Skip to main content

TL;DR

Simpson's Paradox: When Aggregation Reverses Trends: A canonical quantitative trading interview question at intermediate difficulty. Commonly asked at Two Sigma, Citadel, DE Shaw, Point72, WorldQuant.

By Valenke Exam Prep Team·Last updated 2026-06-01
intermediateStatistical Inference & Estimation

Simpson's Paradox: When Aggregation Reverses Trends

Asked at: Two Sigma, Citadel, DE Shaw, Point72, WorldQuant

Problem
Two trading strategies are backtested over two market regimes (bull and bear): | | Strategy A Win Rate | Strategy B Win Rate | |---|---|---| | Bull market (800 trades for A, 200 for B) | 60% | 70% | | Bear market (200 trades for A, 800 for B) | 30% | 40% | | Overall | ? | ? | (a) Compute the overall win rate for each strategy. (b) Which strategy has the higher overall win rate? Does this contradict the per-regime results? (c) Explain the paradox. Which strategy is truly better? (d) A PM asks you to deploy Strategy B because "it wins more often." What do you say?

Ready to practice for the Valenke Finance Exam?

Adaptive practice powered by Item Response Theory targets your weak areas. Start with 3 free sessions.

Start free practice →