What is pairs trading?

Some companies are so similar that their share prices tend to travel together. Two freight railroads serving the same region, two off-price retailers chasing the same shoppers, two large banks exposed to the same interest rates. Day to day their prices wobble, but over months the relationship between them is fairly stable.

Pairs trading is built on that relationship. Instead of asking whether one stock will go up or down, it asks a narrower question: the two have drifted unusually far apart — will the gap between them close again?

The basic idea

• Pick two stocks with a long, measurable history of moving together.
• Measure their normal relationship — how far apart they usually sit, and how much that distance typically wobbles.
• When the distance stretches well beyond its usual range, a pairs trade takes the two sides at once: it buys the stock that has fallen behind and sells short the one that has run ahead.
• If the relationship holds and the gap closes, both sides can contribute. If the gap keeps widening, both sides can lose — which is why every pairs setup needs a pre-set exit for when it goes wrong.

Why it is called market-neutral

Because a pairs trade holds one stock long and a similar stock short, a broad market move tends to push both sides in the same direction and largely cancel out. What is left is the bet on the gap. That is the appeal — and also the limit: it removes much of the market's direction, not the risk that the two companies stop behaving alike.

What can go wrong

• The relationship breaks: one company changes (a merger, a lawsuit, a new business line) and the old link no longer applies.
• The gap keeps widening before it closes — sometimes far enough to hit the stop.
• Costs add up: two sets of trading costs, plus a fee to borrow the stock that is sold short.
• Short selling can lose more than was put in, because a stock's price has no ceiling.

How PairsCopilot fits in

PairsCopilot runs one statistical model every US market evening over a list of pairs that have each passed a qualification test. It publishes how far each pair has drifted from normal and how long similar gaps have taken to close, the same for every reader at the same time. It is data — the research, the decision and the risk stay with the reader.