How the stator magnet might "warp" past the sticky spot. Domains only turn so far, then they would bounce back in to posistion. Bring the stator accross a weaker magnet so that it drags the domains behind it, the domains will face towards the spot that the stator magnets was just 10ms ago, so you'll probably have to move in fast. This is hard to imagine because it's not physicaly dragging it, it's draging the domains in time. After the stator gets past the stickey spot, while thanks to time it didn't have to work and lose energy through it all, it stops and then the domains spring back (because they're allowed time) and push it out with a net gain in energy

Here are some pic's to explain, the red square is the stator magnet, the blue and red lines are the domains of the weaker magnet, and the dotted red square is the place in time that the domians are reacting to.

starts like this, with the weak magnets domains in their natural direction.
(http://img148.imageshack.us/img148/219/untitledlf4.png)

the domains react but after a lag time, so it's as if they're reacting to where the magnet used to be, the dotted red square.
(http://img412.imageshack.us/img412/3598/untitled2od4.png)

where the magnet stops to wait for the domains to give up their reaction to the area it was 10ms ago.
(http://img148.imageshack.us/img148/2667/untitled3gb4.png)