52-week-high breakout

Thesis (rejected): Stocks closing at a new 52-week high on volume

1.5× average should continue higher for 20-60 days. George & Hwang 2004 anchor effect.

Code: research/high_52w_breakout_validate.py Report: research/high_52w_breakout_report.json (n=18,183)

Why it failed: Edge is consistently negative across every horizon tested in our modern-era (2010+) sample.

George/Hwang's finding doesn't replicate cleanly in modern data with our filter. Possible explanations: (a) academic version uses anchor distance, not a strict breakout event, (b) post-2010 the signal has been arbitraged by trend-following quants, (c) requiring volume confirmation paradoxically selects names whose move is already crowded.

Verdict: Cohort skipped. The breakout-on-volume signal is in the same family as our coil-breakout cohort (which works) but lacks the prior- thrust + tight-coil structure that makes coil's right tail tradeable.