Although I am loathe to extrapolate aviation data to the nautical world the physics involved are very similar. The main difference is the incompressibility of the fluid that a boat
propeller works in.
In powered operation the
propeller meets the water just on the concave side of the leading edge. The concave side of the prop "paddles" water aft and some water rushes around the leading edge. In air this creates
lift thanks to bernoulli. In water it must create some small drag. A big difference is that most boat propellers are not airfoil cross sections because the medium is incompressible. But ultimately there is a thrust vector, most nearly at right angles to the blade.
Lot's of interesting things must be happening on a props that are not mounted along the longitudinal axis of the boat - i.e. the shaft angles down. In this case the blades going up and the blades going down see a different angle of attack to the oncoming water and therefore must produce different thrust - mathematical ice cream headache...
When the prop freewheels water strikes the blade on the convex side. The blade rotates and the unbelievable thing is the thrust vector goes negative. Negative thrust is drag. Added to this is the hp required to turn any machinery such as a
gearbox and
engine.
Another difference in boats is that in most cases you can disconnect the machinery, at least the
engine and only turn the relatively easy to turn
gearbox.
So bottom line is that you have to follow your manufacturers recommendations or do your own testing and weigh the factors.
One final thought regarding the autogyro. It works right? But the forward thrust engine is used to overcome the drag of the rotor. If the engine fails energy stored in the rotating rotor is converted by angle of attack to create a flair for a safe landing.
How well do you think that would
work if the gyroplane's rotor was locked in place?
For an interesting
cartoon about windmilling and the forces have a look at this.
Drag From Windmilling Propeller