A gap based only on autocorrelation
Question
A researcher chooses a train-test gap after seeing where sample autocorrelation first becomes insignificant. Why is this not a complete leakage policy?
Short answer
A noisy significance cutoff does not establish that information leakage is absent. Start from feature availability and the exact label intervals; those give deterministic exclusions. Then assess residual dependence and sensitivity to additional gaps using development data. A data-dependent gap can itself be selected to favour performance, so freeze and evaluate the policy as part of the complete modelling procedure.
Original question and answer written by SelfMadeQuant.com.
See the hint and the full worked solution in the question bank (free account).
Open the worked solutionLessons behind this question
Start with a free lesson: Expectation, Variance and Covariance is part of a free topic. Create a free account, no card needed, to read it in full with worked solutions.
Read it free