Don’t like it. In the subsequent decision/analysis tree I then have to track both (or even more) the branches. A few ambiguous nodes and from one line through the graph you have half of the graph full. You can prune some of the branches by careful work with getting new data and/or by a cutoff for low-probability possibilities, but that’s it.
Ambiguity adds uncertainty and a lot of work, and in case of high stakes also a lot of fear/anxiety.
