Maximum-likelihood degree formula for generic polygon Wachspress models

Let $A \in \mathbb{R}^{(n-2) \times n}_{\geq 0}$ be generic among matrices having a cell whose fibers are $n$-gons. Let $C \subset C_{\max}$ be a cell, choose $b \in \operatorname{int}(C)$, let $P_b$ be the corresponding fiber, and set $Q=B^T \cdot P_b$. Denote by $M_{\mathbb{C}}(Q)$ the Wachspress model of $Q$.

Wachspress ML-degree conjecture. For a generic $n$-gon $Q$, the maximum likelihood degree is

$$\operatorname{MLdeg}(M_{\mathbb{C}}(Q))=(n-1)(n-2)+(n-3)(n-5)-1.$$

The conjecture gives the predicted maximum likelihood degree for generic polygonal Wachspress models. The paper presents it as an expected formula and does not establish it unconditionally.