You are totally correct in that we cannot reproduce the exact same coin toss due to the minute differences in force, wind, etc. However, for the level of precision that we assign to coin tossing, we assume that these are identical repetitive events because they are close enough.
In the case of a diver fatality, the differences between individual events is orders of magnitude different than those between coin tosses..
You may "teach" statistics, but only in the mathematical sense.
You should try working with some public health epidemiologists sometime. Events do not have to be repetitive or without confounding variables to have statistically meaningful results.