Correlation-adjusted risk during the London–NY overlap
Posted: Sat Sep 05, 2026 3:14 pm
Correlation-adjusted risk is the part of session planning I used to skip until a green EURUSD scalp and a green GBPUSD scalp turned into one red USD day. Same dollar idea, two tickets, double the drawdown when USD flipped. During London–NY overlap I now treat shared USD exposure as one risk bucket — not three independent opportunities.
The problem is simple. EURUSD, GBPUSD, and XAUUSD often move as expressions of the same USD impulse in the overlap. If I risk 0.3% on each because each chart looks clean, I am not risking 0.3%. I am risking roughly a full percent on one narrative, with fills and spreads stacked on top. When the idea works, the equity curve looks brilliant. When it fails, the daily soft stop arrives in one correlated wave.
How I run it now:
1. Name the idea before the first click. Example: “USD weak into NY open” or “USD reclaim after London fade.” If I cannot name it in one line, I do not stack pairs.
2. Cap aggregate risk on that idea. Typical cap for me on overlap: about 0.4%–0.5% total open risk across correlated USD expressions — not 0.4% per chart. Challenge accounts sit tighter.
3. Pick one primary expression. Usually the cleanest structure plus the best spread and liquidity for my feed. That might be XAUUSD on a reclaim, or EURUSD on a session level — not both at full size. A second pair, if any, is flat or token size inside the aggregate cap.
4. Treat GBPUSD as correlated to EURUSD until structure proves otherwise. Same for gold versus USD majors when the move is clearly dollar-led rather than metal-specific.
5. If I am already in one USD expression and a second chart prints an A+ trigger, I reduce the first, skip the second, or accept concentration — I do not pretend they are hedges.
What I refuse during overlap: opening EURUSD and GBPUSD long at full risk because both broke resistance; adding gold on the same USD-weak story without cutting major risk first; counting stop distance in pips per pair while ignoring that one USD spike can tag all three stops.
Desk habit that keeps this honest: a one-line pre-overlap note — “USD idea; primary pair; aggregate risk max; second ticket: none or token.” Blank note means no second correlated ticket.
This is not about trading fewer ideas forever. It is about not paying tuition twice for the same wrong read. Expectancy after costs improves when losers are single-expression failures, not cascade failures dressed up as diversification.
Correlation is not a constant. On some days gold decouples, or cable has UK-specific flow. I still start from shared USD risk in the overlap and only relax the cap when structure and catalyst clearly diverge. Default is one expression. Stacking needs a written reason — not a second chart that also looks good.
If your journal shows clusters of same-direction majors and gold stopped out within minutes of each other, you do not need a new indicator. You need an aggregate risk line and the discipline to pick one chart for the USD idea.
The problem is simple. EURUSD, GBPUSD, and XAUUSD often move as expressions of the same USD impulse in the overlap. If I risk 0.3% on each because each chart looks clean, I am not risking 0.3%. I am risking roughly a full percent on one narrative, with fills and spreads stacked on top. When the idea works, the equity curve looks brilliant. When it fails, the daily soft stop arrives in one correlated wave.
How I run it now:
1. Name the idea before the first click. Example: “USD weak into NY open” or “USD reclaim after London fade.” If I cannot name it in one line, I do not stack pairs.
2. Cap aggregate risk on that idea. Typical cap for me on overlap: about 0.4%–0.5% total open risk across correlated USD expressions — not 0.4% per chart. Challenge accounts sit tighter.
3. Pick one primary expression. Usually the cleanest structure plus the best spread and liquidity for my feed. That might be XAUUSD on a reclaim, or EURUSD on a session level — not both at full size. A second pair, if any, is flat or token size inside the aggregate cap.
4. Treat GBPUSD as correlated to EURUSD until structure proves otherwise. Same for gold versus USD majors when the move is clearly dollar-led rather than metal-specific.
5. If I am already in one USD expression and a second chart prints an A+ trigger, I reduce the first, skip the second, or accept concentration — I do not pretend they are hedges.
What I refuse during overlap: opening EURUSD and GBPUSD long at full risk because both broke resistance; adding gold on the same USD-weak story without cutting major risk first; counting stop distance in pips per pair while ignoring that one USD spike can tag all three stops.
Desk habit that keeps this honest: a one-line pre-overlap note — “USD idea; primary pair; aggregate risk max; second ticket: none or token.” Blank note means no second correlated ticket.
This is not about trading fewer ideas forever. It is about not paying tuition twice for the same wrong read. Expectancy after costs improves when losers are single-expression failures, not cascade failures dressed up as diversification.
Correlation is not a constant. On some days gold decouples, or cable has UK-specific flow. I still start from shared USD risk in the overlap and only relax the cap when structure and catalyst clearly diverge. Default is one expression. Stacking needs a written reason — not a second chart that also looks good.
If your journal shows clusters of same-direction majors and gold stopped out within minutes of each other, you do not need a new indicator. You need an aggregate risk line and the discipline to pick one chart for the USD idea.