If you want to test bend radius you need to have a line with two eye splices with both of them on the same side of the pull and a bend radius on the other end of the pull. You will have to pull twice as hard to break it though.
Not sure I understand what you are saying. In my last photos above I created two pieces, each with two splices. So, I had two pieces exactly as you say above "a line with two splices with both of them on the same side of the pull'. And exactly as you say the bend radius was 'on the other end' - eg it was a 1:1 bend radius created by the two joined splices. So, I think/thought I was doing exactly what you are suggesting here. Are you suggesting something different . . . if so please explain.
edit: perhaps what you are suggesting is that I could use the same two pieces, but instead of interlocking the end splices, I interlock the middle part, with the two splices (on one piece) both pulling from one end and similarly for the other piece. That would halve the loads on the splices (or double it on the 1:1 junction). That's more like the question of bend radius in a working sheave - than what I was trying to get at - bend under a stationary loop.
If you have a line going through an end fitting and it is carrying 100% of the load then the bend radius will matter. But in the eye of a splice, probably not so much.
That quote from Colligio is re life lines and I am quite sure was intended to apply at the bend radius for the 'main' line spliced around the outside of the fitting/thimble. They go on to say "Our pull testing has
shown that, with a proper splice, the Dyneema line breaks at the bend" . . . . note that's specifically a stationary splice. My test results seem to directly contradict this statement.
Looking at samson's literature . . . it is a little ambiguous . . . but suggests more like what you say and my testing shows . . . 3 or 4:1 for a 100% loaded working line (like a mooring bollard), and for a stationary splice loop they are concerned with the angle at the throat of the splice (so the ratio of splice length to hardware width) rather than the bend radius. . . . except they do say this "The diameter on fixed pin terminations should be at least 3 times the rope diameter (i.e. the bending radius for 1/2" ropes should be 1-1/2")." and I am wondering by "fixed pin termination" they mean in effect a spliced loop.
But it could be that some people have mis-interpreted the industrial bend radius guidance to apply to fixed spliced loops when it was not intended to.
As another obvious point, if you are using a knot, then the bend radius is truly insignificant because you have lost so much strength in the knot.
Yes, of course