# conditional probability chain rule?

I am aware of the general conditional probability rule which says that

$P(ABCD) = P(A|BCD)P(B|CD)P(C|D)P(D)$

But is there any situation where one can write

$P(A|D) = P(A|B)P(B|C)P(C|D)$ where $A,B,C,D$ are random variables.

Thanks

-
Try $A\subseteq B\subseteq C\subseteq D$. –  Michael Greinecker Feb 13 '12 at 22:56
@suresh Per faq, usage of signature is not recommended, so I removed it. –  Sasha Feb 13 '12 at 22:57
@Sasha, I was not aware of this. Thanks for pointing it out –  suresh Feb 13 '12 at 23:02
Actually, you'd want $A,B,C,D$ to be events rather than random variables. –  Michael Hardy Feb 13 '12 at 23:39
In general, you run into the problem that you can have both events $A$ and $D$ occur without $B$ and $C$ occuring. If the events are nested, this problem can not occur. –  Michael Greinecker Feb 13 '12 at 23:40

## 1 Answer

This works with Markov chains. It's essentially the definition of a Markov chain.

-