.

That makes PCPA feel more like a pathway requirement built around application, while PA feels more like a timed actuarial analytics exam with communication as a first-class skill.

The strongest PCPA preparation comes from combining statistical fluency with project discipline. Candidates need to be comfortable with model comparison, validation logic, feature or data issues, and writing a concise technical explanation that stays tied to the business problem.

If you are preparing from the SOA side for comparison purposes, the best way to understand PCPA is to see it as closer to a predictive-analytics workflow assessment than to a classic multiple-choice actuarial exam.