17 August 2026
Trading
8 Minutes Read

Statistical Arbitrage: The Complete Guide to Pair Trading

Statistical arbitrage looks for stock pairs that have historically shown related price movements. Traders monitor these relationships for periods when prices diverge. They may then use trading strategies based on an expected convergence. Retail traders can study such strategies using NSE trading data, statistical tools, and suitable risk controls.

💡 Quick Answer
Statistical arbitrage looks for stock pairs whose prices have historically moved together, then studies periods when they diverge. Pair trading takes opposite positions in the two securities on an expected convergence. Neither correlation nor cointegration guarantees prices will converge, so relationships can widen further and produce losses.

What is statistical arbitrage?

Statistical arbitrage uses relative price relationships between correlated securities rather than relying only on market direction. Traders look for stock pairs whose prices move apart from their historical relationship. They may take positions based on an expected relationship between the securities.

For example, ITC and HUL may show related price movements over a period. If their relationship changes, traders may study the price divergence before taking positions. However, the relationship may not return to its earlier pattern, creating a risk of losses.

Statistical arbitrage often uses mean reversion as part of its analysis. However, historical relationships may change due to company, sector, or market factors. Retail traders should understand these risks before using such strategies.

How Does Pair Trading Identify Market Opportunities?

Pair trading can help identify market opportunities by comparing stock pairs with historically related price movements. Traders may assess historical correlation and look for price divergence between the securities. However, a correlation above 0.7 does not guarantee that prices will converge again.

Traders may take opposite positions by buying one security and shorting the other. They may then monitor the spread and close positions if the relationship changes as expected. However, the spread can widen further, resulting in losses.

Pair trading may involve NSE trading, short selling, and margin requirements. Traders should understand the applicable rules and risks before using such strategies. The Navia All-in-One App provides market data and trading tools for analyzing securities.

Mean Reversion and Cointegration

Mean reversion describes a tendency for a stock pair’s price relationship to move back toward its historical range. However, cointegration does not prove that prices will always converge.

For example, two securities may show a stable long-term relationship based on historical data. Traders can test this relationship using statistical methods before considering a spread trading strategy.

The Z-score measures how far a spread has moved from its historical mean. Traders may use a chosen Z-score level as part of their strategy, but it does not guarantee a profitable trade.

How Are Stock Pairs Identified?

Pair trading analysis can compare two securities with similar historical price movements. Analysts may study their historical correlation and price relationship over time.

However, correlation alone does not show that two securities will continue moving together. Market conditions and company-specific factors can change these relationships.

Therefore, correlation should serve as one analytical measure rather than a standalone basis for trading decisions. NSE trading data can support this type of historical analysis.

How Does a Pair Trading Strategy Work?

Pair trading typically involves analyzing stock pairs with historically related price movements. Traders may back test these relationships and use statistical methods, such as the Engle-Granger test, to assess cointegration.

A Z-score can help measure how far a spread has moved from its historical mean. Traders may use predefined thresholds when designing a spread trading strategy. However, these thresholds do not guarantee that prices will converge.

Pair trading may involve taking opposite positions in two securities. Traders should consider position sizing, margin requirements, transaction costs, and potential losses. A market neutral structure does not eliminate market or strategy-specific risks.

Back testing can help assess how a strategy performed on historical data. However, historical results do not guarantee similar outcomes in live trading.

Backtesting Before Live Trading

Backtesting can help evaluate a pair trading strategy using historical NSE trading data. Analysts may review factors such as win rate, returns, drawdowns, and profit factor.

However, fixed thresholds such as a 1.5 profit factor or 45% win rate do not suit every strategy. Historical results also cannot guarantee similar outcomes in live markets.

Therefore, use backtesting as one part of strategy evaluation. Also consider transaction costs, liquidity, execution conditions, and market changes before using any strategy.

What Rules Apply to Pair Trading?

  • Statistical arbitrage and pair trading are subject to applicable securities market rules. Traders must follow the requirements for short selling, margins, position limits, and order execution.
  • Short selling: Short selling is permitted subject to applicable NSE and regulatory requirements. Traders should check the latest eligible securities and applicable conditions before placing orders.
  • Margin requirements: Margin requirements depend on the security, position, market conditions, and applicable exchange rules. Therefore, avoid using a fixed percentage for every pair.
  • Position limits: Traders must follow applicable position limits and other restrictions for the securities they trade.
  • Intraday trading: Choosing an intraday structure does not remove applicable trading risks or regulatory requirements. Traders should understand settlement, margin, and position rules before trading.
  • For any specific platform, verify the applicable trading rules and margin requirements before placing orders.

Getting Started With Pair Trading

Start by learning how pair trading works before considering live trades. Paper trading can help you understand how stock pairs and correlated securities behave over time.

You can study historical price relationships and calculate a Z-score to assess how far a spread has moved from its historical range. However, no fixed Z-score guarantees a valid trading signal or profitable outcome.

Keep a record of your analysis and review the strategy’s results, costs, and risks. If you later consider live trading, understand the applicable margin, settlement, and market requirements first.

Analyse NSE stock pairs and market data on the Navia All in One App

Common Mistakes in Pair Trading

Do not skip backtesting when evaluating a pair trading strategy. Historical relationships can change, and correlated securities may move apart for extended periods.

Cointegration can also weaken or break over time. Therefore, traders should regularly review their assumptions and strategy performance. Trading costs, including NSE trading charges, can also affect overall results.

Keep detailed records of each trade and account for applicable costs. This can help retail traders evaluate strategy performance more realistically without assuming that past relationships will continue.

Conclusion

Statistical arbitrage and pair trading involve strategies that use relationships between stock pairs and mean reversion. However, these strategies carry risks, and historical relationships may change over time.

Traders can use cointegration, backtesting, and trading signals to evaluate potential strategies. They should also consider transaction costs, margin requirements, liquidity, and applicable SEBI rules.

Market opportunities may arise when price divergence occurs between related securities. However, no strategy guarantees profits. The Navia All-in-One App provides market data and trading tools that can support analysis of NSE trading activity.

Key Takeaways

  • Statistical arbitrage looks at relative price relationships between correlated securities rather than market direction, but historical relationships may change due to company, sector, or market factors.
  • A correlation above 0.7 does not guarantee that prices will converge again, and correlation alone does not show that two securities will continue moving together.
  • Cointegration does not prove that prices will always converge, and a chosen Z-score level does not guarantee a profitable trade.
  • Fixed thresholds such as a 1.5 profit factor or 45% win rate do not suit every strategy, and historical results cannot guarantee similar outcomes in live markets.
  • Pair trading is subject to applicable requirements for short selling, margins, position limits and order execution, and a market neutral structure does not eliminate market or strategy-specific risks.

DID YOU FIND THIS INTERESTING?

We’d love to hear from you –

yes or no feedback form

Frequently Asked Questions

How much capital is required to start pair trading?

There is no universal capital requirement for pair trading. The amount required depends on factors such as the securities selected, position sizes, applicable margin requirements, transaction costs and the trading approach used. Investors should understand the applicable risks and costs before considering the strategy.

How long is a pair trade typically held?

The holding period can vary depending on the strategy, securities selected, market conditions and the behavior of the price spread. There is no fixed timeframe within which a price divergence will necessarily reverse.

When is a pair trade considered?

Statistical arbitrage strategies may use statistical measures such as Z-scores to identify deviations in a spread. The interpretation and thresholds used can vary by methodology. A statistical signal does not guarantee that the spread will revert.

What is the biggest risk in pair trading?

Mean reversion may not occur as expected, and relationships between securities can change over time. Factors such as corporate events, changes in fundamentals and market conditions can affect the relationship between a pair.

Can I backtest a pair trading strategy?

Yes. Backtesting can be used to study how a strategy would have behaved using historical data. However, historical results do not guarantee future performance, and statistical relationships identified in historical data may change.

What correlation is required for pair trading?

There is no single correlation threshold that is appropriate for every pair trading strategy. Correlation measures how securities have moved together historically, while cointegration examines a different statistical relationship. Both should be evaluated in the context of the strategy and data used.

How do I identify potentially cointegrated stock pairs?

Statistical tests such as the Engle-Granger methodology can be used to examine whether a long-term statistical relationship exists between securities. Results depend on the data, methodology and period analysed.

Is pair trading legal in India?

Pair trading and statistical arbitrage are strategies that involve buying and selling securities, subject to applicable market rules, exchange regulations, SEBI requirements, margin requirements and restrictions on short selling. Investors should be aware of the rules applicable before trading.

What is a good profit factor for pair trading?

There is no universally applicable profit-factor or win-rate threshold that guarantees a successful strategy. Historical performance metrics should be evaluated alongside drawdowns, transaction costs, market conditions and other risk measures.

How often should stock pairs be re-evaluated?

The statistical relationship between securities can change over time. The appropriate review frequency depends on the methodology, securities and timeframe being analysed. Historical relationships should be periodically reassessed rather than assumed to remain constant.

Can leverage be used in pair trading?

The availability and amount of margin depend on the securities, trading segment, applicable exchange and regulatory requirements. Leverage can increase both potential gains and losses, so investors should understand applicable margin requirements and risks before using it.

How is trading signal strength measured?

Z-score is one statistical measure that can be used to assess how far an observation is from its historical mean. Different strategies may use different thresholds and methodologies. A higher Z-score does not by itself guarantee that a price spread will revert.

DISCLAIMER: This article is for educational purposes only. It does not constitute investment or trading advice. Navia Markets Ltd does not recommend statistical arbitrage or pair trading strategies and does not guarantee profits. Statistical arbitrage and pair trading involve substantial risk, including potential losses exceeding capital deployed. Market opportunities may not persist. Backtesting results may not replicate live performance. Retail traders must understand the risks, test thoroughly, and only trade with capital they can afford to lose. SEBI-registered advisers can provide personalized guidance. Understand margin calls, shorting rules, and position limits before beginning. Past performance does not indicate future results.