A better unit vector

[Raoul-Kima]

At first glance this looks to me like another way to do what I did in the thread over here:

Divergence /Treedepth issues with unit_vector

I’ve seen that there were several discussions about divergence/treedepth problems when using unit_vector for circular statistics in the past year. E.g.: Has this issue been resolved in the meantime? If not, I think I might be able to explain the cause and give a solution. If there’s interest I can write It up and post some theory and a small case study here.

Is that correct?