Probability and Random Variables
We consider a basic set Ω, whose elements consist of all possible outcomes of an experiment, irrespective of whether this experiment can actually be performed or is only imaginable. A single performance of this experiment is called a realization. It yields an element ω in Ω.
We now want to identify events as certain subsets of Ω. One may think of events as the sets {ω}, which contain one single element ω, and as such represent elementary events.
However, one may also think of other sets which contain several possible outcomes, because the probability that the outcome of a realization belongs to a certain set of outcomes might also be interesting.
A more detailed mathematical analysis reveals that in general not all subsets of Ω can be considered as events to which one can assign a probability. Only for certain subsets, which can be made members of a so-called Borel space, can one always consistently introduce a probability. Here, a Borel space is a set β of subsets of Ω, for which: