Tell us how you trade and we'll point you to the right research segment.
Talk to Our Team →Start with our beginner-friendly guides on market basics, order types, and risk management before you place your first trade.
Start Learning → Browse All Articles →Pairs trading involves simultaneously taking a long position in one instrument and a short position in a related instrument, structured so that the resulting combined position profits from a change in the relationship between the two rather than from the overall market moving in either direction. Because gains from one leg are intended to be roughly offset by losses on the other whenever the broader market moves as a whole, the strategy is generally described as market-neutral, at least in its intended design. This piece works through how a pair is actually selected, how the spread between the two instruments is built and monitored, how this approach differs from a momentum-based strategy, and where the relationship underlying a pair can break down in practice.
Complete Guide
Every Section
Takeaways
The market-neutral property of pairs trading comes directly from its structure: holding a long position in one instrument and a short position in another, ideally of similar size, means that a broad move in the overall market that affects both instruments similarly largely cancels out across the combined position. What remains is exposure specifically to the difference in how the two instruments perform relative to each other, rather than exposure to the market’s overall direction.
This neutrality is a design goal built into the position’s structure rather than a guaranteed outcome, and it depends heavily on how well-matched the two legs actually are in practice. A pair where the two instruments do not actually respond similarly to broad market moves will retain some genuine directional exposure even while nominally structured as long one side and short the other, which undermines the very property that makes the strategy attractive in the first place.
Choosing which two instruments to pair together is the single most consequential decision in the entire strategy, since the whole approach depends on the two instruments having a genuine, persistent relationship rather than a coincidental historical similarity. Instruments operating in the same sector, facing similar underlying economic drivers, are the most common starting point for identifying candidate pairs.
Beyond a shared sector, a more rigorous approach looks specifically for cointegration between the two price series — a statistical property indicating that while each instrument’s price may wander unpredictably on its own, the specific difference or ratio between the two tends to revert back toward a stable long-run relationship over time. This is a more demanding and more specific requirement than simple historical correlation, and it is generally considered the more reliable foundation for actually building a pairs trade around.
Two instruments can show a high historical correlation in their price movements without actually being cointegrated, meaning their prices tend to move in the same direction on average without the specific spread between them ever reliably reverting to a stable relationship. Relying purely on correlation risks building a pair around two instruments that happen to have moved similarly over a specific historical window for reasons that have little to do with any genuine, persistent economic relationship between them, which is precisely the kind of coincidental pattern that tends not to hold up once relied upon going forward.
A useful habit is testing candidate pairs across more than one historical period, rather than confirming a relationship over a single window and assuming it will hold going forward unchanged. A pair that appears cointegrated across one specific stretch of history but not across an earlier or more recent stretch is a weaker candidate than one that shows the same reverting relationship consistently across multiple, distinct historical periods, since consistency across different periods is a stronger indication of a genuine, structural relationship rather than a pattern specific to one particular window.
Once a pair is chosen, the strategy is built around continuously tracking the spread between the two instruments — commonly the price difference or the price ratio, depending on how similarly scaled the two instruments are — and comparing the current spread against its own historical average and typical range of variation.
A spread that has moved unusually far from its historical average, by a margin considered statistically meaningful relative to how much it typically varies, is treated as a signal that the relationship between the two instruments has temporarily stretched and may be due to revert. The specific threshold used to define “unusually far” is a deliberate design choice within the strategy, balancing how often a signal is generated against how confident that signal genuinely reflects a stretched relationship rather than ordinary day-to-day noise.
The entry logic for a pairs trade typically triggers when the spread crosses the chosen unusual threshold in one direction — going long the relatively weaker-performing instrument and short the relatively stronger-performing one, on the expectation that the spread will eventually revert back toward its historical average, closing the gap that triggered the entry in the first place.
The exit logic then typically closes the position once the spread has reverted back to somewhere close to its historical average, capturing the profit from that reversion, or alternatively closes the position at a defined loss if the spread continues widening well beyond the entry point rather than reverting as expected, since a spread that keeps widening indefinitely suggests the underlying relationship may have genuinely broken down rather than simply stretched temporarily.
A time-based exit is sometimes added as a further safeguard alongside these two rules, closing a position that has neither reverted nor breached the stop-loss threshold after a defined holding period has elapsed. This addresses the practical concern that capital tied up in a spread that is neither reverting nor clearly breaking down is capital not available for a fresh opportunity elsewhere, even if the existing position has not technically triggered either the profit-taking or loss-limiting rule on its own.
Pairs trading and momentum-based approaches rest on essentially opposite assumptions about how prices behave. A momentum strategy assumes that a recent trend, once established, tends to persist for some further period, and positions itself to benefit from that trend continuing. Pairs trading instead assumes that a specific relationship between two related instruments tends to revert back toward its typical historical pattern after stretching too far, and positions itself to benefit from that reversion happening.
This makes the two approaches structurally complementary rather than interchangeable — a strategy well-suited to genuinely trending conditions is not necessarily well-suited to a mean-reverting relationship between two related instruments, and vice versa. Some practitioners run both types of strategies simultaneously across different parts of a portfolio specifically because the conditions that favor one often do not favor the other, offering some diversification benefit across strategy types rather than relying on either approach exclusively.
The central risk in pairs trading is not market direction, given the strategy’s structural neutrality to broad moves, but the possibility that the historical relationship the entire trade is built around genuinely changes or breaks down rather than simply stretching temporarily before reverting. A shift specific to one of the two instruments — a change in its underlying business, a shift in its competitive position relative to the other instrument in the pair — can permanently alter the relationship rather than producing a temporary, reversible divergence.
Because this kind of structural break is not always immediately obvious while it is happening, a pairs trade that keeps widening well past its historical typical range can be genuinely difficult to distinguish, in real time, from an unusually large but still ultimately temporary divergence that will eventually revert. This is precisely why a defined stop-loss on the spread’s own continued widening, rather than an indefinite hold waiting for reversion that may never come, is treated as an essential part of the strategy’s risk management rather than an optional addition.
Getting the relative sizing of the long and short legs right matters as much as selecting the pair itself, since a mismatch in sizing directly undermines the intended market-neutral property of the position. Simply matching the number of units on each side is rarely sufficient if the two instruments have meaningfully different price levels or different sensitivities to broad market moves.
A more careful approach sizes each leg so that the rupee value of exposure, adjusted for each instrument’s own sensitivity to broad market moves, is roughly balanced across the long and short sides. This calibration is worth revisiting periodically over the life of a position, since the relative sensitivities of the two instruments can themselves shift gradually over time even while the pair otherwise continues to behave as expected, and a sizing ratio that was well balanced at entry can drift out of balance well before the spread itself shows any obvious sign of reverting or breaking down.
Pairs trading tends to work best in markets or sectors with a genuinely stable, well-understood set of relationships between related instruments, and in conditions that are not dominated by an overwhelming, market-wide directional force that can distort even fundamentally sound pair relationships for an extended period. A period of broad, indiscriminate market stress, where correlations across nearly everything spike toward one regardless of underlying fundamentals, tends to be a genuinely difficult environment for pairs strategies built on more typical, calmer relationships.
Conditions of moderate, range-bound volatility, where individual instrument relationships have room to drift and revert without being overwhelmed by an overriding market-wide trend, are generally considered more favorable for this kind of strategy. Recognising which kind of environment currently prevails, rather than applying the strategy identically regardless of conditions, is part of using pairs trading with realistic expectations about when it is likely to actually perform as intended.
It is designed to be, since offsetting long and short positions are meant to cancel out broad market moves, leaving exposure mainly to the relationship between the two chosen instruments. Actual neutrality depends on how well the two legs are matched.
Correlation measures whether two prices tend to move in the same direction. Cointegration measures whether the specific spread between them reverts to a stable long-run relationship, which is a more demanding and generally more reliable basis for a pairs trade.
Pairs trading assumes a stretched relationship between two related instruments will revert toward its historical pattern. A momentum strategy assumes a recent trend will persist. The two rest on essentially opposite assumptions about price behavior, which is why they tend to perform well in different market regimes rather than the same one.
The risk that the historical relationship underlying the pair genuinely breaks down rather than temporarily stretching before reverting, which is why a defined exit if the spread keeps widening is considered essential rather than optional.
No. It tends to work better in conditions of moderate, range-bound volatility and struggles during periods of broad, indiscriminate market stress where correlations across most instruments move toward one, since that is precisely when historically stable relationships between paired instruments are most likely to break down rather than simply stretch and revert.
Explore our Equity Tips Provider service or get in touch with our research team.