Kazhdan-Lusztig polynomials and subexpressions (Q2215782)

Let \((W,S)\) be a Coxeter system. To any pair of elements \((x,y)\) of \(W\), \textit{D. Kazhdan} and \textit{G. Lusztig} [Invent. Math. 53, 165--184 (1979; Zbl 0499.20035)] associated a polynomial \(h_{x,y} \in \mathbb{Z}[v]\). In their original paper, Kazhdan and Lusztig [loc. cit.] conjectured that the polynomials \(h_{x,y}\) have non-negative coefficients. Kazhdan and Lusztig's positivity conjecture [loc. cit.] was proved in general in [\textit{B. Elias} and \textit{G. Williamson}, Ann. Math. (2) 180, No. 3, 1089--1136 (2014; Zbl 1326.20005)]. The proof is via a study of Soergel bimodules associated to the underlying Coxeter system. The aim of the paper under review is to explain a strategy to use Soergel bimodules to better understand the combinatorics of Kazhdan-Lusztig polynomials. In particular, the authors refine an idea of \textit{D. D. Deodhar} [Geom. Dedicata 36, No. 1, 95--119 (1990; Zbl 0716.17015)], whose goal is a counting formula for Kazhdan-Lusztig polynomials. This is a consequence of a simple observation that one can use the solution of Soergel's conjecture to make ambiguities involved in defining certain morphisms between Soergel bimodules in characteristic zero (double leaves) disappear.