We say that two events, and , are independent when learning that has occurred does not change your probability that occurs. That is, . Equivalently, and are independent if doesn't change if you condition on : .
Another way to state independence is that .
All these definitions are equivalent:
by the chain rule, so
and similarly for .