How to read a market heatmap without fooling yourself

A market heatmap answers one question extremely well: where did money move, and how much of it. Tile area encodes market capitalisation, colour encodes return over your chosen window, and the sector grouping turns thousands of instruments into a picture your visual system can read in about two seconds. That compression is the feature, and it is also every one of its limitations.

The first thing the image hides is the time window. A heatmap coloured on one day and the same heatmap coloured on twenty sessions frequently tell opposite stories, because a violent single-day reversal dominates the daily view while contributing almost nothing to the month. Always know which window you are looking at, and make a habit of flipping between a short and a long one before drawing a conclusion.

The second is that area and colour are measuring different things and the eye multiplies them anyway. A large tile with a mild colour represents far more capital movement than a small tile that is glowing. Since attention is drawn by saturation rather than by size, the most eye-catching tiles on any heatmap are usually small-cap moves that are irrelevant to where the market’s money actually went.

The third is dispersion within a sector block. A sector that reads uniformly green is a sector where a common factor moved everything, which is information about the factor and not about the companies. A sector where half the tiles are green and half red, with no size pattern, is a stock-picker’s market inside that industry. Those two pictures suggest completely different trades and look similar at a glance.

Use the map to generate questions rather than answers. A block that is red while its obvious peers are green is a question — an idiosyncratic event, an index rebalance, a single large holder unwinding. The heatmap tells you where to point the screener; it is not itself a reason to act, and it has no idea whether any given move is news or noise.

One practical habit: check the map at a five-day window before the open, then again at one day after the close. The first tells you what the week has been doing and where positioning has quietly accumulated. The second tells you what today actually changed. The difference between those two images is more useful than either one alone.