The hat-check problem goes something like this: there are N letters that correspond to N envelopes. They get put in randomly. What is the probability that none of them are in the correct envelope (and also, the probability that at least one is in the correct envelope)?
I understand that the approach is the derangements approach, but I initially tried it in the following way and don't fully understand the flaw on why you can't calculate P(all letters are in wrong envelopes) directly.
My approach was the following. The first letter has probability N-1/N of being misplaced. Then, the second envelope has probability N-2/N-1. This continues until everything cancels, leaving P(all letters are in wrong envelopes) = 1/N and hence P(at least one letter is correct) = 1 - 1/N. However, this simply isn't right. I understand the formula and derivation for derangements, but what is conceptually wrong with this approach?
My thoughts are that the error could come from only considering one particular sequence, but I'm really not too sure.