Why correlation is a regime rather than a constant, how a book of eight instruments can carry one bet, and what to measure instead of counting tickers.
Portfolio intelligence is the step between a market view and what you actually own. It asks whether the positions in a book are as diversified as they look, and what single condition would hurt all of them at once.
The most expensive mistake in portfolio construction is treating a correlation as a property of two assets. It is a property of the regime. Two holdings that have spent a year moving independently can converge to one position in a week of stress — and that week is exactly when diversification was supposed to help.
We have measured this directly. Bitcoin followed gold closely in calm markets — a 60-day correlation of 0.58, a five-year high — and the link vanished when it mattered: on 158 stress days, gold moved −0.12 % while Bitcoin moved −1.94 %. The correlation was real and the diversification was not.
A book can hold eight instruments and carry one bet. In our own record, three crypto positions turned out to be the same wager — Bitcoin direction expressed through different tickers. The position list showed three lines; the risk was one.
The test is not how many names are held but how many independent things can go wrong. Apex looks at overlapping drivers rather than counting tickers: which positions share a regime sensitivity, which share a funding condition, and which would be hit by the same headline.
When we measured nine instruments across four axes — behaviour against equities, against gold, against the dollar and against rates — almost everything moved together. The whole matrix sat at the 91st percentile of its own history, which is itself the finding: in that regime there was very little genuine diversification available anywhere.
The column that mattered was the control one, behaviour against equities. Without it, a rise in the gold correlation looks like an asset changing character when it is really the regime changing underneath every asset at once. The only holding that had genuinely decoupled was energy — and it decoupled from everything, including the things it is usually explained by.
What single condition hurts everything I hold? Not the worst case for each position separately, but the one event that reaches all of them. In a book with high internal correlation, that condition exists whether or not it has been named.
Is this correlation from the current regime or the average? A twelve-month average hides exactly the weeks the number is needed for. Apex reads correlation in the regime that is live now, and shows where that differs from the longer window.
Which positions share a driver rather than a sector? Two energy names share a sector. An oil position and an inflation-sensitive equity share a driver. The second pair is the more dangerous concentration because it does not appear on any sector breakdown.
Exposure by instrument and by driver, correlation in the current regime rather than the average of the last year, regime sensitivity, overlapping positions, and what the engine itself is holding right now — including the trades that went against it.
It is not portfolio management and not advice. Apex holds no client capital, gives no allocation recommendation, and trades only its own paper account so the method can be tested forward in public.
All numbers on this page were published in Apex’s own record at the time; they describe the past under stated conditions and promise nothing about the future. Research · Learn · The Apex day