Beat-persistence pre-earnings drift

Thesis (real but too small to productize): When a company has beaten EPS consensus on each of the last 3 quarters, the stock drifts up in the 3-10 trading days before the next print as positioning builds.

Code: research/beat_persistence_validate.py Report: research/beat_persistence_report.json (n=7,660 events)

Why we're parking it: Signal exists but the magnitude is too small to support a defined-risk options structure or a meaningful stock-only position.

Even the tightest bucket (n=506) shows only ~2 % edge across a ~6-day hold. After options friction and entry slippage that probably collapses to ~0.5-1 %. Bounce, coil and PEAD all show 25 %+ options edge on larger samples — beat-persistence isn't competitive.

Verdict: Not added to the watch-list. The fact that beats persist before the next print isn't useful as a standalone trade because analyst-revision activity is largely priced before our signal fires. Keep the validator in the tree for future re-litigation if better analyst-revision or whisper-number data becomes available.