I looked at that power point you referenced. The failure modes are different in the ppt vs the paper. In the paper, the mode of failure was primarily in the transition from the bury to the "clear" line due to uneven stressing on the fibers at that location...the eye held though. That indicates that the loops on either end were strong enough to shift the weak point somewhere else in the rope, IE causing a failure in the rope itself as a function of the best-case splicing or end loop construction, not due to the rope being wrapped around something too narrow. the ratio for D was at least 3:1, indicating a useful method for retaining strength of the sling.
The ppt test where D = 1:1 or 2:1 yielded a different failure mode where the rope failed at the point it was wrapped around something, indicating that the weakest point was at the bend around a pin, not at the splice point.
I don't think the two sources contradict each other. The failure modes were different in what seems to be a clear result of a function of size D.
What we can say with some confidence is that the tighter the bend, the more the rope is prone to failure at the point of attachment, vs overloading the min (useful) breaking strength of the rope itself.
on another note I went to ace hardware over lunch to see a 3/8" quick link, since 3/8" is a little more than 3:1 for 7/64ths, and the size and weight made me realize why you're fighting 3:1 or 4:1 so much lol!
Anyone got any ideas on a carabiner or quick link style of attachment that's 3/8" in diameter that's made of aluminum or something lighter than steel? There's no way I'm using 2 3/8" quick links!