Hi Paul, I detect a degree of frustration here. So let me try and help.
First there are 2 types of visualisation.
1) The static rendering used for commercial /architectural applications. This type uses accurate grids where each point is calculated from the photometric file for the fixture for each fixture. This requires quite "maths hungry" processing and can take from a few seconds to several minutes (even hours) for large complex lighting designs. Commercial programs like AGI 32 is an example of this where the calculated results should be within 2% of actual recorded light meter readings, as large multi-million dollar contracts can be signed on the backs of these results.
2). Then there are the "real time" rendering programs which can be applied to large numbers of lights that can change all their attributes in (virtually) real time. This by necessity has to be a compromise, if the maths is to be crunched in "real time" in most cases this is an "indication or what one could expect to see"
The point is, its an indication, not actual. I certainly would not place any of the real time visualisation programs in front of a client and say "this is exactly how it will look" what I do say "this will give you an idea of what to expect" and" wow" them with all the lights moving/changing etc.
To answer your points about Capture.
Think of it as proportional lighting design ( not actual). Whatever the % you set, all the fixtures will be "in proportion" your Par 16 will be about 6% of the intensity of the 1K so you have to adjust the background / light output smoke etc to get the effect you want. The program will not tell you if you have /have not enough light. As a LD you should be able to assess that by experience.
Hope this helps
cheers
Steve