factual
Overlapping return windows shrink the true sample to a few
Correlating valuations with overlapping multi-year forward returns, such as monthly windows starting in consecutive months, does not create independent observations, so the true sample is only four or five independent segments, which is too small to support strong conclusions and invites overfitting.
factualpending
Speaker
JesseEvidence Quote
“those are not different samples, those are not independent of each other, because they're overlapping.”
Created: 11/11/2026, 6:43:50 PM
Your notes
Loading notes...