Correlated positions are one trade

Sizing each trade at 1% is necessary. It is not sufficient, because five 1% positions that all move together are a single 5% position wearing five names. This lesson is about seeing through the names.

What correlation means here

Correlation measures how much two things move together, on a scale from −1 (always opposite) through 0 (unrelated) to +1 (always together). You do not need to compute it precisely; you need to know which of your positions are, in practice, the same bet.

Some typical relationships (they drift over time, so check rather than assume):

Pair	Typical correlation	What it means
ES and NQ	strongly positive, often 0.85 to 0.95	Long both is one long-equities bet
A large-cap tech stock and NQ	strongly positive	Adding it to an NQ long adds size, not diversity
BTC and ETH	strongly positive, often 0.7 to 0.9	Long both is one long-crypto bet
BTC and most altcoins	positive, often higher in sell-offs	Alts usually fall harder when BTC falls
EUR/USD and USD/CHF	strongly negative, often −0.8 or lower	Long EUR/USD and short USD/CHF are both short-dollar
EUR/USD and GBP/USD	positive, often 0.6 to 0.8	Long both is mostly one short-dollar bet
Gold and silver	positive, often 0.7 to 0.8
Crude oil and energy stocks	positive

The dangerous ones are the pairs that look different on a watchlist and behave the same in a sell-off, because sell-offs are exactly when correlations rise.

Key idea: Group your positions by the underlying bet they express (long equities, short dollar, long crypto), and count risk per bet, not per ticker.

Adding up correlated risk

Suppose you hold three positions, each risking 1% of the account to its stop. How much are you really risking on the day the theme turns?

For n positions each risking r, with the same pairwise correlation ρ, a reasonable estimate of the combined risk is:

combined ≈ r × √(n + n × (n − 1) × ρ)

Correlation between the three	Combined risk (each at 1%)
0 (unrelated)	1% × √3 = 1.7%
0.5	1% × √(3 + 3) = 2.4%
0.9	1% × √(3 + 5.4) = 2.9%
1.0 (identical)	1% × √9 = 3.0%

Three unrelated 1% positions behave like a single 1.7% position. Three positions at 0.9 correlation behave like a 2.9% position, almost exactly as if you had put 3% on one trade. Since the day they all stop out is the day correlation spikes towards 1, the honest number to use for a themed group is the plain sum.

Worked example: the "diversified" book

A trader has:

Long MES, risk 1%.
Long MNQ, risk 1%.
Long a large-cap semiconductor stock, risk 1%.
Long BTC, risk 1%.
Long ETH, risk 1%.

Five positions, five markets, "only 1% each". Group by theme:

Long risk assets / equities: positions 1, 2, 3. Sum = 3%.
Long crypto: positions 4, 5. Sum = 2%.
And crypto has itself traded with a positive correlation to equities for long stretches, so on a bad risk-off day the honest total is closer to 5%.

That is a 5% bet on "risk assets go up today" made by someone who thinks they are running a diversified book at 1% per trade. One bad session, all five stops hit, 5% gone: half of the drawdown budget many plans allow for a whole month.

Rules that fix it

Define your themes. Write down the four to six underlying bets you actually take (long equities, short equities, long dollar, short dollar, long crypto, long energy, and so on).
Cap risk per theme. A common cap is 2% to 3% of the account across all positions in the same theme, regardless of how many names.
Treat a hedge as a hedge only if it is real. Long EUR/USD plus short GBP/USD is partially hedged; long EUR/USD plus long GBP/USD is doubled up.
Watch for hidden themes. A long in an airline and a short in crude oil are both "oil goes down" bets. Long gold and long a gold miner is one bet with two tickers.
Assume correlation goes to 1 in a crisis. That is when the theme cap does its job.

Why it matters more with leverage

Capital efficiency (Module 4) is what makes this easy to do by accident. With forex margin at 30:1 you can hold six major pairs on a small account without the platform complaining. If five of them are short-dollar, a single US data release moves all six at once. The platform's margin check did not stop you because it counts margin, not themes.

Try it: List every position you hold right now (or held last week). Next to each, write the theme it expresses. Sum the risk per theme. If any theme is above 3% of the account, you have found a position you did not know you had.

Recap

Correlated positions are one bet with several names; count risk per theme, not per ticker.
Three unrelated 1% positions behave like about 1.7%; three highly correlated ones behave like 3%.
Correlations rise in sell-offs, so the honest risk for a themed group is the plain sum.
Cap total risk per theme (2% to 3%) and define your themes in writing.
Margin checks count deposits, not themes; leverage makes accidental concentration easy.