In this case, some flux lines are lost inside the iron between the STATOR MAGNETS, because they have different B field values. This causes a decreasing field density, and an increase on the variation of energy as the MOVING MAGNET moves:
The force magnitude decreases and has an irregular profile, but it's always directed to impulse the MOVING MAGNET on the right direction.
CLOSING THE LOOP:
I think all the process is based on a local effect between the MOVING MAGNET and the TRACK of STATOR MAGNETS, so a closed loop must work too:
To demonstrate this fact, let's do a closed loop like this one:
In this case, each MOVING MAGNET has a different position in respect of the STATOR MAGNETS, so the total torque must be zero if the additive effects of each MOVING MAGNET compensate.
The flux lines inside the magnets are like these ones:
The energy on STATOR MAGNETS increase clockwise near the MOVING MAGNETS and anti-clockwise on all MOVING MAGNETS and STATOR MAGNETS when no MOVING MAGNET is in between.
see the torque (N*m/m) on each arm of this 'clock' of MOVING MAGNETS:
The torque about 0,0 may be near to zero if the system doesn't work, but as you can see, the STATOR MAGNETS always give some positive torque to the MOVING MAGNETS, so the rotor can turn without any other energy to apply!
As an example of the working of this device, imagine the same closed loop, but with only one MOVING MAGNET in between, like this: