[...]
I didn't say that. Mathematics of probability allow you to set up scenarios of various states of information, i.e. also what the probability *would* have been if this and this information was known.
Probability estimation always assumes a vantage point; with some information available and other information not available at the vantage point. Hence, a probability is not necessarily universal, but depends strongly on the vantage point. Bayes was probably the first to put that into a formal context.