For 2a, my inequalities seem to match those of the solutions. For 2b, I think my proof does argue what the grader took a point off for: I say that since we're given that the Nash equilibrium must exist, a solver that loops over all subsets must find it.