PairScanner

Start here

What is pair trading?

The idea

Pair trading combines buying one asset with selling another short. The mean-reversion approach looks for an unusual gap between related assets and tests the idea that the gap will narrow.

A gap opens, then narrowsFictional paths · indexed to 100 at the start
Asset AAsset B
A gap opens, then narrowsFictional indexed paths. At the entry marker A lags B; after entry A rises and B falls, narrowing their relative gap. This shows a possible outcome, not a forecast.100110Entry hypothesisEarlierLater
  1. ObserveA has lagged behind B.
  2. Pair the positionsBuy A · sell B short.
  3. Follow the relationshipThe gap may narrow or widen.

Buying A creates the long position. Selling B short means selling borrowed shares, then buying them back later. The result comes from both positions together.

How a pair trade makes or loses money

Assume the trader expects A to catch up relative to B. Both positions start at $1,000. Select a scenario to see how each leg contributes.

Buy A10 × $100$1,000 long
Sell B short20 × $50$1,000 short
Explore both outcomes

Gap narrows

A: $100 → $104B: $50 → $49
Long A+$40
Short B+$20
Pair result before costs+$60

Gap widens

A: $100 → $96B: $50 → $52
Long A−$40
Short B−$40
Pair result before costs−$80

Fictional examples, before fees, borrowing charges, financing and dividends owed on the short leg. Gross exposure is $2,000; this is not the cash committed or the margin requirement.

Must the market go up?

No. With the same starting positions, A outperforming B by three percentage points produces $30 before costs in either of these fictional scenarios.

Asset AAsset B
Both assets rise
Both assets riseFictional path: A ends at 104, B at 101, both indexed to 100 at entry. A outperforms B by three percentage points.96100104EntryExit
A +$40
B −$10
+$30

Before costs · same starting positions

Both assets fall
Both assets fallFictional path: A ends at 99, B at 96, both indexed to 100 at entry. A outperforms B by three percentage points.96100104EntryExit
A −$10
B +$40
+$30

Before costs · same starting positions

Market neutrality is an objective. Equal dollar amounts do not remove differences in market sensitivity, sector exposure or changing relationships.

Three terms you will meet

Spread

The relationship you measure

A−β × B=spread

For a simple price-difference model: A = $100, B = $50 and β = 2 gives a spread of $0. PairScanner uses log prices; its β is not a share count.

Cointegration

Does the relationship hold?

Stable fluctuations versus driftA fictional orange line fluctuates around zero; a dashed blue line drifts upwards.+0+4EarlierLater
Around equilibriumDrifting away

A cointegration test looks for evidence of a stable combination of price series over the period tested. This sketch illustrates the idea; it is not a statistical test result.

Z-score

How unusual is the spread?

If the mean is 0, the standard deviation is 1 and the spread is −2, then Z = −2. That means two standard deviations below the mean, not a probability of profit.

What can go wrong?

An unusual gap does not have to close. The relationship can change, and a pattern found in historical data can also be a coincidence.

The spread can keep moving awayFictional spread · arbitrary units, not a live Z-score
The spread can keep moving awayFictional spread: after an entry at minus two, it drops to minus four instead of returning to zero.-4-2+0EntryEarlierLater

Relationship breaks

A takeover or a change in business can create a new equilibrium.

Borrowing and costs

Borrow fees, dividends and recalls can change the trade.

Uneven execution

One filled leg leaves exposure until the other fills.

Losses on an individual short position can be unlimited in theory. Margin calls and forced exits are possible; a stop order does not guarantee an execution price.

How to explore without placing a trade

  1. Read an example

    Inspect the charts and their dates.

  2. Check the relationship

    Compare tests, context and limits.

  3. Build a watchlist

    Follow a hypothesis before placing orders.

Read a dated analysis first. These public studies show different conclusions, including a failed cointegration test, and are available without an account.

Discover pairs beyond the names you know

You do not need to know which companies to pair. PairScanner’s algorithm finds candidates within its supported universe and performs the statistical calculations automatically. Open the scanner: the pair analyses are already there.

Pair-trading tools at a glance
What you getProfessional softwareOther retail softwarePairScannerReady-to-use pair analysis
AccessSubscriptionLicence and data costs depend on the offerFree + paidFeatures depend on the plan$0Sign up with just your email
Finding pairsMarket screening and quantitative research tools.Screeners, charts and indicators, depending on the software.Pairs discovered automatically. Including companies you have never searched for.
CalculationsAdvanced analytics and customizable models.Built-in indicators and analysis tools; pair features vary.The algorithm does the calculations automatically: cointegration, hedge ratio, spread and Z-score.
Historical depthHistorical datasets, with coverage depending on the licence.History varies by market, data provider and plan.Five years of daily data for pair analysis, subject to available common history.
Getting resultsResearch workflows and automation options.Scans, scripts and alerts, depending on the software.Results calculated daily. No code to write, no pair to enter manually.
Your workflowBroad market research and analysis tools.Charting, technical analysis or specialized tools.Pair scanner, analysis and monitoring in one place. Watchlists and Z-score email alerts included.

PairScanner: no card required. Creating a free account does not start a paid subscription or authorize automatic charges.

General overview: features vary by software and plan. PairScanner is an analysis tool; orders stay with your broker.