DC, the bows you make with that design are great, and are a strong testament to how good the design can be. I think you're spot on, straight limbs at full draw is ideal. If your gut says string angle is just a symptom, I'm curious, does it have any feelings about what's the cause?
Not really but possibly the actual lever length of the limb. A "lever" is the direct measurement from the fulcrum(grip) to the point of effort(where the string hits the bow) ignoring any bends. With a straight bow the lever gets shorter as the limb bends. Bad. With a reflexed bow the lever gets longer as the bow bends. Better but can be unstable and by the time you get the bow braced you've used up most of the bend. With an RD the deflex gives you your brace height without using up much bend and by moving the tips back it also stabilises the bow some so you can use more reflex. If you can arrange things so that the limb is straightest at FD then I think you would have maximised the lever length.
Also if you can move the the bend toward the grip you can shorten the actual string length. By that I mean the part of the string that isn't touching the bow. Then your braced lever length is short because the lever is measured to where the string hits the bow. This gives you a short lever when the bow is easy to pull and the lever lengthens as you draw it when it gets harder.
That was a lot of typing for a "not really" but those are my thoughts. I get the impression that some others here are swishing roughly the same thoughts around in their heads.
All that said if I understood the physics of bending a beam with a string maybe string angle would mean something 
I agree with most of that, and some of it is what I was trying to say earlier. You say that a reflexed bow's lever length gets longer as the bow bends, which seems quite at odds with me saying that reflexed bows usually get extra short as they bend. But I suspect if we further clarify what we mean, we actually agree about most of the underlying principles here. I'm realizing the explanation of my drawing a couple pages back leaves a lot to be desired.
There are two different ways to use the word reflex that I think make the conversation muddled. Reflex can refer to where the limb curves away from the string. But when I talk about "overall" reflex, I'm talking about how far the tips are in front of the handle when unbraced. I'm using it in the sense people use it when they say like, "This bow kept 2 inches of reflex after tillering". They're just measuring how far the tips are beyond the handle, irrespective of the curves the limbs may have taken to get there. I think that difference in language is important to pin down. As D/R bows show us, a bow's limbs can be reflexed along their entire length without the bow having any overall reflex (if the handle is deflexed enough). Similarly, a bow's limbs can be entirely deflexed along their length, like they took set, but the bow could still end up with overall reflex if the handle is reflexed. So these are quite distinct concepts.
In the explanation of my drawing, I should've been more clear, but I was meaning that the farther limbs bend past straight, the shorter their lever length gets. As my drawing shows, increased overall reflex does mean the bow has to bend farther to get to full draw. For most common primitive-type bows, this increased bend means shorter lever length. This is because the main working area is usually pretty close to straight when unbraced, and it starts bending past straight as soon or nearly soon as you start to brace it. Heck most bow's working areas have some string follow, and are bent a little past straight before you even touch it to brace it. This means that the more overall reflex they start with, the farther past straight they will bend, and the shorter lever length they will have. A bow that has overall reflex due to recurves is better off than a bow with reflex with straight limbs, as recurves can counteract some of the loss, but it would always be even better if the main working areas weren't bending so far past straight. Hence some of the benefits of a deflexed handle. I think we agree on all that, right?
But the bow design we've been talking about, the one Stick Bender posted pictures of, doesn't fit in to this usual category. When unbraced, the working area isn't straight or near it, it's significantly curved away from the string. Essentially the whole limb is a working recurve. If we think of the movement of a limb in 3 stages: Pre-straight, straight, past straight, these bow limbs start pre-straight. Their lever length gets longer as they bend toward straight. Then if they go past straight, it starts getting shorter again. The more deflex in the handle, the later this turning point happens.
I agree with all that you said about string contact and the lever length elongating as you draw. That's a recipe for a fat f/d curve. Though I suspect if you take that concept too far and tried to have the string contacting a large proportion of the limb at brace, some other inefficiencies would rear their head. But I won't get in to that now.
Moving on to lever length and string angle, they're closely connected. You'd have to really work to design a bow with a long lever length and bad string angles, so they're mostly two sides of the same coin. So you could look at lever length as the one that matters, and string angle just as a easy quick glance diagnostic. But I'm reluctant to think of string angles as merely a symptom. The direction the string is pulling on the limb pretty directly matters, and the angle that the string makes at your fingers matters too. As the angle at the fingers gets sharper, you put significantly less force on the tips. I don't want to get to math-y, but the pictures here do a good job of explaining.
https://roperescuetraining.com/physics_angles.php And you can see in my drawing how much better the angle at the fingers is for the deflex handled bow, even with the same limb length.
Also important to keep in mind is something that stuckinthemud was saying in his original post. Looking at a bow braced and at full draw, you can't easily tell how much overall reflex it has when it's unbraced. Overall reflex/deflex doesn't
necessarily matter to the shape of the bow at any point during its use. But I will have to get in to my thoughts on that and why it's interesting at some other time, if anyone is interested.
