Purpose
Correlation Map shows how closely a group of names has actually moved together over a window you choose.
It exists to answer one uncomfortable question about a book: how many positions do you really have? Five names that correlate at 0.9 to each other are not five positions. They size like one, they draw down like one, and they will all be wrong on the same morning.
Best used for:
- Checking concentration before adding to a theme, rather than after.
- Finding genuine diversifiers: a candidate that correlates weakly to what you already hold.
- Pairing: two names that normally track each other and have temporarily come apart.
- Understanding why a book that looked spread out moved as a single unit.
What you see on the chart
A matrix of the names you select, each cell carrying the correlation between that row and that column over the chosen window.
- Near 1: the two have moved almost identically.
- Near 0: no reliable relationship over the window.
- Negative: they have moved opposite one another, which is what a genuine hedge looks like.
The diagonal is always 1, because every name correlates perfectly with itself.
How to read it
- Read the clusters, not the individual cells. A block of names all high to each other is one exposure wearing several tickers.
- High correlation is a sizing problem, not a reason to sell. If three holdings cluster, size them as one position rather than three, or accept the concentration knowingly.
- A low or negative cell is where diversification actually lives. That is the only place in the matrix where adding risk genuinely spreads it.
- The window changes the answer, and it should. A short window shows the current regime; a long one shows the structural relationship. Disagreement between them is itself informative: a pair that is normally uncorrelated and is currently at 0.9 is in a risk-on or risk-off regime rather than trading on fundamentals.
- Expect correlations to rise in a selloff. Almost everything correlates toward 1 when markets fall, which is exactly when a diversified-looking book stops behaving like one. A calm-market reading understates how a book will behave in stress.
How the engine works
Each pair's correlation is computed from their returns over the selected window, so the figure describes co-movement in percentage terms rather than in price. The window is yours to set, and the whole matrix recomputes when you change it.
Honest limitation: correlation is backward looking by construction. It describes the window you picked and nothing else, and it says nothing about cause. Two names can correlate because they are in the same business, because they share an index, or because the same flows happen to be moving both, and the number is identical in all three cases.
Key settings
- Window: the lookback the correlations are measured over.
- Names: the set of symbols in the matrix.
Availability
Works across asset classes, including crypto. Not available on pairs. Included from the Pro plan up.
Correlation Map is an AI Trading Desk indicator. See it on a live chart in the AI Trading Desk, or browseevery guide.
Educational and analysis purposes only. Not financial advice. See theterms.