Min Max Uniforms Correlated

This question considers the minimum and maximum of two independent uniform variables on an interval. The candidate is asked to reason about the joint behavior of these order statistics, and then to condition on one event involving the maximum while constraining the minimum. The first part focuses on computing a conditional probability involving inequalities on both random variables; the second part requires obtaining the correlation between the minimum and maximum, capturing how they co-move despite coming from independent inputs. This type of setup is common in quantitative research and trading interviews when testing intuition about transformations of basic distributions and dependence induced by sorting or ranking.

Solving it leans heavily on geometric reasoning in the unit square, interpreting joint events as regions and turning probability questions into area computations. From there, the candidate must translate the conditional probability into a ratio of such areas or integrals, and then derive expectations, variances, and covariance of the transformed variables. The interviewer is watching for comfort with joint densities, conditioning, order statistics, and the mechanics of computing correlation, as well as the ability to organize a multi-step calculation cleanly.

What it tests

When analyzing functions of independent random variables, especially order statistics like the minimum and maximum, the joint distribution is often best understood geometrically. For two independent uniform variables, their joint distribution is uniform over the unit square, and events involving their minimum and maximum correspond to simple geometric regions (rectangles, triangles, or squares) within this square. The probability of such events is proportional to the area of the corresponding region. This geometric approach generalizes: for any pair of independent continuous random variables, the joint probability of events defined by inequalities can be visualized and computed as the measure (area, volume, etc.) of the region defined by those inequalities. The underlying reason is that independence makes the joint density a product, so integration over a region is just the region's measure times the density.