harrym @harrym - 8mo
The history determines the environment and therefore the current prob distribution for whatever happens next. When that distribution is sampled at an event, you have zero control over the value of the sample you get. That's pretty relevant. The finiteness of timelines is irrelevant (in fact its doubtful they are infinite, at least in the negative time direction). 100 heads is fine, of course, just unlikely.
