Good that you ask, i realize that it was sort of wrong. It should be more like this:
a = 100/360;
pos = value;
r = thisComp.layer(“GEAR”).transform.rotation;
pos[0] += a *(r.value-r.valueAtTime(inPoint));
pos;
it adds something that is proportionnal to the rotation variation to the x component (component 0) of the position. For a = 100/360 it will add 100 pixel for one complete turn, etc.
The first 3 lines simply define some variable to be used.
pos stores the current value (before expression).
The the variation is added to the component 0.
To calculate a variation, one needs a reference value, i assumed that the reference should determined at the layer’s inPoint, but you might need something else.
Then the last line, just ‘pos’, tells what is the final result to be retained.