Qm
Statistics

The strategy that wins in every regime but loses overall

confoundingsimpsons-paradoxcausal-inference

A desk compares a New strategy (N) against the Old one (O), counting a day as a win if it ends profitable. They split days into calm and volatile markets.

  • Calm days: N wins 95 of 100 (95%); O wins 180 of 200 (90%).
  • Volatile days: N wins 130 of 200 (65%); O wins 55 of 100 (55%).

Pooling everything, N wins 225 of 300 (75%) while O wins 235 of 300 (about 78%), so O looks better overall.

Which strategy is actually better, and why does the pooled table point the other way?

Your answer

Solving needs a free account

Answers, streaks and solutions unlock when you are signed in. Reading the question and the hint stays free.

Create a free accountSign in

Discussion

Sign in to join the discussion · reading is open to everyone

💡 Discussion rules

  1. No full solutions here. Hints and approaches only.
  2. Complexity, edge cases and intuition are the point.
  3. Interview experiences are welcome. Respect your NDAs.

Loading discussion…

Learn the concepts

The theory behind this question.

Bad Controls and Post-Treatment BiasConfounding and the Backdoor CriterionConfounding in Financial DataDirected Acyclic Graphs for CausalityDouble Machine Learning

Related questions

The batter with the higher average each year, but not overallWhen disaggregating is the wrong moveThe ad channel that converts better on every deviceThe school that passes more students in every subjectDid the university admit men at a higher rate?Simpson's paradox: which ad creative truly converts better?
All questions →