There are pairs of seemingly unrelated spaces such that Schubert calculus on the two spaces `match’ in some concrete sense. This duality, called 3d mirror symmetry, is best observed in equivariant elliptic cohomology. For example, the dual of the 12-dimensional $T^*Gr(2,5)$ is a certain 4-dimensional Nakajima quiver variety. However, the set of cotangent bundles of homogeneous spaces, or even the set of quiver varieties are not closed for 3d mirror symmetry. In this talk, based on a joint work with Y. Shou, we present a larger pool of spaces: Cherkis bow varieties. Superstring theory predicts that bow varieties are closed for 3d mirror symmetry. The combinatorics necessary to play Schubert calculus on bow varieties includes binary contingency tables and tie diagrams. The existence of an operation (called Hanany-Witten transition) gives bow varieties extra flexibility. We will illustrate the combinatorics and geometry of bow varieties with examples, and we will calculate cohomological and elliptic Schubert classes (rather, `stable envelopes’) to explain the `matching Schubert calculus’ phenomenon.