Using a shield to assist swap of polarity in a magnet

[Low-Q]

I have done some tests with one single magnet which I guided inbetween two other magnets. I split the experiments in two sections. On in attraction mode and one in repelling mode.

First experiment went like this:
The "rotormagnet" feels a pull towards the statormagnet when it is outside, feels nothing in the middle, and are pushed back when it is closing the shield.

I swapped polarity, and the opposite forces are applied.

Green arrows are the direction of the force when using one single magnet as "rotor".

EDIT: I forgot to say that it was no counterforce when I tried to rotate the magnet insode the shield. After a few attemts I finally made an easy way to rotate it without the magnet getting stucked to the walls inside the shield all the time

Maybe a continous array of magnets will behave differently, as the magnetic field then will be forced from being spread in all directions around the disc magnet, into a field which is allways parallell to eachother?

Vidar