User talk:Matt.matolcsi

Hello Matt,

here is an idea to help you to understand the formula : as you wrote it , 1/n comes from the uniformaty of the distribution (le best candidate can be the 1st, the 2nd etc... with the same probability). Now the term (r-1)/(j-1) is the probability to chose the candidate number j knowing the fact that the best candidate is the one with number j : if j is the best, the event "I chose j" is exactly the same that the event "At least one of the r-1 first candidate is the best of the j-1 first of them". This because of the policy the manager fixed at the begining of the story. This event has probability (r-1)/(j-1). To finish, the sum comes from the fact than the best one is number 1, or 2, ..., or n.

Bye, Vivien.