> That's very limited frequentist perspective. Just because you can't re-run the experiment doesn't mean you can't assign probabilities to events

One time events have no probability, they happened exactly as they happened and no outer outcome can be assigned probability besides 0