The topological string is a mathematical framework for counting curves on geometric spaces, organized by the complexity (genus) of those curves. At each level of complexity, the counting functions can have problematic singularities near special points in the geometry called conifold points. The "conifold gap" condition is a powerful constraint saying that most of those potential singularities must actually vanish, leaving only the mildest possible singular behavior. Proving this condition for a given geometry is hard because the cancellations are not automatic and require genuinely new mathematical input at every level of complexity.
The paper establishes the conifold gap for a specific but important geometric setup: counting curves in the projective plane that have controlled tangency with a fixed cubic curve (an ellipse-like shape). This is called a relative Gromov-Witten theory. The authors connect this relative counting problem to a simpler local counting problem (where the cubic boundary is essentially removed) using a known correspondence between the two theories. The key new tool they introduce is an algebraic identity, called a central difference identity, which reorganizes complicated combinatorial sums into a form where regularity near the conifold point becomes transparent. The crucial objects are "completed cycle traces," which are generating functions with a specific elliptic (doubly periodic) structure that makes them especially well-behaved.
The main payoff is a proof that the relative theory inherits the gap condition from the local theory through a careful inductive argument that moves across increasing levels of curve complexity. Along the way, the authors also extract additional concrete results: explicit finite formulas for the Taylor coefficients of the counting functions, identities relating the finite parts of these functions at different genera, and a systematic reconstruction of the so-called holomorphic ambiguities, which are the free parameters that must normally be fixed by hand using physical or geometric intuition. The argument is algebraically rigorous at every fixed genus and should be applicable to other geometric settings with similar structure.