The paper studies a question in combinatorics and group theory that is connected to a long-standing open problem: does every finite tournament (a complete directed graph, where each pair of vertices has exactly one directed edge between them) have a certain symmetry-extension property called EPPA? EPPA roughly asks whether any finite structure can be embedded into a larger finite structure where every local symmetry extends to a global symmetry of the whole object. The authors translate this combinatorial question into an algebraic one involving partial permutations, asking when partial permutations satisfying certain constraints can be extended to full permutations whose orders involve only primes from a chosen set.
The key technical contribution is reformulating these extension questions in terms of what the authors call "EP problems," which are combinatorial objects whose behavior is controlled by an associated algebraic structure called a fundamental group. The main results show that whether a solution exists depends almost entirely on this fundamental group. When the fundamental group is trivial or cyclic, solutions always exist. The authors also produce concrete examples where no solution exists, even allowing permutations whose orders are powers of a single prime, which shows that certain hoped-for extension properties genuinely fail.
The paper then focuses on a special class of EP problems organized by how many generators are needed and how the underlying combinatorial structure "wanders." Problems with a single wandering parameter always have solutions for any set of primes. For problems with two wandering parameters, the authors show odd-prime solutions always exist and give a complete characterization of exactly which ones fail to have solutions involving nilpotent groups. These results simultaneously recover and generalize earlier work by Huang, Pawliuk, Sabok, and Wise, and push the state of knowledge closer to resolving the tournament EPPA problem.