“The algorithm we study is very similar, except we choose a random spanning tree based on the standard linear programming relaxation of TSP.” That’s so subtle

“The algorithm we study is very similar, except we choose a random spanning tree based on the standard linear programming relaxation of TSP.” That’s so subtle to suggest but incredible hard to prove https://arxiv.org/abs/2007.01409 https://arxiv.org/abs/2007.01409