FWIW I have only got a single white screen event sometime in early spring during boot on my intel 4770k iGPU self built system running nightlies on bare metal.. And was able to get past with a two letter command that I have since forgotten and an update two nights later fixed it.
Even the nightlies are pretty stable these days… But if it’s working for you right now, I don’t see the point of using nightles.
For many of us in higher latitudes north of the equator a new release (why don’t we do intermediate releases such as Beta 5.1?) would be awfully nice right now. Since there are longer darker evenings to do projects such as migrate onto Haiku.
Good question, actually.
When there are major improvements a new beta and otherwise minor release?
What is major and what is not? There will be a lot of debates…
Preparation of the release itself takes time, the name won’t change that. So, it won’t accelerate the cycle. Though, it could avoid to reach beta127. 
I think setting two release windows per year and set goals. One major goal per release, two if first goes well, then time needed to fix relative bugs and regressions, few visuals improvements and other bug fixes to make people happy and to wait the release window. If a window is missed don’t push commits for major changes.
Like that you are accelerating the cycle a bit and people know when they can expect new one. It’s not important for forum followers but, for those who will check progresses once a year, it can make a difference.
Because the beta releases are already intermediate releases to R1. It would make no sense to make intermediate releases of an intermediate release. Just get the disk image from Beta5 and run pkgman update or softwareupdater to get the latest bugfixes. There is nothing release-worthy there.
The only time we did one (for R1 alpha 4.1) was because there was not enough testing on Alpha4, and there was a major regression making the install media unbootable on several machines. So we made a fix release before pressing the DVDs.
Remember that the developers work on their free time (well, most of them) and so they work on whatever they want. Otherwise, it feels too much like a paid job and we’ll go do something else for fun.
There are two options to fix that:
When I say major goal, it’s not like making 3d accel working for all cards magically.
For me, something major is rather a change that will have an impact on what is coming next. I consider that changing memory allocator is a good example. Debugger and HaikuDepot improvements that we have seen lately are some others.
I didn’t mean everyone has to work exclusively on that major goal either but when that one is achieved, it’s time to prepare the release.