Pairs trading is a bet on a relationship rather than a direction. You find two instruments that normally move together, wait until they drift unusually far apart, then buy the one that is lagging and sell the one that has run ahead. If they return to their usual gap — as correlated things tend to — you profit from the two legs closing, whether the market as a whole rises, falls or does nothing. The direction of the market barely matters; the gap between the pair is the trade.
What it is
Pairs trading, sometimes called a spread or relative-value trade, works on the gap between two correlated instruments rather than the price of either one. You hold two positions at once — long one, short the other — so a broad market move lifts or drops both roughly equally and cancels out. What you are actually trading is the difference between them, and the bet is that the difference returns to its normal range.
Two things make a pair worth trading:
- Correlation — the two normally move together, so the spread stays in a fairly steady band.
- Mean reversion — when the spread stretches, it tends to snap back.
Get both and you have a relationship to trade. Lose either and you have two unhedged bets.
How it works
You track the spread — the gap between the two instruments — and measure how far it usually strays from its average. When it stretches past that, say two standard deviations, you take the trade: long the one that has lagged, short the one that has run ahead, in sizes that balance. When the spread returns to its average, you close both legs, and the convergence is your profit.
Spread = price of A - (ratio × price of B); trade when it stretches, close when it reverts.
Because the two legs offset, the trade is roughly market-neutral. A sell-off that drags both down at once should barely touch you. What you are exposed to is the relationship breaking, not the market falling.
A worked example
Take EUR/USD and GBP/USD, which usually move nearly in step, so the spread between them normally sits in a tight band. One week EUR/USD stalls while GBP/USD climbs, and the spread stretches to two standard deviations wider than usual. You go long EUR/USD and short GBP/USD, £10 a pip each. Over the next fortnight they converge: EUR/USD gains 60 pips, GBP/USD gives back 40. The long leg makes £600, the short leg £400, £1,000 in total, and you never needed to know whether the dollar was rising or falling. The losing version: they never reconverge, GBP/USD keeps climbing, and you close both at a £700 loss. The bet was reversion; when the relationship breaks instead, the trade breaks with it.
Why it matters
Pairs trading appeals because it strips out market direction — the hardest thing to predict — and leaves you trading a narrower, steadier relationship. Done with discipline it can earn in flat and falling markets alike. The catch sits in the one assumption: that the correlation holds. Correlations that lasted for years break without warning when something fundamental changes for one instrument. When that happens the spread does not revert, it runs, and both legs can bleed at once. Market-neutral is not risk-neutral.
Common misconceptions
- Market-neutral does not mean safe. If the correlation breaks, both legs can lose together and the hedge evaporates.
- Correlation is not permanent. Two instruments that tracked each other for years can decouple in a week. Yesterday's correlation guarantees nothing.
- Two positions, two lots of cost. You pay spread and swap on both legs, every trade, which quietly eats a thin reversion edge.
- Reversion can outlast your account. The spread can stretch further before it snaps back, if it snaps back at all, and margin does not care about your thesis.
The Karnek note
A pairs trade lives or dies on both legs behaving, so you want to see them together in real time. Karnek reads the live account from your terminal and shows every open position, the equity and the drawdown as the spread moves. It is read-only by design and can never open, close or alter a trade — it only reports what yours are doing.
Written and reviewed by the Karnek Research team. Last updated August 2026.
Educational content only - not financial advice. Past performance does not predict future results. Trading carries significant risk.