I remember that we had a lecture in which during the inductive step, we proved P(n) by breaking the n+1 terms into two groups. The below question is marked as an incorrect approach on the practice sets. When is using such an approach (breaking n+1 things into two groups) valid?
