The strategy is a cross sectional mean reversion equity stat arb that uses PCA to isolate idiosyncratic mispricings. Each month it forms a universe of the 20 most liquid US equities by dollar volume with a price filter above 5, then uses the most recent 60 daily closes to model common risk factors. It log transforms and de means prices, extracts the first three principal components as statistical factors, and for each stock fits an OLS regression on these factors to estimate a factor explained fair value. The alpha signal is the latest standardized regression residual (z score): stocks with residual z scores below -1.5 are treated as abnormally cheap versus the factor model and are expected to mean revert. At each monthly rebalance (decision made pre market), the portfolio takes short positions in the selected negative residual names with weights proportional to the magnitude of their negative z scores, normalized so the selected basket sums to full gross exposure (subject to maintaining a 5 percent free cash buffer). There are no explicit long legs, stop losses, or time based exits beyond the next scheduled rebalance; positions are refreshed monthly using updated universe membership and factor fits.