Data and research interviews.
The trend that reverses
The question
A strategy outperforms in every market regime measured separately but underperforms overall. How?
Have a go before you open the solution. If you are stuck after a few minutes, the first paragraph below is usually enough of a nudge.
Show the worked solution
The answer
Simpson's paradox: the weighting across regimes differs from the within-regime comparison.
If the strategy is mostly active during regimes with poor returns generally, its aggregate result is dragged down even while it beats the benchmark inside each regime.
Aggregation collapses the group sizes, and when group membership correlates with both the treatment and the outcome, the pooled comparison can reverse.
The fix is to decide which question you are answering. For allocation you care about the aggregate; for signal quality you care about within-regime performance.
Naming the confounder explicitly is what separates a good answer from 'the numbers are weird'.
What it is testing
Whether you look for the confounding variable rather than distrusting arithmetic.
Try these next
The third correlation
StandardA and B are 0.9 correlated, B and C are 0.9 correlated. What range is possible for the correlation of A and C?
Why the top performers disappoint
Warm-upA fund's best traders this year are usually worse next year. Does that mean the firm is failing to develop them?
How long to trust a Sharpe ratio
HardA strategy shows a Sharpe of 2 over six months. How much should that convince you?
The backtest that will not survive
StandardYou tried two hundred signal variants and kept the best. Its backtest Sharpe is 2.5. What do you actually believe?
Preparing properly
Live opportunities
Interviewing soon? See who is hiring
Open vacancies we track at quant firms, grouped by role and city. Adverts are anonymised.
Hiring for these roles? Brief the desk on a hire or check your fit for a role.
Research desk
Speak to a quant market specialist
Ask what a firm really pays, how a desk hires, or where your profile sits. One reply from the research desk, in confidence, no obligation.