An increased number bears no relationship to increased realism.
This is an overstatement also. It's a relationship very much dependent on the instrument and very many other things besides. But its still an important relationship.
I've bough string libraries only realize that they lack a decent soft layer, and that therefore the instrument just doesn't hold up for any soft exposed passages.
Worse, the fake dynamics then give it this uncanniness that your average listener wouldn't necessarily understand as unrealistic, but the uncanniness has this, kind of ... unpleasant ... quality.
And at least arguably worse still, I've bought instrument that sounded in the demos like they had decent dynamics, but that turn out to have simulated dynamics on 1 or two dynamic layers. And so again when I try to write something in a style that doesn't minimize the dyanmics as in the demos, its not that its obviously 'unrealistic', but that this is the fake dynamics can really give this quality that I find particularly unpleasant in its uncanniness - precisely because its not always immediately obvious that its the 'fake' dyanmics giving rise to this uncanny quality. Its like an itch that you can't quite locate in order to scratch it.
Of course, there are lots of contexts in which the number of dynamic layer don't matter. I mean you can always just add lots of fff trombones, in which case the dynamic nuances of the p -> mp of your lyrical oboe isn't going to matter so much. But I find I tend to like to (try to) write in a way to leans on dynamics perhaps a bit more that your average demo.
Similarly, the OT exp B winds do sound truly wonderful - so long as you play within the constrains of the single dynamic layer. But this is a constraint that I think would really bother me at times given how I like to write - not that I wouldn't also love to have them for the times I would be ok with staying within those dynamic constrains.
So in this context, I'm sure you'll agree that 'how many dynamic layers?' it at least a fair question, no?