An arXiv preprint reports an exact answer for the fine-scale size of the extreme points of a random shape. Hausdorff dimension is a way to describe how a set occupies space at very small scales. For an ambient dimension d of at least 2, a stability index alpha between 0 and 1, and uniformly distributed directions, the extreme-point set has Hausdorff dimension equal to the ambient dimension minus one, multiplied by alpha, with probability one in the model.
The result is paired with a second conclusion: the Hausdorff measure taken at that same dimension is finite almost surely, both for the full image generated by the model and for the extreme-point set itself. The paper therefore gives not only a precise dimension, but also a bound on how much set remains at that critical scale.
A shape built from random arrivals
This is a theorem about a mathematical probability model, not an empirical sample. The construction starts with independent standard exponential variables. Their cumulative sums give the arrival times, while independent uniform directions on the sphere supply the model's directions; the two sequences are independent.
The weights attached to the countably many contributions decrease asymptotically like k raised to the power negative one over alpha. Because the reciprocal of alpha is greater than one, the weights are summable almost surely.
Two bounds meet
To identify the relevant set, the analysis introduces a field X that turns directions into points of the shape. On the good-direction set G, the image X(G) equals the exposed-point set. The extreme-point set is bounded between these exposed-point images and the full image of the sphere, allowing the dimension question to be studied through X.
The lower bound comes from examining how X changes along tangent directions when the angular separation is small. The characteristic exponent of those tangential increments is bounded above and below by constant multiples of the separation times the tangent-vector length raised to alpha. A negative-moment estimate controls the singularity of small increments, and the resulting energy argument establishes that the extreme-point set has dimension at least the target value.
To prove the matching upper bound, the paper covers the sphere with small cells and separates large jumps from small jumps at each scale. Each cell image can be covered by a controlled number of sets, while the diameter of each set is controlled by the variation from the small jumps. A moment estimate bounds the combined cost of the large-jump count and small-jump variation. This gives an upper bound of the same value for the full field image and, consequently, for the extreme-point set.
With both inequalities in place, the two sides meet: the exact Hausdorff dimension is the ambient dimension minus one times alpha, almost surely. The adaptive covering also yields finiteness of the Hausdorff measure at that dimension, not only for the full field image but also for the extreme-point set.
What the result leaves open
That last result has a clear limit. Finiteness does not establish that the critical measure is positive, and the supplied discussion leaves positivity unresolved in dimensions d of at least 3. The exact dimension theorem itself is stated for uniform directions, with the lower-bound argument relying on the tangential non-degeneracy available in that setting.
For the upper bound, the discussion says uniformity can be weakened to a uniform spherical-belt estimate, a condition that limits how much spectral measure can lie in a narrow belt around a direction. Under that condition, the upper-bound theorem and critical-measure finiteness extend. This is not an exact-dimension result for arbitrary non-uniform spectral measures.
Other questions remain open: for d of at least 3, the correct gauge function is not known, and the model's packing dimension and multifractal properties have not been determined.
The front matter identifies the work as arXiv:2608.28004v2 in probability, dated 3 September 2026. It is a preprint result for a specified mathematical construction and offers no empirical evidence about people, animals, or real-world populations.
Paper data and sources
Original title: Hausdorff Dimension of the Set of Extreme Points of a Random Countable Stable Zonotope
Authors: Maksim Kukushkin
Journal/Repository: arXiv
Status: Preprint, not yet peer-reviewed
First online: 2026-08-28
DOI: Not available
Original paper · Full text