Volume climax (single-day extreme volume + down day in drawdown)

Thesis (rejected): A single bar with volume >= 3-5x the 20-day average closing meaningfully negative is the textbook "capitulation" event. In a name already in drawdown, this should mark the exhaustion low.

Code: research/volume_climax.py Report: research/volume_climax_report.json (n=20,767)

Why it failed: Edges hover around zero across every variant — including the strict "vol ≥ 5x + dd ≤ -30 % + OOS 2020+" subset (n=1,081) which is slightly negative at 5-20 d.

The "capitulation volume" intuition does NOT hold at the daily-bar resolution. A single high-volume down day is far more often the middle of a slide (institutions distributing) than the bottom. Volume flow over many bars (OBV divergence) IS predictive — but single-day spikes are not.

Verdict: Skip. The signal needs context — pairing the volume spike with a gap (gap_capitulation) or a pivot reclaim (failed_breakdown) or a divergence (OBV) is what works, not the spike by itself.