Pairs Trading Basics: Hedging One Position Against a Correlated One

A pure directional bet on one stock or contract puts all of your risk on one question: will it go up? A pairs trade splits that into a narrower question. You go long one asset and short a related one, and you're really betting on how the two move relative to each other.

That sounds safer, and sometimes it is. But it swaps one set of risks for another, and sizing the two legs correctly is where most beginners go wrong.

The mechanics

The classic version takes two assets that usually move together, such as two competitors in the same industry or two related ETFs. When their price ratio stretches far from where it normally sits, you buy the laggard and short the leader, expecting the gap to narrow. If the whole market falls, the long loses and the short gains, so the broad move partially cancels. What remains is exposure to the spread.

The word to underline is partially. Two assets with a high historical correlation can still diverge for company-specific reasons, and the relationship that held in the past is an assumption, not a guarantee.

A relative-value thesis (hypothetical)

Suppose two competing medical device makers, Company A and Company B, sell similar products. You believe A is taking customers from B. A pairs expression is short B and long A. If you're right, B underperforms A. If the whole sector sells off, the short on B offsets some of the loss on A.

Now the sizing. Equal dollars is the simplest approach: $10,000 long A and $10,000 short B. But if A has a beta of 0.8 and B has a beta of 1.2 (hypothetical numbers), equal dollars leaves you net short market exposure, because the short leg is more sensitive. To beta-match, you'd need about $8,000 of beta exposure on each side, so $10,000 long A (0.8 x 10,000) pairs with roughly $6,667 short B (1.2 x 6,667 = 8,000). That's the difference between dollar neutral and beta neutral, and it's why a real pairs trade usually carries some leftover market exposure.

Applying it to futures: ES versus NQ

ES and NQ are both US equity index futures with different composition, so a spread between them is a way to express a view on tech versus the broader market. The sizing problem is the same. A hypothetical example: if ES sits at 6,000 and NQ at 21,000, one ES contract controls 6,000 x $50 = $300,000 of notional, and one NQ controls 21,000 x $20 = $420,000. A notional-matched spread would be about 7 ES against 5 NQ. With micros, 7 MES against 5 MNQ lands in the same place ($30,000 and $42,000 per contract, so 5 MNQ is $210,000 and 7 MES is $210,000).

Matching notional doesn't match risk. NQ typically moves more in percentage terms than ES, so a notional-neutral spread can still be directionally long or short the market. Volatility weighting or beta weighting is a more careful alternative, and the right ratio is something you measure, not copy from an article.

Calendar spreads

A calendar spread is long one expiry and short another in the same product, such as a nearer and a later ES expiration. The price relationship is tied to the cost of carry, so it tends to move less than outright price, but it isn't fixed. Margin treatment for spreads varies by broker and exchange, so check the actual margin requirement with your broker and the CME product page before assuming a discount.

Risks that remain

• Basis risk: the pair stops behaving like a pair. The relationship you modeled breaks, and both legs can move against you at once.
• Timing risk: a thesis can be right and still take far longer than your capital or patience allows.
• Short-side costs: in stocks, shorting involves borrow availability, borrow fees, and the possibility of a squeeze.
• Double costs: two legs mean two commissions and two sets of slippage.
• Margin: running both legs usually needs more margin than one. Margin is collateral, not a cap on loss.

For stock pairs in a margin account, rules around day trading are in transition. FINRA's pattern day trader requirements were replaced by intraday margin rules effective June 4, 2026, with brokers allowed to transition through October 20, 2027, so some brokers may still apply the older rules for now. The Investor.gov glossary entry explains the change. Futures accounts are different: they run on exchange and broker margins, not this securities rule.

Write the exit before you enter

For every pair, define three things up front: the spread level where the thesis is wrong, the maximum dollar loss you'll accept on the combined position, and a time limit. If the spread hasn't moved by then, close it. Without those, a pairs trade tends to drift into a position you hold only because you're already in it.

To check whether a spread idea has any positive expectancy at all before risking money, run your numbers through the edge calculator using realistic costs for both legs.