Overview

Metrics for the Individual × Sparse cell — event signals where most (date, asset_id) cells carry factor == 0 and the few non-zero cells mark events. The schema is {0, R} — zero on non-event entries, any real value otherwise (R is unrestricted — positive, negative, or any magnitude). Common forms: {0, 1} for a pure event flag, or {0, R} for any real-valued magnitude; both flow through (see caar for the input-form table).

Metric	Role	Page
Cumulative average abnormal return — \(t\)-test or BMP \(z\)-test	Primary	caar
Skewness / hit-rate / win-loss diagnostics on per-event returns	Profile	event_quality
MFE / MAE order-statistic excursion within an event window	Profile	mfe_mae
Event-window horizon decay	Profile	event_horizon
Herfindahl-Hirschman index (HHI) on event dates — flags clustering that violates BMP/CAAR independence	Profile	clustering
Non-parametric Corrado rank test — robust to non-Gaussian returns	Profile	corrado

The Common × Sparse cell shares this column contract but routes through a distinct procedure (_CommonSparsePanelProcedure — per-asset ordinary least squares (OLS) \(\beta\) on the broadcast dummy → cross-asset \(t\)). Its metrics live under Common continuous.

When events cluster on the same date (earnings season, macro release), prefer caar.bmp_test(kolari_pynnonen_adjust=True) over the vanilla \(t\)-test — the clustering HHI tells you when this matters.