Re: Correlation caps for prop candidates: weekly audit
Posted: Fri Oct 23, 2026 6:27 am
The leg method above also catches the opposite problem, where a ticket that feels like a hedge quietly adds to a theme. Crosses are where this hides, because each one carries two non dollar legs.
Say the book holds long EURUSD, long GBPUSD, long gold and short EURGBP, each at 0.5R. The euro column nets to zero, but GBP reads +1.0R and USD reads -1.5R.
The chart shows that made up book as currency columns against a 1R cap. The EURGBP short felt like a euro trade, yet it removed the euro exposure and doubled the pound instead, so the book is really a full size long GBPUSD with gold on top.
The cooldown rule from the first post works well with columns too. After a full stop on long GBPUSD, the block would apply to any new trade that adds GBP or subtracts USD, not only to the pair itself.
One limit is that R in different pairs is not quite the same risk. A 0.5R stop on gold and a 0.5R stop on EURGBP rarely get hit by the same kind of move, so the column works as a cap rather than a forecast.
Say the book holds long EURUSD, long GBPUSD, long gold and short EURGBP, each at 0.5R. The euro column nets to zero, but GBP reads +1.0R and USD reads -1.5R.
The chart shows that made up book as currency columns against a 1R cap. The EURGBP short felt like a euro trade, yet it removed the euro exposure and doubled the pound instead, so the book is really a full size long GBPUSD with gold on top.
The cooldown rule from the first post works well with columns too. After a full stop on long GBPUSD, the block would apply to any new trade that adds GBP or subtracts USD, not only to the pair itself.
One limit is that R in different pairs is not quite the same risk. A 0.5R stop on gold and a 0.5R stop on EURGBP rarely get hit by the same kind of move, so the column works as a cap rather than a forecast.