AWC,
The trick is striking a balance of security, convenience, and authenticity.
In order to be secure, the mail needs to underlap the plate by a couple of finger widths, more or less, depending on how much motion is expected. By the time you figure a safe underlap from the elbow, and the armpit, there is not really much that can be removed in between.
I am pretty sure I can justify removing the mail on the top of the arm. I believe that is what we are seeing in the Hastings Manuscript illustration.
What you suggest is essentially two separate gussets; one for the armpit, and another for the elbow. If we were dealing with an articulated elbow, and the amount of mail needed for the elbow was therefore rather small, I think that would be a good choice. Here, though, we have to make sure we cover all of the potentially open places around a "floating" elbow, front and back. That requires quite a bit of mail. By the time we are talking about such a large and sleeve-like elbow gusset, it just seems (to me anyway) more reasonable to connect it to the armpit gusset.
That said... This whole area of gussets and authenticity is largely
terra incognita. There's very little information to go by, and we really don't know what they actually did. I am making what I think and hope are reasonable decisions, but we may never know for sure what was usual and typical historically.
Mac