The great advantage of such a hooked and blooded Fisherman s knot, is not that that the tails make two additional sharp U turns - it is that the second U turns, the hooked ones, take place in the space in between the two nubs, so they do not increase the cross area of the knot. Regarding this, what Mr Lehman has just tied “is not so stupid”. The two sliding halves press the two embraced bights from both sides, so, as the pressure increases, the friction forces in between them increase, too.
The centrally positioned heavy interlocking of the two tails seemed [i]“a mess”/i to some people - who would have eaten their hat by now. However, the most interesting thing in this knot, in comparison to similar sliding / interpenetrating bends (1), is that, in a sense, it is three knots in line- so its total cross section is not larger than the cross section of each one of them.
A problem with all those bends with crossed tails, in that they can seldom be side-symmetric - and be able to be inspected as easily as the side-symmetric bends can be. ( One other reason I like side symmetric bends is that they require one, and one only, picture, of their one side. 
) The bends shown in the attached pictures are side symmetric - and they, too, belong to the same quite broad class of Fisherman s hooked bends, as explained in the previous post.