A remark on probability spaces (Q751694)

The question underlying this paper is: How does one usually construct a probability space to represent a random experiment? Starting from elementary events, which become the points of a set \(\Omega\), one is led to the interpretation of the singletons of \(\Omega\) as events. Since the Kolmogorov model does not require this interpretation, one might well ask whether it is more general than the model obtained from the elementary events. The authors show that, in the general (i.e. not necessarily countable) case the answer is in the negative.