It's my understanding that NVFR was originally intended as a means of letting VFR aircraft get home a bit after last light, or launch a bit before first light.
As we all know, though, it's possible - probable even on a lot of nights - that flying at night can involve no visible horizon, not a lot of ground reference features and very limited attitude information outside from stars, ground lights, moonlit terrain or whatever.
My point is that for an adequate safety margin on many NVFR flights, you need to be as good at instrument flying as an IFR pilot - probably better at hand flying on instruments seeing as you probably won't have an autopilot.
So, peuce, the line between visual and instrument flying blurs significantly in the NVFR world as I'm sure you know. I'm not suggesting anything in particular about what happened in this Moree accident, but I know that I've flown on a lot of dark nights where some aspects of a circuit were far more instrument than visual.