What a Shoeprint Can Tell You

An examiner first reads what the manufacturer built in: tread pattern, size, heel shape, logo on the heel. These soles are searched for in reference databases containing thousands of patterns linked to brands and models. The field narrows from any shoe in the world to 1 model - but hundreds of thousands of pairs of that model are made.

Then come the things the factory never planned. Glass left a cut across the tread, a pebble got wedged in a groove, 1 lug chipped off on a curb. These marks appear by chance - in a random place and on a random day - so they do not match on 2 pairs of the same model. They are what support the conclusion, “that very sole,” rather than “one like it.” But this set of features does not stay fixed: as a person walks, some scratches are added and others wear away - so it matters how much time passed between the print and the seizure of the shoe.

Wear sits between these 2 levels. When you walk, your foot does not roll strictly through the center: one person rolls it outward and wears down the outer edge of the heel, while another wears down the inner side. Look at your old pair - one side is almost certainly more worn. You have seen this 100 times and never thought of it as a feature, but for an examiner, wear sits in the middle: it is not unique like a cut from glass, but it is not the same across the whole production run either.

One print tells you about the shoe. Several in a row tell you about the gait. From a trail of prints, examiners measure stride length, the width of foot placement, and the angle at which the toe points outward. For any one person, these measurements stay fairly consistent.

The same trail can show whether someone was walking or running. When running, weight usually falls on the front of the foot: the toe is pressed deeper than the heel, the stride is noticeably longer, and loose soil is thrown backward, against the direction of movement. A sharp turn leaves a smeared, twisted print. You can read this from the geometry, without a lab.

The hardest part is the final step: “this print was left by this very pair.” Matching the model and size is easy to check. But a conclusion about a single pair rests on random features, and there is no common standard: the examiner decides for herself how many matching cuts are enough. In 2016, the President's Council of Advisors on Science and Technology (PCAST) reviewed the method and concluded that there is not enough research to confirm the reliability of this conclusion.