In 2D, the great circle route - in other words the magenta line, should always be most efficient.
Don't agree. If the front is in the middle, and perpendicular to your 400 mile journey, and the crosswind component from the right before passing the front is as strong as the crosswind component from the left after the front, it's better to simply let yourself blow away to the left first, as this will be perfectly compensated later on.
When sailing at sea, you do the same thing. Let yourself wash one way with the tide, and the other way six hours later. Trying to compensate for the tide only costs in distance traveled.
However, practically speaking:
- Winds aloft forecasts are notoriously unreliable. (Tidal flow is much more predictable.)
- The perfect situation as above will almost never appear, which makes the calculation on how much to compensate much harder. It's far easier to follow the magenta line, or calculate your plog so that it follows the imaginary magenta line.
- Airspace, intersections, reporting points, placement of navigation beacons and general predictability of your (filed) route will usually be more important than saving a few liters of fuel.