Forex Statistical Arbitrage Made Simple: A No-Code Playbook

Years ago, when I first got interested in markets, I showed an experienced trader my research set up – machine learning frameworks, Kalman filters, Hurst exponents. If it had a fancy name and sounded complicated, I was all over it.

“What’s your edge?” my trader friend asked.

I launched into a passionate explanation about predicting the euro’s next big move based on some complicated technical signals.

He nodded politely, then asked, “So you’re trying to outguess every bank, hedge fund, and trader on which way EUR/USD is heading?”

This was slightly deflating, but turned out to be a pivotal moment in my career.

“What if,” he suggested, “instead of predicting where one market is going, you just bet on the relationship between two returning to normal after it gets out of whack?”

That’s the essence of statistical arbitrage – trading relationships rather than direction.

This approach has made more money over the course of my career than any other, in both pro and independent settings. And it is still a viable approach today.

And the beauty is, you can implement basic stat arb strategies without writing a single line of code.

What Is Stat Arb in Forex?

Statistical arbitrage in forex is all about exploiting temporary mispricings between related currency pairs rather than predicting absolute direction.

Instead of asking “Will the euro rise or fall?” (a brutally difficult question), you ask “Has the relationship between the euro and pound temporarily gotten out of balance?” (a more answerable question).

Think of it like this: directional trading is trying to guess the weather; stat arb is measuring the temperature difference between two nearby thermometers.

This approach is particularly appealing to systematic traders because it’s based on data and probabilities rather than gut feelings. You’re not trying to predict big macro trends or news events – you’re simply betting that historically relationships tend to remain stable over time.

This is not always a good bet, however, and no relationship is ever truly “stable” in trading.

Finding Related Currency Pairs: Your Stat Arb Hunting Ground

The foundation of any forex stat arb strategy is identifying pairs of currencies that typically move in tandem because they share similar economic drivers. These correlations create opportunities when one currency temporarily overshoots while its partner lags.

Don’t think that these relationships are stable and unchanging through time – they’re driven by all sorts of things that are difficult to predict. And really, that hints at the real game of stat arb trading – figuring out when to bet on a relationship being stable and when to pass.

Here are some of the most well-understood relationships:

The Commodity Currency Twins: AUD and NZD

The Australian and New Zealand dollars often (but not always) behave like synchronized swimmers. Both are export-driven economies heavily influenced by commodity prices and global risk appetite. Their geographical proximity and similar economic structures mean they tend to rise and fall together.

If you chart AUD/USD and NZD/USD together, they’ll typically move in near lockstep, with high correlations.

When one suddenly outperforms or underperforms the other without a fundamental reason, it signals a potential trading opportunity.

AUD/NZD – Image: TradingView

Oil-Linked Currencies: CAD and NOK

The Canadian dollar and Norwegian krone are both heavily influenced by oil prices due to their countries’ significant oil exports. When crude prices rise, these currencies typically strengthen (and vice versa).

CAD-NOK spread. Image: TradingView

The Japanese yen, conversely, tends to weaken when oil prices rise since Japan imports most of its energy. This creates interesting pairs like CAD/JPY, which often tracks oil price movements closely.

If oil surges but CAD/JPY doesn’t budge (or moves the “wrong” way), it signals a divergence that might revert once the correlation reasserts itself.

European Neighbors: EUR and GBP

The eurozone and United Kingdom have deeply intertwined economies, so EUR/USD and GBP/USD tend to trend together with high positive correlation. Despite Brexit, the fundamental economic ties remain strong.

EUR/GBP – Image: TradingView

When one of these pairs starts behaving differently from the other, it can signal a potential mean-reversion opportunity.

Safe Havens: JPY and CHF

The Japanese yen and Swiss franc often move together as both are traditional safe-haven currencies. During risk-off periods (market stress or volatility), both tend to strengthen as investors seek safety.

If one safe haven suddenly strengthens much more than the other without a country-specific reason, it may present an opportunity.

The key is to focus on currency pairs with comparable risk exposures. Their strong linkage is the statistical foundation for forex stat arb strategies.

Correlation and Mean Reversion: The Two Pillars of Stat Arb

Stat arb in FX rests on two core concepts: correlation and mean reversion.

Correlation: Finding the Dance Partners

Correlation measures how similarly two currency pairs move. If two pairs are highly positively correlated (near +1.0), they usually rise and fall together; if highly negative (near -1.0), one tends to rise when the other falls.

A practical first step is looking at a correlation matrix for major pairs over a recent period (e.g., 90 days) to spot strong relationships. Pairs like AUD/USD vs. NZD/USD, EUR/USD vs. GBP/USD, or USD/CAD vs. USD/NOK often show correlations above 0.7.

But correlation alone isn’t enough to justify a trade. It’s just a starting filter that helps you find potential opportunities.

Mean Reversion: The Rubber Band Effect

What we truly seek is mean reversion: the idea that if two related assets drift far apart, they will eventually snap back to their typical relationship.

Think of two mountain climbers tied by a rope. They might wander separately around obstacles, but they’re tethered – sooner or later, the rope pulls them back together.

An intuitive analogy: two drunk friends walking home. They may separate at times, but neither strays too far for long. This is exactly the behavior we want for pairs trading – short-term divergences with a long-term tendency to revert.

In statistics, cointegration is the formal way to confirm such a tether exists. If two currency pairs are cointegrated, they share a stable long-term equilibrium even though each moves a lot day-to-day.

However, I would caution you against using cointegration and other statistical tests – in my experience, they just aren’t that useful in the trading context. You can get a sense of this by calculating the cointegration relationship between two currencies in one period and then comparing it to the next – often, there is no relationship between the cointegrating relationship in the two periods.

Instead, look directly at the thing you’re interested in – for example, how often did a spread cross the zero line in a given period?

Thinking about time scales

Mid to high frequency stat arb in FX is hard for indie traders. Execution difficulties all but ensure this to be a losing approach.

But economic divergences and convergences play out over weeks and months rather than hours or days – so FX can lend itself to slower converging stat arb trades. You just need to have the patience for such an approach.

Implementing Stat Arb Without Code: A Step-by-Step Guide

You don’t need a PhD in math or complex algorithms to practice simple stat arb. A spreadsheet or even manual chart analysis can suffice. Here’s a practical approach:

1. Pick a Pair of Markets

Choose two instruments with a logical connection. It could be:

• Two FX pairs (both vs. USD), like AUD/USD and NZD/USD
• An FX pair and a related commodity, like USD/CAD and oil
• Two currencies against a common third currency, like EUR/USD and GBP/USD

Ensure there’s plenty of historical data to observe their interaction and a fundamental reason they should move together. Don’t just mine for random relationships – you’ll absolutely find stuff that worked by chance that has little probability of working out in the future.

2. Examine the Historical Relationship

Plot their price series over time to verify they usually move together. Many charting platforms let you overlay one chart on another. For instance, if you plot AUD/USD and NZD/USD, you’ll see they track closely, almost like the same line.

Returns to AUD/USD (blue) and NZD/USD (maroon). Image: TradingView

You may need to include the interest rate differential if it’s significant. This is especially true if your hold period is long.

You can also calculate a simple spread:

Spread = Price_A – β * Price_B

Where β is a hedge ratio (often 1 if the currencies are on a similar scale). If this spread fluctuates around a constant mean, it’s a good sign of mean reversion. In practice, this will never be clear-cut or constant.

You can eyeball it or calculate the mean and standard deviation of the spread in Excel. No fancy software required.

3. Define Entry Signals

Decide how far is “too far apart” for these two currencies. Using statistical thresholds helps – a common approach is to use 2 standard deviations from the rolling mean.

For example, if the current spread is greater than 2 standard deviations above its historical mean, you might take that as a sell signal on the spread (short the first currency, long the second). If it’s 2 standard deviations below the mean, that’s a buy signal.

AUD/NZD Bollinger Band Stat Arb – Image: TradingView

Some traders use percentiles or visual levels on a chart (like “the highest divergence in six months”) as a cue.

Pick a threshold that historically often precedes a reversion to normal. Remember that you’re not looking at each currency in isolation; you’re trading their relationship.

4. Execute the Trade in a Market-Neutral Way

Implement the trade by going long one side and short the other in appropriate amounts. Typically, you equalize the dollar value (or risk) of both legs.

For example, if you short $10,000 worth of AUD/USD, you’d also buy about $10,000 worth of NZD/USD for a roughly balanced spread trade.

This hedges out common factors (mainly USD in this example) so that your profit comes from the convergence of AUD and NZD rather than a big USD move.

No complex execution algorithms needed – you can manually place both trades in your platform and monitor the combined P&L.

5. Define Your Exit Strategy

Decide when to close the trade. Common approaches include:

• Exit when the spread returns to its mean (Z-score near zero)
• Use a time-based stop (exit after X days if the trade hasn’t worked)

I prefer not to use stop losses in general, but particularly in stat arb. It doesn’t make sense to use one, since the very thing that you’re betting on (spread divergence) is the thing you’re using to take you out of the position.

Instead, use a time stop informed by historical data – how long did the spread take to converge on average in the past?

A Real-World Example: AUD/NZD

Let’s walk through a concrete example with the AUD/NZD cross.

Historically, the AUD/NZD exchange rate might oscillate roughly between 1.04 and 1.12, centred around 1.08. Now it’s spiked to 1.13 after surprise news favoring Australia (for the sake of the example).

A stat arb trader sees this as an outlier – above the normal range. Without any code, they check past data and confirm 1.13 is indeed a multi-year extreme. They decide to short AUD/NZD, expecting it to fall back toward 1.08.

This can be done directly by trading the AUD/NZD pair if your broker offers it, or by shorting AUD/USD and simultaneously going long NZD/USD in appropriate amounts.

The position might take weeks or months to play out. If the currencies indeed revert (AUD weakens or NZD strengthens), the trader closes for a profit once AUD/NZD falls near the prior mean.

When Pairs Diverge: Understanding Why Relationships Break

Not every divergence is a trading opportunity. Understanding why a normally correlated pair might diverge helps you decide if a signal is valid or a trap.

You’ll never predict this with anything close to 100% certainty, but these are things worth thinking about.

Differing Economic News

Country-specific surprises can jolt one currency while barely affecting the other.

If New Zealand’s central bank announces an unexpected rate hike while Australia’s policy stays unchanged, NZD could jump independently of AUD. This might create a temporary divergence that eventually reverts once the market digests the news.

However, if the news marks a long-term policy divergence (like one country entering a multi-year tightening cycle while the other eases), the gap may persist.

Stat arb works best when the divergence is driven by random events or fundamental over- or under-reaction.

Commodity Price Moves

Currencies tied to commodities can diverge if their underlying commodities move differently.

If oil prices spike (boosting CAD) while gold prices fall (hurting AUD), these “commodity currencies” could temporarily decouple. If these moves reflect short-term supply/demand imbalances rather than structural shifts, the currency relationship is slightly more likely to revert.

Risk Sentiment Shifts

During sudden risk-on/risk-off swings, correlations can break down as safe havens (JPY, CHF) behave differently from risk-sensitive currencies (AUD, NZD).

EUR/USD and USD/CHF might correlate during calm times, but in a crisis, the CHF’s safe-haven status could make USD/CHF behave very differently.

Intervention or Policy Changes

Central bank interventions or major policy shifts can cause permanent changes in relationships.

The Swiss National Bank’s peg removal in 2015 is a famous example – EUR/CHF had been artificially fixed until suddenly it wasn’t. That extreme divergence crushed stat arb traders betting on a stable relationship.

Liquidity and Flow Factors

Sometimes divergences aren’t about fundamentals at all. Market flows – big fund reallocations, option barriers being triggered, or holidays in one region – can temporarily distort one currency.

These are usually short-lived and prime candidates for mean reversion once normal trading resumes.

The key is judging whether a divergence is temporary noise or a fundamental shift. Small, unexplained divergences typically mean-revert – those are your bread-and-butter trades. But genuine regime changes can create persistent divergences that would lose money in a simple mean-reversion strategy.

Risks and Pitfalls: Navigating the Dangers

Statistical arbitrage isn’t a risk-free “sure thing” – far from it! Here are the major pitfalls to watch for:

False Signals and Regime Shifts

A historically reliable relationship can break when the underlying economic regime changes. Correlations that held for years might vanish if macro conditions shift.

For example, from 2014-2016, the AUD-NZD correlation dropped from its usual ~80% to almost 15% due to divergent economic factors. A strategy assuming “they always revert” would have failed during that period.

It helps to ask: Has something fundamentally changed? If one country’s economy embarks on a different path, relationships can reset to a new normal rather than mean-revert.

Whether you can predict this in advance is the million dollar question.

Correlation Breakdown Risk

Even pairs that are typically tethered can diverge far and fast during market shocks. Political events, natural disasters, or sudden shifts in sentiment can decouple currencies that usually move together.

This might be temporary, but the move against your trade can be extreme.

Stat arb is famously negatively skewed – lots of small wins and occasional blow ups.

Diversifying across multiple pair trades can help mitigate this risk. Appropriate sizing helps more.

A classic stat-arb blow-up occurs when traders keep averaging into a diverging trade (because it looks “even more mispriced”) only for the gap to widen further as the relationship breaks down. This is where a time stop can really help.

Execution Costs Matter

Stat arb often targets small pricing inefficiencies. Profits per trade might be modest (think small percentage moves). Trading costs – spreads, commissions, and slippage – can eat a large portion of those gains.

For retail traders, this means focusing on longer-term relationships (avoid overtrading like the plague).

There’s also the consideration of swap rates (overnight interest). In a pair trade, you will pay interest on one leg and receive on the other; if the differential is large, a trade held for weeks could rack up charges that offset your win.

Overfitting and Backtest Illusions

It’s easy to cherry-pick a pair and timeframe where mean reversion looked perfect in hindsight. Be wary of over-optimising your strategy on past data.

You might find one specific parameter that would have yielded steady gains every month for two years – but that could be luck.

Simple often works best: a pair with sound economic rationale and a straightforward entry/exit rule is more likely to hold up than an over-tweaked model.

Psychological Challenges

Don’t underestimate the psychological aspect. Stat arb can lull you into a false sense of safety (“the spread always comes back”). There will be times when it doesn’t, or it takes far longer than expected.

Staring at a negative P&L while “hoping” for reversion is tough. Define a maximum pain point – a time cutoff.

If your strategy shows it wins 60% of the time, accept that 40% of trades will lose. Don’t abandon the approach after a couple of losses or double down after a string of wins.

A Personal Trading Experience

I once spotted what looked like a perfect stat arb setup between AUD and NZD. After tracking the pair for months, I noticed their typical correlation was around 0.85 – extremely high. One day, after some Australia-specific economic data, AUD suddenly strengthened while NZD barely budged.

The spread between them hit 2.5 standard deviations from its mean – a clear outlier based on historical data. I confidently put on the trade: short AUD, long NZD in equal dollar amounts.

For the first few days, the trade went nowhere. Then it started going against me as the spread widened further. I checked the news – nothing fundamental had changed. My analysis suggested this was still a temporary divergence.

So I did what many traders do in this situation – I added to the position. “An even better entry price!” I told myself.

Two weeks later, I was sitting on a significant loss. The Reserve Bank of Australia had shifted to a more hawkish tone in their latest minutes, suggesting their policy path was genuinely diverging from New Zealand’s.

This wasn’t just noise – it was a fundamental shift that changed the equilibrium between the currencies. I finally closed the trade at a serious loss, having learned an expensive lesson.

Most people think the main lesson here is about regime changes – but the real lesson is in sizing and managing positions. Hindsight is wonderful, but could I have really predicted that regime change ahead of time?

Probably not.

Sensible sizing would have saved me. Your first priority as a trader is to live to fight another day.

Practical Tips for Getting Started

If you’re intrigued by statistical arbitrage but unsure how to begin, here are some practical first steps:

Start With Well-Known Pairs

Begin with the most established relationships like AUD/USD vs. NZD/USD or EUR/USD vs. GBP/USD. These pairs have substantial research behind them and tend to exhibit more reliable mean reversion.

Use Basic Tools You Already Have

Most trading platforms offer correlation tools or the ability to overlay charts. TradingView, for instance, lets you create a “spread chart” between two instruments. Excel can easily calculate means, standard deviations, and Z-scores from downloaded price data.

Do Basic Data Analysis

Try to answer questions with simple data analysis:

• How stable is the correlation relationship?
• How frequently did the spread diverge and converge?
• What was the average time to convergence in the past?

Do It Small and Get Feedback

You often see people advising beginners to paper trade. I think this is a terrible idea. You won’t have the same trading conditions in a demo account, you won’t get real market feedback, and you won’t be incentivized to manage it as seriously as you should.

Instead, start small and manage your tiny portfolio as if it were a million dollars. You’ll learn a ton.

Watch Your Costs

Calculate your all-in trading costs (spread, commission, swap rates) before entering positions. Some pairs that look attractive from a pure correlation perspective might have prohibitive costs that erode your edge.

Diversify Across Multiple Pairs

Don’t put all your eggs in one relationship. Trading multiple uncorrelated pair combinations helps smooth your equity curve and reduces the impact of any single correlation breakdown.

Final Thoughts

Statistical arbitrage in FX brings a practical, data-driven mindset to currency trading. Rather than trying to predict which way a currency will move (effectively competing with the smartest minds in finance), you’re betting on relationships returning to equilibrium – a much more tractable problem.

For a systematic trader, it’s an excellent way to develop market insight without requiring complex coding or high-frequency infrastructure. You’re forced to understand not just one currency but the interplay between multiple currencies and the factors that drive them.

The beauty of this approach is its simplicity and accessibility. With just a spreadsheet and a basic understanding of statistics, you can implement strategies that professional funds have used for decades.

Of course, stat arb isn’t a magic bullet. Markets evolve, relationships change, and no edge lasts forever without adaptation. But by focusing on sound relationships, using proper risk controls (primarily sizing), and remaining disciplined, you can put probabilities in your favor.

Remember that stat arb is about consistency over flashiness – grinding out steady gains rather than hitting home runs. It rewards the patient and the methodical – not the complicated. In a trading world often dominated by excitement and emotion, there’s something refreshingly rational about simply betting that two drunk friends will eventually find their way back to each other.