A juggling pattern is a closed walk in a directed graph. Each node represents a state — which hands hold balls, which are in the air — and each edge represents a throw. The pattern repeats, so the walk must close.
Some patterns decompose into shorter patterns concatenated together. Others do not — these are the primes. A prime juggling pattern cannot be broken into two shorter valid patterns. It is irreducible in the same way a prime number cannot be factored.
The question is: how many prime juggling patterns exist for a given number of balls and period? Butler et al. (arXiv:2603.17284) develop a unified framework treating standard, multiplex (multiple balls caught simultaneously), colored, and passing (multiple jugglers) patterns through the same state-graph formalism. They establish lower bounds on prime pattern counts and exact enumerations for several cases.
The key structural insight: the state graph is infinite, but its strongly connected components have tractable structure. Prime patterns correspond to walks that cannot be decomposed into shorter closed walks — the same relationship that prime cycles have to the zeta function of a graph. Counting primes is harder than counting all patterns for the same reason that counting prime numbers is harder than counting integers: the decomposition structure imposes global constraints that local counting misses.
The framework unifies what looked like separate problems. A two-juggler passing pattern, a multiplex pattern where multiple balls land simultaneously, and a colored pattern where balls are distinguishable all live in different state graphs but share the same prime decomposition theory. The mathematics doesn't care what the physical moves are — only the graph structure matters.