I have found it really worthwhile to add a midpoint into the track when none is actually required, to check out what might happen if (say) we went outside/inside the current/weather system when conventional routing wisdom suggested the opposite.  It has to be a race of x miles, large enough that system of current changes over time make it worthwhile - in my part of the world think Syd-Hbt or Syd-Southport distances.
The big question as a navigator is how you decide this might be something to consider?  We use a meteorologist, and a good one, to cast the runes and speculate about big picture data which we then pull apart.  And then work from there with the tweaks.
A great question.