Pairs Trading Strategy Guide: Rules, Tools & Examples

The strategy rests on a concept called cointegration — a stronger condition than simple correlation. Two instruments can be 85% correlated yet drift apart permanently. Cointegrated pairs, by contrast, share a long-run equilibrium relationship: deviations are temporary, and prices are statistically compelled to reconverge. The Engle-Granger cointegration test provides the mathematical framework to distinguish genuine pairs from coincidental correlations.

Consider AUDUSD and NZDUSD. Both currencies are driven by commodity cycles, Chinese demand, and global risk sentiment. Their price ratio has historically oscillated within a narrow band, with divergences exceeding 2 standard deviations resolving within 5 to 15 trading days in the majority of observed cases. The same dynamic applies to XAU/USD and XAG/USD — gold and silver share industrial and monetary demand drivers that periodically fall out of sync before realigning.

The Correlation Coefficient quantifies the linear relationship between two instruments on a scale from -1 to +1. Viable pairs typically require a coefficient above 0.80 over a rolling 60-day window. Below that threshold, the mean-reversion premise weakens considerably, and the strategy shifts from statistical arbitrage into speculative directional trading — a fundamentally different risk profile.

The Z-Score is the operational heartbeat of pairs trading. It measures how many standard deviations the current spread sits away from its historical mean. A Z-Score of 0 means the spread is at equilibrium. A reading of +2.0 signals the spread has widened two standard deviations above the mean — statistically unusual, and the trigger point most practitioners use for entry. The formula is straightforward: Z = (Current Spread − Mean Spread) / Standard Deviation of Spread.

Pairs trading operates on four timeframes of analysis but executes on three. The D1 chart establishes the cointegration baseline and the historical spread mean. The H4 chart identifies the developing divergence. The H1 chart times the precise entry, confirming the spread has stalled or begun reversing before committing capital.

Entry Conditions (all must be met simultaneously):

First, the 60-day Correlation Coefficient must read above 0.80. Any reading below this level disqualifies the setup regardless of the Z-Score.

Second, the Engle-Granger cointegration test p-value must be below 0.05, confirming the pair shares a statistically significant long-run relationship. This test should be recalculated monthly.

Third, the Z-Score of the spread must reach ±2.0 or beyond. A reading above +2.0 means Instrument A is overpriced relative to Instrument B — sell A, buy B. A reading below -2.0 means the opposite — buy A, sell B.

Fourth, on the H1 chart, the Spread Chart must show at least one candle closing back toward the mean, confirming initial reversion momentum rather than a continuing trend.

Exit Conditions:

The primary profit target is Z-Score reversion to 0.0 — full mean reversion. This corresponds to a 1:2 risk-reward ratio on most qualifying setups. A secondary, conservative exit at Z-Score ±0.5 captures partial profit while reducing exposure, aligning with the 1:1.5 minimum target.

Stop-loss placement is non-negotiable: close both legs of the trade if the Z-Score extends to ±3.0 in the wrong direction. This level signals a possible regime change — the historical relationship may be breaking down rather than temporarily dislocating. Holding beyond this threshold transforms a statistical strategy into hope-based speculation.

For the EURUSD/GBPUSD pair specifically, a Z-Score of +2.0 has historically corresponded to spread divergences of approximately 120 to 180 pips in the ratio chart, with mean reversion completing within 3 to 12 days on H4 timeframes, based on backtested data from 2018 through 2023.

Most traders assume pairs trading means buying one lot of each instrument. That assumption is incorrect — and expensive. Because two correlated instruments rarely move in equal pip increments, an equal-lot approach creates hidden directional exposure that undermines the market-neutral premise entirely.

Proper position sizing in pairs trading requires dollar-value neutralization. The goal is for each leg to carry the same monetary exposure so that a 1% move in Instrument A is offset by a corresponding 1% move in Instrument B.

The hedge ratio — calculated via linear regression of one instrument's price against the other over the lookback period — determines the correct lot size for each leg. If the regression slope between XAU/USD and XAG/USD is 1.35, a position of 1.0 lot in gold requires 1.35 lots in silver to achieve neutrality.

Position sizing rules for this strategy:

— Risk no more than 1% to 1.5% of account equity per pair on the combined two-leg position. — Calculate the dollar value of one pip for each instrument and adjust lot sizes so both legs carry equal dollar exposure. — Maximum concurrent pairs: 3. Beyond this, correlation between pairs themselves can create portfolio-level concentration risk — particularly when multiple pairs share a common driver such as USD sentiment. — In volatile macro environments (central bank decision weeks, NFP releases), reduce position size by 30% to 50%. Macro shocks can temporarily overwhelm statistical relationships, triggering stops before reversion occurs.

For a $10,000 account using a 1% risk rule, maximum loss per trade is $100. If the stop is placed at Z-Score 3.0 and entry occurs at Z-Score 2.0, the spread between entry and stop represents one standard deviation of the spread — a measurable, fixed dollar amount that determines exact lot sizes for each leg.