On the twelve-point theorem for \(\ell\)-reflexive polygons (Q2335699)

Given a lattice \(N\subseteq \mathbb{R}^2\) and a lattice \(N\)-polygon \(Q\subseteq \mathbb{R}^2\), assuming that \(Q\) contains the origin \(0\in\mathbb{R}^2\) in its interior and that \(Q\) has primitive vertices, that is, vertices \(\alpha\) such that the line segment through the origin \(0\) and \(\alpha\) meets \(N\) in \(\{0,\alpha\}\), the index \(\ell\) of the polygon \(Q\) is the least common multiple of the integral distances \(\ell_F\) of all the edges \(F\) of \(Q\) to the origin. Let \(Q^{\circ}=\{x\in\mathbb{R}^2: \langle x,y\rangle\geq -1 \text{ for all } y\in Q\}\) denote the polar polygon, where \(\langle\, ,\,\rangle\) is the usual Euclidean inner product of \(\mathbb{R}^2\). The polygon \(Q\) is reflexive if its dual \(Q^{\circ}\) is a lattice \(M\)-polygon, where \(M\) is the dual lattice \(M=\Hom_{\mathbb{Z}}(N,\mathbb{Z})\). Then, \(M\) is a reflexive polygon if and only if its index \(\ell=1\) or equivalently, iff \(\ell_F=1\) for all the edges \(F\) of \(Q\). If \(Q\) is a reflexive \(N\)-polygon, its dual \(Q^{\circ}\) is a reflexive \(M\)-polygon [\textit{V. V. Batyrev}, J. Algebr. Geom. 3, No. 3, 493--535 (1994; Zbl 0829.14023)]. If \(Q\) is a reflexive \(N\)-polygon, the number of \(N\)-lattice points in the boundary \(\beta Q\) of \(Q\) plus the number of \(M\)-lattice points in the boundary \(\beta Q^{\circ}\) of its polar polygon \(Q^{\circ}\) is always \(12\). Assuming this, Batyrev [loc. cit.] gave an exhaustive and detailed classification of all reflexive polygons: They are \(16\) equivalence classes of their standard model and each one has at most \(6\) vertices. In [Electron. J. Comb. 19, No. 3, Research Paper P9, 18 p. (2012; Zbl 1258.52008)], \textit{A. M. Kasprzyk} and \textit{B. Nill} introduced a generalization of the notion of reflexivity: A lattice \(N\)-polygon \(Q\) is \(\ell\)-reflexive iff \(\ell=\ell_F\) for all edges of \(F\) of \(Q\). For this new notion, the adequate generalization of the polar polygon is its dual \(Q^*\) defined as the \(\ell\)-dilation \(Q^*=\ell Q^{\circ}\subseteq\mathbb{R}^2\) of its polar polygon; thus, for \(\ell=1\) the polar \(Q^{\circ}=Q^*\). It follows that if \(Q\) is an \(\ell\)-reflexive polygon, then its dual \(Q^*\) is also an \(\ell\)-reflexive \(M\)-polygon. The main result of Kasprzyk and Nill [loc. cit.] is a generalization of the \(12\)-point theorem to the case of \(\ell\)-reflexive polygons proving, with a purely combinatorial argument, that \(\#(\beta Q\cap N) + \#(\beta Q^*\cap M)=12\). The main result of the paper under review is a new proof of this theorem, showing its relation to toric geometry as in the \(\ell=1\) case, where from the work of Batyrev on mirror symmetry in toric surfaces, the number \(12\) is obtained from Max Noether's formula for the Euler-Poincaré characteristic of the structure sheaf of the minimal desingularized compact toric surface associated to the reflexive polygon \(Q\). The new proof of the \(12\)-point theorem in the higher index case requires translating everything into the language of toric geometry and log del Pezzo surfaces. Once this is done, the appearance of the number \(12\) is again explained by M. Noether formula. There are several byproducts of the author's algebraic geometry approach, including an interpretation of the duality notion for higher index reflexive polygons in the context of toric log del Pezzo surfaces, allowing the author to give an algebraic geometry interpretation of the lattice change used in the original argument of Kasprzyk and Nill [loc. cit.].