A correlation coefficient is a summary of one window of history. It is not a property of two instruments. That sentence is the whole argument, and almost every expensive surprise in a "diversified" book comes from forgetting it.
The three choices hidden in every number
Any correlation figure encodes three decisions, and a figure quoted without them is not interpretable:
- What is being correlated. Log returns, not price levels. Correlating levels finds shared trend, which is why unrelated instruments in a bull market look related.
- The sampling interval. Two instruments that appear uncorrelated on daily data can be tightly linked during a single overlapping session.
- The window. A 20-day and a 200-day reading answer different questions. Neither is the correct one.
Raptor shows all three wherever a coefficient appears. Not for rigour's sake — because a trader who can see the window can tell when the number is about to stop applying.
Three ways a relationship dies
Regime change. A pairing that held through a quiet period frequently inverts when volatility expands. This is the common case, and it is the worst one, because it inverts precisely when position sizes are largest.
Session mismatch. The relationship exists for four hours a day and the daily number averages it into nothing — or the reverse, where a daily reading implies a link that only ever existed during the London overlap.
Common factor removal. Two instruments correlated because both responded to a third thing. The third thing stops moving and the relationship evaporates without either instrument changing behaviour at all.
Stability is the more useful number
A coefficient of 0.8 that has sat between 0.7 and 0.9 all quarter is a different object from a coefficient of 0.8 that was −0.2 six weeks ago. Both display as 0.8.
So the reading that changes decisions is not the coefficient — it is the coefficient plus its behaviour across sub-windows. Where a relationship has broken down, the platform says it has broken down rather than averaging the break away. The intelligence layer weights an unstable relationship down for exactly this reason.
What it means for a book
Positions spread across eight instruments can be one position by exposure. The practical test is not how many symbols you hold; it is what happens to the book under a single move. Group the correlated instruments, then look at the concentration.
And then remember that the grouping was computed over a window, and check the window.
What this does not tell you
Nothing here predicts anything. Correlation describes how two instruments have moved together over a chosen period. It is not causation, a stable reading is not a promise, and a relationship that held last month may not hold today.
Related: Market Intelligence · Risk Intelligence
Trading leveraged products carries a high level of risk to capital. Technology can improve analysis and controls; it cannot remove market risk.