There's more reasons why we put the half hitch (or Marl) in the pieces being chunked off than to dissipate the load.

One of the chief reasons is that we want to minimize the distance between the rope's attachment to the cut piece and the block. Whilst TA&SOPR video series demostrated that the fall distance doesn't change it failed to demonstrate with a dynomometer at the block the difference of having a close to block half hitch. If the half hitch is only 1' away from the block then there's going to be less force on the rope than for example if the bowline is say 5' away. Imagine as the block comes off the fall distance of the further tie point will result in additional force at the block. Whilst this whole theory is debatable without evidence I can only also conclude with a parallel example that of a swing. Same weight but further the swing seat is from the fulcrum more pendulum force is exerted. Tests should be done.

Another is that we do not want the cut piece to invert, likely to happen if the bowline was tied past the centre of the piece.

Another is we want the block to come down long and not sideways spinning and carrying on.

Another is security of having two ties ... but you can have more than one half hitch along the piece.

Regarding the second question of the bowline failing ... of course there's a higher chance without the half hitch as there's more force on the knot. But that said it shouldn't fail unless you have overloaded the system and bombed off more than 10% of the ropes SWL!

So if you are rigging pieces off and your bowlines are failing by breaking the rope then cut smaller pices. However if your bowlines are slipping off then do another round turn ... I tend to go round twice that way even less force is put on the knot.

Also there should be tension on that rope as it goes over, a slack knot is asking for it.