Let the events $E_1, E_2, E_3, ... E_n$ be independent. $P\{E_K\} = p_k$. Find, $p$, the probability that none of the events occur.

The answer for this in the book is: $p = (1-p_1)(1-p_2)(1-p_3)...(1-p_n)$

Why is the answer not: $p = 1-(p_1)(p_2)(p_3)...(p_n)$

  • 1
    $\begingroup$ The product of the $p_i's$ is the probability that all of the events occur so your formula computes the probability that at least one of the events does not occur. $\endgroup$ – lulu Sep 18 '17 at 23:03
  • 1
    $\begingroup$ if one of the $p_i=0$ your p is going to be 1 in your formula it's absurd because of one of the event, all independant, it gives a certitude on p $\endgroup$ – Isham Sep 18 '17 at 23:09

Because $$(p_1)(p_2)(p_3)\cdots(p_n)=P(E_1\cap E_2 \cap E_3 \cap \cdots \cap E_n)$$

and the complement of your event space “none of the events occur” is not “all of the events occur,” so you cannot use $P(A')=(p_1)(p_2)(p_3)\cdots(p_n)$ in the formula $P(A)=1-P(A')$.

The actual complement of your event space is “at least one of the events occurs.”

Your event $E$, “none of the events occur,” is equivalent to “$E_1$ does not occur and $E_2$ does not occur and $E_3$ does not occur and . . . and $E_n$ does not occur.”

Using $P(E_k')=1-p_k$, we have

$$\begin{align} P(E) &= P(E_1' \cap E_2' \cap E_3' \cap \cdots \cap E_n') \\ &= P(E_1')\,P(E_2')\,P(E_3')\cdots P(E_n') \\ &= (1-p_1)(1-p_2)(1-p_3)\cdots(1-p_n) \\ \end{align}$$


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.