For a Bowline with spread apart force in loop this is my choice:
.
i think Bowline is based on a Half_Hitch, that is designed to pull INLINE;
not spread on cross-axis, so would change orientation of tool to match force flow.
.
So i think, as we switch to cross-axis loading across the loop;
we'd turn the inline structure of the Bowline to match.
.
This might reference as a 'Cowboy' visually because 'tail outside of eye':
BUT; mechanically it is the same structure; both 'tails'exit the same side; both tails sandwich between primaries.
.
also note: how the longest tail; that is usually on outside of ring it forms,
>>is now on the inside; just as now it is the tail and not the sPart;
>>thus another reverse, in this reverse situation!
.
noting pull forces are different in this form; there is no sPart pull that is 2x tail/release(so more like Sheet Bend in that aspect)
>>not visualizing this as binding not rather a 'splayed'/spreading force
.
strain spreading on loop would go to solid side primarily, from any stretch , shift, adjustment in Bowline as weaker leg i'd think.
.
And to original question:
i'd 'DBY' any or these commonly by habit.
>>Double/Round Turn Bowline with Yosemite finish
(favor the inversion method for easier threading thru dbl.rings)