Market Neutral
Pairs Trading / Statistical Arbitrage
Not a bet on either stock going up or down. A bet on the gap between two correlated stocks snapping back to normal after it stretches too far.
Set the thresholds, watch the spread trade itself
Stock A and Stock B move together most of the time. The bottom strip tracks the z-score of their spread against its own recent history. Set how far that z-score has to stretch before you'd open a trade, and how close to normal it has to come back before you'd close it. Every slider move re-runs the full session instantly.
How it works
- The trade is on the spread, not on either stock. Buy one leg and sell the other in roughly equal size, so a move in the overall market that pushes both stocks the same direction mostly cancels out. What's left is the relationship between them.
- The z-score measures how stretched that relationship is right now. It's the current spread compared to its own recent average, expressed in standard deviations. A z-score of 2 means the spread is about two standard deviations wider than its recent norm, unusual, not impossible.
- Entry and exit are two different thresholds on purpose. Entering at a wide stretch (say 2.0) and exiting near normal (say 0.4) captures the reversion instead of trying to catch the exact peak or the exact return to zero, which nobody can time consistently.
- This only works if the relationship is real. Two stocks in the same sector with a genuine business link tend to snap back together. Two stocks that happened to move together for a few months by coincidence can drift apart permanently, and every signal built on their old relationship stops meaning anything.
Where this breaks
The relationship itself breaks, and the z-score can't tell you that in real time
A z-score is calculated from the pair's own recent history, so when the relationship genuinely changes, a merger falls through, one company's business model shifts, a sector rotation splits two former peers, the spread doesn't revert. It becomes the new normal, while the strategy keeps reading the growing gap as an even bigger reversion opportunity and adding to a trade that's fighting a real structural change instead of noise. The strategy has no built-in way to tell "unusually stretched but still the same relationship" apart from "the relationship stopped being true."