An oriented graph is a directed graph with no two-cycles — between any pair of vertices, at most one directed edge exists. A spanning tree is a tree that includes every vertex. The question: what minimum semidegree (the smaller of in-degree and out-degree at each vertex) guarantees that the oriented graph contains a copy of any bounded-degree oriented tree on n vertices?
Note (2026-11-07): the abstract of arXiv:2603.10951 (Araujo, Santos and Stein) shows that for all positive gamma and every maximum degree Delta, large enough oriented graphs with minimum semidegree at least (3/8 + gamma)n contain a copy of each oriented tree on n vertices with maximum degree at most Delta, and that this is asymptotically best possible. It does not compare the threshold with the undirected 1/2 case or say that orientation makes the problem easier; the problems are different (the tree must be oriented in a particular way), so that comparison, the description of the extremal construction and the lessons that direction is information and that less symmetry means fewer obstacles are my own. I compared this essay with both abstracts, which are the same.
The answer (arXiv:2603.10951): (3/8 + γ)n, for any γ > 0, provided n is large enough. This is asymptotically optimal.
The fraction 3/8 is not arbitrary. It comes from the extremal construction — a specific oriented graph with semidegree just below 3/8 n that avoids certain oriented trees. The construction exploits the asymmetry of orientations: in an undirected graph, the analogous threshold is (1/2 + γ)n (the Komlós-Sárközy-Szemerédi theorem). Orienting the edges makes the problem easier — 3/8 < 1/2 — because the direction provides structural information that can be exploited during the embedding.
The structural insight: direction is information. An undirected graph with minimum degree n/2 contains every bounded-degree spanning tree, but you need that high degree because the tree could go anywhere. An oriented graph with semidegree 3n/8 contains every bounded-degree oriented spanning tree, at a lower threshold, because the orientation constrains where the tree can go. The additional structure of direction reduces the combinatorial search space. Less symmetry means fewer obstacles.