Market-Neutral Pairs
Statistical-arbitrage research: co-moving pairs, spread z-scores, and the honest out-of-sample test.
What this is: a research view of co-moving stock pairs. We long the cheap leg and short the rich one so net market exposure is ~0 — the spread, not the market, is the bet. What this isn't: a trade signal. The out-of-sample column shows that selecting pairs by recent backtest return has no reliable forward edgeon this universe — so nothing here trades. It's the validation that tells us what to fix before any market-neutral strategy goes live.