Q1: I have better experiences with 2 trains that serve the 2 station on one line, then 2 separate lines with 1 train each. But is my personal subjective observation, no exact science. 😉
Q2: If C and D are on one common line C-A-D, you don’t need to connect C and D directly. If you have 2 crossing lines: line C-A-D and E-A-F, then it might me beneficial to make a connection between the surrounding cities (eg C-E, D-F), but it might put in danger your 2 main lines. I would not make the extra connection unless the stions are more the 20-minutes away from city A
Q3: one big central station reduces the necessity (and time wasting) to get from station to station by bus or tram. Several smaller stations around the city can encourage city grow, but I would maintain a central hub.
Q4: Passengers station near residential quarter is ok, but cargo lays better near industry.
Q5: you don’t know. The program will select at random an new destination. There is no general habit, fixed work or other pattern programed for the moment. So for instants each “day” a person is leaving fore work it will be another work place. So, with this kind of behaviour, it is impossible to predict how many people are going to take the train to a nearby city. I would personally like to see this changed, and for instance have that people have near-fixed work places. Only Then one can make traffic statistics and predictions.
I hope my humble remarks can help you in developing a railway network. Succes and have fun! Train Fever is no exact science, but lot of testing en trying to figure out how things work. 😉