The thesis
If two stocks are exposed to the same underlying business drivers — same sector, similar market cap, comparable margin profile — their price ratio should be relatively stable over time. When the ratio diverges meaningfully without a fundamental cause, the spread tends to mean-revert. Pairs trading is the systematic exploitation of that mean reversion: long the underperformer, short the outperformer, hold until the spread re-converges.
Academic basis
Gatev, Goetzmann, and Rouwenhorst (2006) in Review of Financial Studies documented that a simple distance-based pairs trading strategy on US equities produced positive risk-adjusted returns over 1962–2002. Subsequent work (Do and Faff 2010) found the unadjusted profitability declined post-2002 as the strategy became more crowded, but cointegration-based variants and sector-constrained variants continued to perform after costs.
How Alpha Suite implements it
- Sector pre-filter — only same-sector pairs are considered, eliminating most spurious correlations from cross-sector pair selection.
- Distance ranking — within a sector, pairs are ranked by historical price-ratio variance; the lowest-variance pairs are candidates.
- Cointegration test — candidate pairs must pass an Engle-Granger or Johansen cointegration test to qualify for live trading; non-cointegrated pairs are dropped.
- Spread divergence trigger — when the standardized spread crosses a threshold (typically 2 standard deviations from the rolling mean), the pair becomes a live signal.
- Half-life sizing — estimated mean-reversion half-life informs position duration and time-stop.
- Stop-out logic — if the spread continues to widen past the entry threshold without showing reversion within the half-life window, the position is closed at a fixed loss rather than holding indefinitely.
When it fires
Pairs trading is a low-beta, market-neutral strategy that produces small consistent gains in stable markets and can struggle during regime shifts when within-sector correlation breaks down. The Alpha Suite implementation is designed to coexist with the directional strategies — pairs adds Sharpe ratio without correlating to insider buying or sector momentum signals.
What it does not catch: Structural breaks where the historical relationship genuinely no longer applies (M&A, capital structure changes, sector restructurings). Cointegration testing helps screen these out but does not eliminate the risk; the time-stop and stop-out logic are the second line of defense.
See live signals on the dashboard
Recon (free) gives you read-only access to today's ranked signals across all 10 strategies. No card, no commitment — sign up to see the model output for yourself.
Open the Dashboard