Three Separate Positions, One Single Bet: Pair Correlation and Hidden Risk Concentration

What is correlation and why is it not a constant number?

Correlation is a statistical coefficient that measures how similar the price movements of two assets are to each other. Its value ranges between -1 and +1.

As the coefficient approaches +1, it means the two instruments move in the same direction to a similar extent. As it approaches -1, they move in opposite directions: as one rises, the other falls. A coefficient close to zero indicates that there is no meaningful connection between the movements of the two assets.

Up to this point is textbook knowledge. The truly important part, which most traders skip, is the second part: the correlation coefficient is not constant. A relationship that is 0.60 during one period may rise to 0.90 or drop to 0.20 when the market regime changes. The coefficient also varies depending on the period and timeframe over which you calculate it — a 20-day correlation and a 200-day correlation can show the same pair very differently.

And the most critical point is this: correlations rise especially during times of stress. When the market is calm, every pair has its own story — inflation data from Europe moves the euro, employment numbers from the UK move the pound, China's PMI moves the Australian dollar. But when a single major driver comes into play, these separate stories fall silent and everything turns in the same direction at the same time.

In other words, diversification disappears right when you need it most. On a day of sharp movement when you expect different parts of your portfolio to balance each other out, those parts stack on top of each other instead of balancing out. Diversification that seems real in a calm market turns into an illusion in a sharp market.

Common denominator logic: count currencies, not pairs

In forex, every instrument consists of two currencies. EUR/USD is not a single asset; it is a ratio between the euro and the dollar. Therefore, taking a position in a pair is actually taking two positions at the same time: long in one currency, short in the other.

Being long in EUR/USD means euro long and dollar short. When you apply this simple translation to all your open positions, the true structure of your account emerges.

Let's return to the initial example:

• –EUR/USD long = EUR long + USD short
• –GBP/USD long = GBP long + USD short
• –AUD/USD long = AUD long + USD short
• –Total: one unit long in three different currencies, three units short in the US dollar.

Real risk calculation: How does 1% quietly become 3%?

Now let's move on to the numbers. Consider a $10,000 account. The trader acts disciplined and risks 1% of their capital on each position, which is $100. They open three positions: EUR/USD long, GBP/USD long, AUD/USD long. A stop level is set for each, and the risk in each is $100.

How the account looks in their mind: three independent trades, each experiencing its own fate. Some will win, some will lose, and on average the risk will be diversified.

The reality of the account: if the dollar strengthens sharply, all three positions move into loss simultaneously. Stops are triggered not sequentially, but almost within the same minutes. Total loss is $300 — that is, 3% of the capital.

This result is largely the same as taking a direct 3% risk on a single trade. The only difference is that the trader never planned it that way. No one sat down in the morning and said, "Today I am committing 3% of my capital to a single dollar bet." That decision was made three separate times on three separate screens, each time with innocent-looking 1% steps.

Finding a catalyst to trigger this is not difficult either. US non-farm payroll data will be announced on September 4 at 12:30 GMT, and the Fed interest rate decision on September 16 at 18:00 GMT. In such events, the dollar single-handedly sets the direction, and that direction hits all three positions at once.

You might say that correlation is not exactly +1, so all three stops won't always be triggered together — that is true. But when planning, the realistic assumption is the worst-case scenario, not the average scenario. On stress days when correlation is already rising, all three stops going off together is not an exception, but the expected situation.

Gold, indices, and crypto: concentration is not only on the currency pair side

Believing that this problem is limited to currency pairs is a common misconception. Even if instruments like XAU/USD and US100 belong to different classes by name, they can become correlated with your currency positions through dollar and risk appetite channels.

Gold is already priced in dollars; being long XAU/USD is, by definition, a position against the dollar. An index position, even if it is not a direct dollar bet, can fall on the same side as high-beta currencies like AUD through general risk appetite.

This week's data provided a concrete example of this. Following the PCE data released on August 26, the dollar strengthened across the board and US Treasury yields rose. The result appeared simultaneously on three different pair charts: EUR/USD fell to a multi-day low around 1.1650; GBP/USD pulled back below 1.3600; USD/JPY turned upward. Three separate screens, three separate stories apparently, but a move driven by a single driver.

On the crypto side, the relationship is looser and less predictable. Bitcoin pulled back briefly on the same data, but then recovered and closed August 27 at $79,027, up +0.59% on a daily basis. So the macro driver created an impact there too, but the persistence of the impact was different. This demonstrates exactly why correlation is a question of "how much and when" rather than "yes or no".

The practical takeaway is this: group your portfolio by driver, not by asset class. The sentence "I have one forex, one gold, and one index position" sounds diversified. The sentence "All three are sensitive to the direction of the dollar" tells the truth.

The opposite error: Risk is not eliminated with negatively correlated positions

Some traders who notice correlation try to solve the problem in reverse: since stacking positions in the same direction increases risk, let's balance it out by opening positions in the opposite direction.

A classic example is holding EUR/USD long and USD/CHF long positions simultaneously. These two instruments historically tend to move in opposite directions because both have the dollar on one side, and the dollar is in opposing positions. EUR/USD long means dollar short; USD/CHF long means dollar long. The two positions largely cancel each other out on the dollar side.

The result is not the elimination of risk, but the doubling of cost. While net directional risk is largely neutralized, the spread for both positions is paid, and if positions are held overnight, two separate swaps accrue. In other words, the account continues to generate transaction costs without betting in any direction.

Moreover, this neutralization is not complete either. The relationship between EUR/USD and USD/CHF is not a perfect -1 and changes over time. The remaining residual risk — essentially the relationship between EUR and CHF — is a position the trader did not consciously choose. So in exchange for the double cost you pay, you actually carry a third bet without knowing what it is.

If you are not sure about the direction of a position, the right move is not to cover it with an opposite position, but to reduce or close the position. This is a less exciting, but much cheaper solution.

Tools and practical details that make calculation easier

The most laborious part of these steps seems to be splitting lot size in correlated positions. Dividing a %1 risk into three and calculating a separate lot size for each position is both slow and prone to error when done manually. The Position Calculator on the site is useful at this point: you can enter the account size, risk percentage, and stop distance to generate lot values for each leg separately. Consulting the AI Assistant when extracting the currency breakdown of your portfolio or questioning whether two instruments are linked to the same driver also speeds up this analysis. To see which news will hit which group of positions, the Economic Calendar is the most direct way to map dates to currencies.

There is a small but important detail on the implementation side: splitting a position into three depends on the minimum lot step allowed by your account. In small accounts, when you divide a %1 risk into three parts, the resulting lot value may fall below standard account steps, making splitting practically impossible. Therefore, account types that support micro lots make this method viable. On the XM side, the fact that Micro and Standard accounts can be opened with a minimum deposit of $5 and support micro-lot trading is a detail that practically enables splitting correlated positions in small capitals. For those who want to compare conditions on the broker side, the Broker Inquiry tool and Broker Rankings pages can be a starting point.

A warning is also in order: lot splitting does not turn a bad idea into a good idea. It only keeps the space that same idea occupies in your account to the extent you planned. The quality of your analysis is a separate matter; correlation management does not replace it.

Conclusion: risk is measured by the number of drivers, not the number of positions

The true risk carried by an account cannot be understood by looking at the number of open positions. If all five positions are tied to a single macro driver, that account carries one idea, not five ideas — just at five times the size.

Therefore, when looking at the position book, the question to ask should not be "how many trades do I have", but "how many different drivers am I exposed to". Events such as US non-farm payrolls on 4 Eylül, the ECB interest rate decision on 10 Eylül, and the Fed interest rate decision on 16 Eylül can tell you in advance which positions in your portfolio they will move simultaneously. Making that mapping is much cheaper than facing a surprise after the event.

Correlation risk is not a technical detail, but a matter of choosing the correct unit of measurement for risk management. Risk percentage per trade alone is an incomplete metric; what complements it is total exposure tied to the same driver. A trader who tracks these two together at least knows how much risk they are taking — which is one of the few things that can be controlled in a business where knowing the outcome is impossible.

This content is for general information purposes only and is not investment advice; leveraged trading carries high risk and you may lose all of your capital.