A pitch map smooths away the one ball that decided the session

Heat maps bin, blur and average, and every one of those steps deletes the outlier that changed the match.

Look at a pitch map long enough and you stop seeing deliveries.

You see a warm blob. That blob was produced by dividing the surface into cells, counting balls per cell, then applying a smoothing kernel so the picture reads as continuous rather than as a scatter of dots. Three deliberate choices sit inside it: cell size, smoothing radius, colour scale. Change any one and the same over produces a visibly different graphic. Nobody publishes those settings.

Bin width is the sly one. Make the cells small and the map looks noisy and unusable. Make them large and two distinctly different lengths merge into one hot region, which is exactly what happens with the back of a length ball and the genuinely short one. Plenty of maps treat those identically. For a batter they are two entirely separate propositions.

The deeper issue is that the map is a density estimate, and density is the wrong summary for a contest decided by exceptions. Test sessions turn on one ball that did something different. In a density plot, one ball is a faint smudge or nothing at all. The graphic is built to show central tendency and the thing you wanted was the tail.

I am not arguing against pitch maps. They are honest about a bowler's habitual area and they are quick to read, which matters at broadcast pace. I am arguing that they get used to answer a question they cannot answer, which is why a wicket fell.

Two changes would help and neither is expensive. Overlay the wicket balls as individual marked points on top of the density, so the exceptions stay visible. And show the same over at two bin widths side by side, which immediately reveals whether a hot region is a real cluster or a smoothing artefact.

There is a small ethical edge here. Coaches use these images in review meetings and players read them as evidence. A blurred picture presented as measurement carries more authority than a scatter of raw dots, and it deserves rather less.

Raw dots are ugly and truthful. The smoothing is a design decision, made for legibility, that we have quietly promoted into a finding.

Ask for the settings next time. Watch how few people can tell you what they were.