How do the following compare in terms of autofocus accuracy for night photography:
Mirrorless cameras like the Sony NEX-5R, in hybrid autofocus mode.
The NEX-5R, in contrast-only autofocus mode.
A typical Nikon or Canon SLR in live view mode.
The SLR when using the OVF.
Note that this question is only about accuracy (what percentage of photos are mis-focused?), not the autofocus speed (I can certainly wait for a second for the camera to focus). In other words, I want to avoid situations like Is it normal for autofocus to produce blurry photos under low light? .
Am I correct that the order is (4) > (1) > (2) > (3), where > means "is more accurate than"? Or, to put it differently, phase-detection > hybrid > contrast-only?
Am I further correct in assuming that anything other than a contrast-only system works well most of the time (say, produces correctly focused photos 99% of the time)?
As I said, this question is about night photography (I have had no issues with mis-focus during daytime, with any of my cameras). It's also about a static scene like a landscape, and not, say, sports (which is why I specifically excluded speed of focus from this question). The scene does have a decent amount of contrast (bright streetlights and dark fields), unlike, say, astrophotography.
When I said SLR, I mean typical digital SLRs exclude specific brands, and Sony SLTs from this discussion, so that we discuss the big picture (pun not intended) and not get side-tracked into discussing the pros and cons of each of these camera systems.
Similarly, when I said mirrorless camera, let's please discuss only the NEX, again to keep the discussion focused (pun not intended) and not get into Sony vs Olympus vs Panasonic minuteae.