Jersey Funds with 97.5% Certainty

This quant interview question is about modeling the total cost of multiple independent, continuously distributed risks and linking that to a required confidence level. It uses normal distributions and inverse cumulative probabilities to turn an informal statement about being "sure enough" into a precise numerical requirement. In quant prep terms, it sits squarely at the intersection of continuous random variables and risk aggregation.

It trains your understanding of how independent normal variables combine, and how their parameters govern the distribution of a sum. It also reinforces comfort with z-scores, quantiles, and tail probabilities, and with moving fluently between real-world wording and probability notation. You practice turning a verbal confidence target into a concrete capital level.

This matters for quant interviews because portfolio-level questions, PnL aggregation, and capital adequacy all rely on the same ideas. Interviewers use such problems to test whether you can quantify uncertainty rigorously, not just compute isolated expectations. For trading and risk roles, being able to translate narrative risk constraints into distributional statements and then into hard numbers is a core quantitative finance skill, and mastering these continuous-variable questions is essential quant prep.

What it tests

Whenever you sum independent normal random variables, the result is itself normally distributed, with a mean equal to the sum of the means and a variance equal to the sum of the variances. This property arises because the normal distribution is closed under addition: the convolution of two normal densities yields another normal density. This closure allows you to replace a problem involving several uncertain quantities with a single, aggregate uncertainty, making it possible to use standard normal tables or inverse CDFs directly. The key is that independence ensures variances add, not standard deviations, and that the resulting distribution's shape remains normal, so all familiar probability calculations apply. This principle is powerful because it reduces a potentially complex joint probability question to a one-dimensional calculation.