A student team from The Ohio State University, named Nova Analytics, has advanced to the semifinal round of the 2026 Society of Actuaries (Society of Actuaries) Student Research Case Study Challenge. The four-member team brings together a mix of mathematical backgrounds, consisting of undergraduate students Yaqi Zhang (Statistics), Xiangying Li (Actuarial Science), Shiyan Gao (Actuarial Science), and Sili Wei (Mathematics). The group conducted their research under the guidance of advisor Dr. Linfeng Zhang and co-advisor Dr. Kenneth Ng.
The Interplanetary Insurance Project
The 2026 SOA case challenge required participating student teams to step into a fictional scenario: designing a comprehensive insurance product for the Cosmic Quarry Mining Corporation, a mining entity operating across three distinct planetary systems. To address the unique operational environments and systemic risks of space-based mining, Nova Analytics proposed a diversified insurance portfolio. Instead of a one-size-fits-all plan, their framework spanned four distinct hazard lines: equipment failure, cargo loss, workers' compensation, and business interruption. Each line was mathematically tailored to account for localized infrastructure constraints, environmental hazards, and potential operational pauses unique to each planetary system.
The Team's Risk Framework
The mathematical framework required to price these policies had to move past simple linear assumptions and independent risk models.
First, the team utilized Random Forest models—a machine learning technique—to better predict how often claims would occur and how severe they would be. Rather than assuming risk changes at a constant, linear rate, these models allowed the team to capture complex, non-linear relationships in the data. Once these baseline risks were established, they ran Monte Carlo simulations, generating thousands of possible future outcomes to map out the overall probability distribution of the company's aggregate financial losses.
A major challenge in risk management is accounting for domino effects, where one disaster triggers another. Standard models often treat different types of insurance claims as independent events, but in the real world, a major environmental disruption might simultaneously cause equipment failure, cargo loss, and a halt in business operations. To fix this, Nova Analytics applied a statistical technique called the Iman-Conover method. This allowed them to impose a realistic dependency structure across their data, ensuring that the model accurately simulated correlated extreme events across different planetary systems.
Finally, the team looked closely at the "tail exposure"—the worst-case scenarios sitting at the extreme edges of their loss distributions. Using Value-at-Risk (VaR) and Tail Value-at-Risk (TVaR) metrics, they calculated exactly how much backup capital the corporation would need to hold in reserve to survive severe financial shocks. To ensure the entire framework was robust, they stress-tested their equations against shifting macroeconomic trends, verifying that their pricing remained stable even when face-to-face with unexpected inflation or rapid operational growth.