1
$\begingroup$

Using the inclusion exclusion principle - http://www.proofwiki.org/wiki/Inclusion-Exclusion_Principle - if I set $n=2$ I get the following -

$$P(A_1 \cup A_2) = P(A_1) + P(A_2) - P(A_1 \cap A_2) + P(A_1 \cap A_2)$$

when the correct answer should be -

$$P(A_1 \cup A_2) = P(A_1) + P(A_2) - P(A_1 \cap A_2)$$

I have the term $P(A_1 \cap A_2)$ at the end the first equation due to the last part of the inclusion exclusion principle - $$(-1)^{n-1}P(\cap_i^n A_i)$$

It seems that I shouldn't be including that if I want to have the correct answer...but surely I have to include it as I can't just drop an arbitrary term from some formula...so what am I missing?

|cite|improve this question
$\endgroup$
  • $\begingroup$ The set $A$ is not in the union on the left, so should not appear on the right side. $\endgroup$ – coffeemath Nov 30 '12 at 12:41
2
$\begingroup$

In both of the places where you wrote $A$ you probably mean $A_1$; I’m going to assume so.

When $n=2$ the last term is

$$(-1)^{n-1}P\left(\bigcap_{i=1}^nA_i\right)=(-1)^1P\left(\bigcap_{i=1}^2A_i\right)=-P(A_1\cap A_2)\;,$$

just as it should be. Somehow you added an extra term that is not present in the expression on the cited Proof Wiki page.

|cite|improve this answer
$\endgroup$
  • $\begingroup$ Oops, well then I would have $$P(A_1 \cup A_2) = P(A_1) + P(A_2) - P(A_1 \cap A_2) - P(A_1 \cap A_2)$$. If the last term is correct then it is the second last term in my expression that is incorrect. I am getting that term $- P(A_1 \cap A_2)$ from the second term of the inclusion exclusion principle $\sum_{i<j}^n P(A_i \cap A_j)$. Have I made a mistake by including this term? $\endgroup$ – csss Nov 30 '12 at 12:48
  • $\begingroup$ @csss: No, you still have one term too many: you’ve written the last term twice. $\endgroup$ – Brian M. Scott Nov 30 '12 at 13:03
  • $\begingroup$ I have wrote it twice because when I let $n=2$ that is what the inclusion exclusion formula gives me. I get $-P(A_1 \cap A_2)$ twice. Once from the 2nd term of the IE formula and once from the last term of the IE formula. How come you don't get two $-P(A_1 \cap A_2)$ terms? $\endgroup$ – csss Nov 30 '12 at 13:15
  • 1
    $\begingroup$ @csss: You’re not reading the formula correctly. When $n=2$ the only terms are $\sum_{i=1}^2P(A_i)$ and $-\sum_{1\le i<j\le 2}P(A_i\cap A_j)$. That second term is the term $(-1)^{2-1}P\left(\bigcap_{i=1}^2A_i\right)$. $\endgroup$ – Brian M. Scott Nov 30 '12 at 13:24

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.