ACT 490 · Year 4 · Semester 1 · 3 credits · Integration & Practice

Capstone I: Data-Driven Pricing & Reserving

Bridge classical actuarial reserving and pricing with modern machine learning pipelines to deliver defensible, ASOP-compliant decisions on real-world industry data.

Full practice exam · 30 Q · 120 min

Season 1 · 8 episodes

The Solvency Threshold: Horizon Mutual's Capstone Defense

When state regulators threaten receivership over a thirty-million-dollar commercial auto deficit, a newly appointed lead actuary must construct a fully reproducible, ASOP-compliant reserving and pricing pipeline combining classical loss development with neural networks to survive the ultimate external audit panel.

Protagonist · Maya Lin, newly appointed Lead Reserving and Pricing Actuary at Horizon Mutual

Setting · Horizon Mutual's corporate headquarters in Chicago and the State Insurance Commissioner's hearing chambers

Stakes · Horizon Mutual faces mandatory regulatory supervision and liquidation if Maya cannot prove the solvency of its reserves and justify a forty percent commercial auto rate overhaul before a hostile panel of appointed chief actuaries.

Pick an episode

Episodes

Season finale: Maya delivers a unified, reproducible waterfall and bootstrap reserve range to the external practitioner panel, demonstrating that machine learning micro-reserving and classical Bornhuetter-Ferguson reconcile perfectly, securing regulatory clearance and saving the carrier from insolvency.

From the lecturer: Welcome to ACT 490, your senior capstone. Over the next fourteen weeks, you transition from student to credentialed practitioner. You will work in small actuarial consultancy teams analyzing real market data from the CAS Loss Reserve Database. Do not watch these lecture films passively from behind a keyboard. Keep an engineering notebook and an active Git repository alongside every session. Before running any script, sketch out triangle flows, write down loss development factor equations, and define your parameter assumptions explicitly on paper. Although this course does not target a single exam, it synthesizes the core technical syllabi of CAS Exams 5, 7, and 8, alongside SOA General Insurance exams and Actuarial Standards of Practice (ASOP Nos. 12, 20, 23, 41, and 43). In practice, algorithms do not sign actuarial opinions—you do. Treat every data reconciliation, loss distribution assumption, and uncertainty interval as if you were defending it before an insurance commissioner or corporate audit committee. Let us build rigorous, reproducible, and explainable models.

Cold open

At six fifteen on a Monday morning, Maya Lin finds an official notice from the State Department of Insurance taped to the door of Horizon Mutual's actuarial suite. Horizon's commercial auto surplus has eroded by twenty-eight million dollars over four quarters, and the regulator demands a completely audited, ASOP-compliant reserve reconstruction and rate indication within eight weeks, or Horizon will be placed into administrative receivership.

preparing lecture…

Transcript

Define the scope of an integrated reserving and pricing capstone project under ASOP No. 41, 43, and 56
Formulate the mathematical boundary conditions linking unpaid claim liabilities to future premium rate indications
Establish a reproducible data engineering and modeling lifecycle integrating classical actuarial methods with machine learning pipelines
Draft an actuarial project charter with defined data validation checks, hindsight testing criteria, and model risk governance