A random planar map is a graph drawn on the sphere — vertices, edges, faces — chosen uniformly at random from all maps of a given size. Specific substructures appear in these maps: faces of particular sizes, vertices of particular degrees, self-loops, patterns of connectivity. How often does a given pattern appear? And what does the distribution of its count look like?
The paper on asymptotic normality of pattern counts in random maps (arXiv: 2603.19485) proves that the count of many patterns satisfies a central limit theorem: for large maps, the pattern count is approximately Gaussian, with explicit formulas for the mean and variance.
This is not obvious. Random maps are highly correlated combinatorial structures — the presence of one pattern constrains what can appear nearby. Unlike independent coin flips, where the CLT follows from independence, pattern counts in random maps require analyzing dependencies that extend across the entire graph. The proof must show that these long-range correlations decay fast enough that the pattern count behaves like a sum of weakly dependent random variables.
The result is a distributional structure theorem for random combinatorial objects. It says that despite the complex, correlated structure of random maps, the macroscopic statistics (pattern counts) are asymptotically Gaussian. The microscopic correlations average out.
The through-claim: randomness normalizes. Random maps are intricate, with long-range geometric correlations. But the frequency of any fixed pattern across the map, when properly scaled, follows the simplest possible distribution: the Gaussian. The central limit theorem extends from sums of independent variables to counts of dependent structures embedded in random geometries. Complexity at the micro level produces universality at the macro level.
2603.19485. Probability / random maps / central limit theorem / pattern enumeration / combinatorics.