Correlation is the position size you didn’t know you took

A trader opens a long EURUSD at 1% risk. An hour later the setup looks even better, so a second long

goes on — GBPUSD this time, at another 1%. On paper the account risks 2% across two trades. In

practice it risks close to 2% on one idea, because both currencies are being priced against the same

dollar.

That is correlation: the part of two trades that is the same trade. It is not a subtle statistical

footnote. It is the difference between a risk budget that holds and one that quietly doubles.

Why dollar pairs move together

EURUSD is mostly the euro against the dollar; GBPUSD is mostly the pound against the dollar. When the

dollar strengthens, both fall, and they tend to fall on the same days, at the same times, for the same

reason. The pound's own story matters, but when the driver is dollar-wide — a US inflation print, a

Federal Reserve press conference — the two positions behave like one bigger position in the same

direction.

The same logic links AUDUSD and NZDUSD, USDCHF and USDJPY (in the opposite direction), and pairs that

share a commodity exposure such as USDCAD and oil. Knowing which of your trades share a driver is more

useful than knowing the exact correlation number.

Measuring it without fooling yourself

Correlation is a number between -1 and +1 that describes how two series moved together over a chosen

window. Two cautions matter more than the arithmetic:

• The window decides the answer. Hourly returns over a month and daily returns over two years can

give different numbers for the same pair of instruments, because different drivers dominate at

different horizons.

• Correlation drifts. It rises in a crisis. Positions that were independent for six months can

become one bet in an afternoon, which is exactly when it hurts.

Our correlation matrix is a reading aid, not a rule: it shows how pairs

have moved together recently, so you can see the overlap before you take it.

The practical version: count the bets, not the tickets

Before the second order goes in, ask one question — is this a different idea, or the same idea with a

different symbol? If it is the same idea, the sizes add. Three correlated 1% trades are not three

trades at 1%; they are one trade at 3%.

The workable discipline is:

• Group open positions by their main driver (dollar, euro, risk appetite, a single commodity).
• Cap the risk per group, not per ticket. A common choice is the same 1–2% you would allow a single

trade.

• When a group is at its cap, a new setup in that group waits — or replaces the weakest position

instead of joining it.

What correlation cannot tell you

It cannot tell you which of the two trades will work, and it is not a forecast. It also does not

describe the risk of an uncorrelated set of positions that all stop out at once because the market

gapped — that is a different failure, the one a drawdown measures after the

fact.

What it does do is keep you honest about how much you have on. The position size you wrote down is not

the position size you are running; the second number includes everything that moves together.

Education only, never financial advice.

Frequently asked questions — Correlation is the position size you

Are EURUSD and GBPUSD always correlated?

No. They share a driver — the dollar — so they often move together when dollar-wide news is the story,

and they can diverge completely when the story is the euro or the pound. Treat the correlation as a

tendency to check, not a fixed number.

How much risk should two correlated trades carry in total?

Whatever you would allow one trade in that direction. If a single position in that group would carry

1% of equity, two correlated positions should share that 1%, not double it.

Does negative correlation cancel risk?

Only partly. Two positions that move in opposite directions reduce directional risk, but they still

carry spread, swap and the risk that the relationship breaks at the worst moment. A hedge is a

position, with its own costs.