A plane triangulation is a way of drawing a network of points (vertices) on a flat surface so that every region created by the connecting lines is a triangle. The question studied here is about coloring the vertices of such a network with two colors, say red and blue, in a way that satisfies a particular balancing property called "polychromatic." The challenge is to understand how unequal the two color groups can be forced to become, a quantity called the discrepancy.
The paper resolves a long-standing conjecture by Asayama and Matsumoto by proving that for any plane triangulation with n vertices, you can always find such a coloring where the difference in size between the red group and the blue group is at most roughly one-third of n. The proof is not built from scratch but instead cleverly combines two recent results from 2026 to get the conjecture as a consequence. For most values of n, the authors also establish a slightly tighter bound than what the conjecture required, meaning the coloring can be made even more balanced than previously hoped.
To support and verify the theoretical results, the authors computationally checked all distinct plane triangulations with up to 12 vertices, a collection of over nine thousand networks. This exhaustive check confirms that the bounds hold in every case and also validates the sharper estimate for n equal to 11, which belongs to the one remaining category of n values where the tightest possible bound is not yet fully settled theoretically. Together, the theoretical proof and the computational census give a thorough and reliable picture of how balanced these colorings can be made.