)
WOOF! 
OK I think I'm convinced.
My problem was seeing a wind vehicle with a prop on it and assuming the prop was driving the wheels.
Took me some time to get my head around the wheels driving the prop.
So the prop is actually trying to drive the vehicle backwards when the wind is from behind.
But the wind pressure on the whole vehicle is able to overcome the torque on the wheels from the prop.
Once it starts moving, the thrust from the prop adds to the wind.
At windspeed, there is no longer any force on the vehicle, but the prop is still pushing.
Still sounds a bit "iffy" to me, it's counter intuitive.
About the third time this topic's been up. Always a good topic for a bit of mental exercise, so many ways of looking at it. Here's another one.
Start off by thinking about the cart in no wind.
Give it a good push up to speed on level ground and release it.
You'll find the propellor is turning in a direction to continue pushing the cart - like a taxiing model aeroplane.
The push of the propellor, thrust using aeroplane terminology, can be turned into power and transmitted through the gear box. The amount of power that is transmitted though the gear box is equal to thrust times the velocity of the cart over the ground.
So the issue of what happens next gets down to how much thrust is generated by a propellor versus how much power is required to turn a propellor.
The thrust of a propellor is proportional to the lift generated by the propellor blade times the cos of the pitch angle of the blade. The power to turn a propellor is proportional to the lift of the blade times the sin of the pitch angle. (no drag in this thought experiment).
When you go through a trial calculation (use a simplified propellor! one that neatly corkscrews through the air) you find that the thrust that can be converted to power is exactly equal to the power required to turn the propellor. Always impressed at how conservation of energy is embedded in all these mechanical calculations where friction is ignored.
So with friction a cart pushed in still air will of course roll to a stop.
Now add wind.
If instead of zero wind we have a slight tail wind then the propellor blades can be run at less pitch and still deliver the same thrust. With less pitch angle it takes less power to turn the propellor. With a bit of power at hand to overcome friction the cart is away.
(All well and good thinking after the event - so impressed with the fellows who came up with the idea in the first place)
Ok, that's great. Now somebody design something that will sail dead upwind and all our energy problems are solved.
In my stupid mind I imagine something with two sails, both on close hauls, one starboard the other port. Should work just fine.
