the possibility of predicting the future

[reor]

I guess probability is the best we can get. There IS something like a "main pattern", you can see examples everywhere in nature, but an exact prediction (100%) will never be possible, due to data loss on every observation. We can't possibly get EVERY needed information and there will allways be loss.

 

As to "free will": There is a nice long thread here:

[scidata]

Psychohistory based on formulae as in "Foundation" is very doubtful. However, computational psychohistory based on models and algorithms is more of a possibility.