Haha! Not a bad summary. I hope this is more helpful than confusing
You can believe both the historgram and the light meter. They do not lie. The histogram is closest to the real thing because it's derived from the actual image file, but it's also important to know that it is a JPEG (it has to be, you can't actually see an unprocessed Raw file) and therefore adjusted by the camera's processing parameters, ie Picture Styles etc.
The only thing I would really be wary of is the Contrast setting in Picture Styles, which can shift the histogram over one full stop left/right with identical exposure levels. To see what I mean, shoot two identical images, one with the Contrast at max and one at minimum and you'll see it clearly enough. The point at which blinkies (highlight over-exposure warning) cut in is also changed with the Contrast setting.
For this reason, some people set everything in Picture Styles to zero so that the resulting JPEG-generated histogram is as close to a neutral Raw as possible but I don't do this because it wrecks the resulting JPEG image if you want to skip any Raw processing yourself and go straight to JPEG output. The important thing is to know what parameters you have set, to keep them constant, and know (by experimentation) how much lee-way and headroom you've got if you want to go tweaking things, eg exposing to the right.
Much the same applies to using a separate hand-held meter, either in reflective mode or for incident readings. The important thing is to know how the thing works, ie to know that whatever you point it at, using whichever method, it will always scramble whatever light you've pointed it at and pin it to to the 18% grey level. It doesn't actually matter if the camera is calibrated slightly differently, maybe for 12% grey, or that different sensors react differently just as different film emulsions do. The point is that it is constant, you know what it is doing and you know how to interpret it because you've experimented with it in combination with your camera.
Phew. That's the theory of it as I can best explain, but is sounds horrendously complicated like that. And while exposure setting is an exact science in theory, in practice it is quite a long way from it. I mentioned some of the tricky variables earlier - a lens diaphragm that is a fraction of a mm out which can make maybe a third of a stop difference at a small aperture like f/16; who says that ISO400 on one camera is exactly 400 on another (often it is not); vignetting - use a wide zoom at low f/numbers and the edges of the frame will be a stop darker than the centre; f/numbers themselves are theoretical and the light transmission actually hitting the sensor, the T/stop, can be a third of a stop out; shutter speeds are not always accurate although I think they are pretty good nowadays, but they used to be notoriously inaccurate at high speeds; the flash output might not be consistent shot to shot and in fact it certainly won't be if you're using a hot-shoe gun and shooting fast before it can recycle fully; and so it goes on, last but certainly not least is how you use the meter (assuming it's accurate!) what you point it at and how you interpret the reading.
So, after all that, if you can get your exposure to within half a stop of optimum, in practise that is actually as good as it gets due to all manner of constantly shifting variables that you cannot hope to keep dead accurate track of.
If you really want or need to be any more accurate than that then you should bracket your exposures, maybe +/- half a stop either way. Just like film! It's very easy with auto-bracketing, even if I might think it's a waste of time LOL
This is a good link - maybe I should have posted it earlier 
they didn't make them easy did they!?!