{"id":"arxiv/2609.10852","url":"https://arxiv.org/abs/2609.10852","tags":["math.FA","math.MG"],"score":0,"title":"Zonoids whose polars are zonoids: the Banach--Mazur distance need not tend to one","domain":"math.FA","source":"arxiv","authors":["Dmitry Ryabogin","Artem Zvavitch"],"reviews":0,"subject":"math.FA","summary":"For every $n\\geq2$, we consider a Gaussian zonoid of revolution $Z_n$ arising from works of Vitale and Mathis. We prove that $Z_n^*$ is also a zonoid and compute the Banach--Mazur distance from $Z_n$ to the Euclidean ball. This distance is independent of the dimension and is approximately $1.10$. Consequently, the supremal Banach--Mazur distance among zonoids whose polars are zonoids does not converge to $1$. The construction also produces a separable real Banach space $X$, not isometric to a Hilbert space, such that both $X$ and $X^*$ embed linearly isometrically into $L_1$.","category":"research","comments":0,"disputed":false,"division":"math","mentions":0,"discussed":0,"published":"2026-09-09","reactions":{}}