0
$\begingroup$

We have a complete set of events $\{A,B,C,D\}$.

$p(A)=0.2$
$p(B)=0.3$
$p(C)=0.4$
$p(D)=0.1$

Two events happen in succession, where each event cannot occur twice. (i.e. when one happens the others' probabilities re-scale to sum up to $1$).
So we have a new complete set of events $\{AB,AC,AD,BA,BC,BD,CA,CB,CD,DA,DB,DC\}$

Since they are all mutually exclusive, the probability of, say, $B$ happening is the sum of probabilities of all events that include $B$.
If I calculate correctly, that is
$p(B)=0.608$.
Also, for $C$, it's
$p(C)=0.715$.

So if I want to know the probability of either $B$ or $C$ happening, or both, I'd sum up all the probabilities of events including $B$ or $C$ or both.
And that's $p(B+C)=0.952$.

That feels correct, but there's this formula that's valid for all events, that says $p(B+C)=1-(1-p(B))(1-p(C))$, i.e. the probability of at least one of them happening is one minus the probability of neither of them happening. But that formula yields $p(B+C)=0.888\neq 0.952$.

What am I doing wrong?

And another question - is there a way I can correctly calculate $p(B+C)$ by knowing only $p(B)=0.608$ and $p(C)=0.715$, without iterating over the whole set just to calculate $p(B+C)$ (or any other such composed event, e.g. $A+D$)?

$\endgroup$
0
$\begingroup$

The formula that gives you .888 assumes $B$ and $C$ are independent events, doesn't it?

$\endgroup$
  • $\begingroup$ Well, I guess it does. I'm a bit ashamed to admit I'm not really sure. But that could well be it. $\endgroup$ – vedran Jun 19 '11 at 12:28
  • $\begingroup$ @vedran: $p(B \cap C) \sim 0.37$ which indeed doesn't equal $p(B)p(C) \sim 0.43$. $\endgroup$ – Marek Jun 19 '11 at 15:53

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.