I think we mostly agree that a slight tilt is probably best. BUT when I said a "more convex rod tip path", that is entirely different than the micro wrist.
Consider the classic windshield wiper cast in which the beginner takes the rod from 9 o'clock to 3 or 4 o'clock for the backcast and then back again to 9 o'clock for the stop. That rotation of the rod through a long arc of greater than 180 degrees, leads to a very convex rod tip path and a very open loop, even if the rod is traveling along the 90 degree vertical plane to the horizon. So loop size is controlled by the both rotation of the rod through an arc which affects the path of the rod tip (convex, straight or concave) and the micro flip of the wrist right after the stop.
You would think that the rotation of the rod through any arc would always cause the rod tip follow an arc or convex path. This would be true IF the rod did not bend. The bending of the rod while the rod is rotating, shortens the "effective rod length" by bringing the rod tip closer to the caster's hand during the rotation. It is the balance of the rod shortening and the rod rotation that causes what could be a convex path into a straight line or sometimes even a concave tailing loop.
Since beginners do not properly accelerate their rods, they tend not to shorten the rod, and the rod tip traces that very convex rod tip path.
The tighter the loop the greater the danger of hitting the rod at or below the tip. Make a very open loop casting in the strict 90 degree plane and the fly stays well away from the rod. A 90 degree plane ALONE does not mean the fly is in danger of hitting the rod. There must also be a tight loop to bring the fly low enough to hit the rod.
Regards,
Silver
"Discovery consists of seeing what everybody has seen and thinking what nobody has thought"..........Szent-Gyorgy