Applied Probability Theory

Applications of the Probability theory (which is pretty straightforward - the same universal weighted sums) is the one of the most bullshittted topic out there.

The main point is that, in principle, uncertainty cannot be reduced ahead of the time, or, to put it simple - probabilities cannot predict or forecast anything, only provide a better view (refined beliefs).

The only valid applications are those which update “weights” with every new event. This, of course, has all the parallels with Neural Nets, Kalman Filters, etc. Again, only reactive, never predictive.

And yes, most what has been written about application of probabilities is utter bullshit, because

• events are not independent (the major cause of bullshit)
• not all possible outcomes has been taken into account
• initially assigned probabilities are wrong (flawed models)

Finally, after many years (not an exaggeration) I found a non-bullshit, non-hand-waving probability course, with video lectures and a textbook