The unit ball in n-dimensional space looks different depending on which norm you use to measure distances. For the so-called p-norm, where p is a number between 1 and infinity, the unit ball changes shape from a cross-polytope (like a multi-dimensional diamond) when p equals 1, to a standard round ball when p equals 2, to a hypercube when p goes to infinity. A natural geometric question is: when you project such a ball onto a lower-dimensional subspace, specifically a flat hyperplane passing through the origin, how does the volume of that shadow behave, and which directions give the largest or smallest shadow?
It has been known for some time that certain special projection directions are extremal, meaning they give either the maximum or minimum projected volume. The question this paper addresses is one of stability: if a projection direction gives a volume that is close to the extremal value, must that direction itself be close to one of the known extremal directions? In other words, does nearly achieving the extreme volume force the geometry to be nearly extremal? This kind of result is called a stability estimate, and it gives quantitative control over how much the geometry can deviate while the volume stays near its extreme value.
The authors prove that the answer is yes, and they make this precise with explicit bounds. Their work is the dual counterpart of a known stability result about sections, which are slices through the ball rather than projections of it. Sections and projections are related by a classical duality in geometry, but the techniques needed for each are different. By establishing the stability result on the projection side, the paper completes a natural symmetry in the theory and provides tools that could be useful in convex geometry, functional analysis, and related areas where understanding the shape of high-dimensional bodies is important.