Hello shadow, Welcome to LessWrong! I think the answers to your questions are as follows:
1) Working out exact probabilities for these kinds of predictions is unfeasible, so we approximate. Rounding to the 5 seems natural to a lot of us, and (I expect) it automatically conveys the approximate-ness of the prediction to most people.
2) Probability is in the mind. Therefore, giving a distribution of probabilities is just a way of saying "I am uncertain about how uncertain I am".
4) Don't think I understand this one, although this might qualify and/or interest you.
As we did last year, use this thread to make predictions for the next year and next decade, with probabilities attached when practical.
Happy New Year, Less Wrong!