Mistake: The inductive step fails from n=1 to n=2. If we assume S=(c1, c2..., ck, ck+1) is a set of k+1 cpus, then when k=2, S=(c1, c2). If we take S-c1 then this is a set of k cpus and we get S-c1 = (c2) should be a set of k cpus (specifically k=1), but we already showed k=1 is P(1) = (c1). So...